epri power electronics-based transmission controllers reference book (the gold book)

EPRI Power Electronics-Based Transmission Controllers Reference Book

(“The Gold Book”)

2006 Progress Report

1012414

ELECTRIC POWER RESEARCH INSTITUTE 3420 Hillview Avenue, Palo Alto, California 94304-1338 PO Box 10412, Palo Alto, California 94303-0813 USA

800.313.3774 650.855.2121 [email protected] www.epri.com

EPRI Power Electronics-Based Transmission Controllers Reference Book

(“The Gold Book”)

2006 Progress Report

1012414

Technical Update, December 2006

EPRI Project Manager

A. Edris

DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:

(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S CIRCUMSTANCE; OR

(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS DOCUMENT OR ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.

ORGANIZATION(S) THAT PREPARED THIS DOCUMENT

EPRI

This is an EPRI Technical Update report. A Technical Update report is intended as an informal report of continuing research, a meeting, or a topical study. It is not a final EPRI technical report.

NOTE

For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or e-mail [email protected].

Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc.

Copyright © 2006 Electric Power Research Institute, Inc. All rights reserved.

iii

CITATIONS This document was prepared by

EPRI

Principal Investigator A. Edris

Contributing Authors B. Andersen L. Barthold L. Gyugyi P. Lips

This document describes research sponsored by the Electric Power Research Institute (EPRI).

This publication is a corporate document that should be cited in the literature in the following manner:

EPRI Power Electronics-Based Transmission Controllers Reference Book (“The Gold Book”): 2006 Progress Report. EPRI, Palo Alto, CA: 2006. 1012414.

v

ABSTRACT EPRI is sponsoring development of a first edition of the EPRI Power Electronics-Based Transmission Controllers Reference Book. The reference book will provide a broad overview on power electronics-based controllers—with information on historical perspectives, basic design considerations, factory testing, site installations, commissioning, operating performance, operation and maintenance, and future trends. The book will assist users in planning, developing, installing, and utilizing this technology to enhance the controllability and increase the power transfer capability of transmission systems.

Progress to Date

Work on the new reference book began in 2005 and will continue through 2007. This 2006 Progress Report incorporates work completed through December 2006 and is intended to provide the preliminary foundation and guidance for further development of the reference book. It includes five completed draft chapters—nearly 300 pages of text—and 10 detailed chapter outlines. Together, these chapter drafts and outlines establish the basic organization and structure of the reference book. Project advisors and contributing authors will find this material useful in planning and developing additional content.

Each chapter is intended to be largely self-sufficient, containing all the necessary information needed for understanding the subject matter. The information will be tailored to potential users to understand the essential principles and issues involved with credible comparisons, sufficient practical data, and accumulated installation experience to help them in the evaluation, potential acquisition, and use of the “right” transmission controller for their specific applications.

Part of a Colorful Tradition

The new EPRI Power Electronics-Based Transmission Controllers Reference Book is being developed in the tradition of the landmark series of EPRI transmission reference books, first published in the 1970s and 1980s, and currently being updated. These books—on overhead line transmission, underground transmission, wind-induced conductor motion, and compact line design—brought together leading experts to compile state-of-science information on advanced research in these areas. Most notable of these volumes was the EPRI Transmission Line Reference Book: 345 kV and Above. First published in 1975 with a red cover, the book became commonly known in the industry as the “EPRI Red Book.” In this tradition, the new reference on power electronics-based controllers will be printed with a gold cover and be referred to as the “Gold Book.”

1

CONTENTS

1 ALTERNATING CURRENT (AC) TRANSMISSION SYSTEMS (DRAFT).............................1-1 1.1 Historical Perspective.....................................................................................................1-1 1.2 Fundamental Relationships of AC Power Transmission ................................................1-4 1.3 Transmission Problems and Needs: The Role of Power Electronics...........................1-35 1.4 Transmission Compensators and Controllers ..............................................................1-37 1.5 Issues and Options for Future Transmission Control...................................................1-57

2 POWER SEMICONDUCTORS AND VALVES (DRAFT) .......................................................2-1 2.1 Basics of Semiconductors..............................................................................................2-1 2.2 Basic Power Semiconductor Switch Types and Their Characteristics...........................2-3 2.3 Comparison of Basic Power Semiconductors ..............................................................2-11 2.4 Trends in Power Semiconductor Characteristics .........................................................2-22 2.5 Functional Switch Requirements in High Power Transmission Controllers .................2-23 2.6 The Need for and Issues of Semiconductor Valves Using Series-Connected Devices2-23 2.7 Design Considerations of Thyristor Valves for Reactor Control...................................2-25 2.8 Design Considerations of Thyristor Valves for Capacitor Switching ............................2-29 2.9 Design Considerations of GTO Thyristor Valves for Voltage-Sourced Converters......2-31 2.10 Design Considerations of Transistor-Based Valves for Voltage-Sourced Converters2-36

3 AC / DC CONVERTERS (OUTLINE) .....................................................................................3-1 3.1 Operating Principles of Switching Converters................................................................3-1 3.2 Role of AC-DC Converters in Transmission Systems....................................................3-1 3.3 Voltage-Sourced AC-DC Converters .............................................................................3-1

4 GUIDE FOR AND DEFINITIONS OF FACTS CONTROLLERS (OUTLINE) .........................4-1 4.1 Basic Categories Determined by the Type of Semiconductor Valves Employed...........4-2 4.2 Categories by Functional Connection ............................................................................4-2 4.3 Categories by Main Application Function.......................................................................4-3

5 SHUNT COMPENSATORS (SVC AND STATCOM) (OUTLINE) ..........................................5-1 5.1 Objectives of Shunt Compensation................................................................................5-1 5.2 Concepts for Controllable Var Generation .....................................................................5-1 5.3 Thyristor Controlled/Switched Schemes for Static Var Compensators (SVCs) .............5-1 5.4 Voltage-Sourced AC-DC Converter Schemes for Static Synchronous Compensators (STATCOMs) .......................................................................................................................5-2 5.5 Compensator System Control Scheme for Network Shunt Compensation....................5-2 5.6 Comparison Between SVC and STATCOM Schemes...................................................5-3 5.7 Application Examples.....................................................................................................5-3 5.8 STATCOM for Arc Furnace Compensation....................................................................5-3

6 SERIES COMPENSATORS (TCSC AND SSSC) (OUTLINE) ...............................................6-1

6.1 Objectives of Series Compensation ...............................................................................6-1 6.2 Concepts for Controllable Series Compensation ...........................................................6-2 6.3 Thyristor Controlled/Switched Schemes ........................................................................6-1 6.4 Voltage-Sourced AC-DC Converter Scheme .................................................................6-2 6.5 Comparison Between the TCSC and SSSC ..................................................................6-2 6.6 Special Application and Control Considerations for Damping Electromechanical Oscillations...........................................................................................................................6-3 6.7 Application Examples.....................................................................................................6-3

7 VOLTAGE REGULATORS AND PHASE SHIFTERS (OUTLINE) ........................................7-1 7.1 Objectives of Voltage Regulators and Phase Shifters ...................................................7-1 7.2 Concepts for Voltage Regulation and Phase Shifting ....................................................7-1 7.3 Thyristor-Controlled Voltage Regulator..........................................................................7-1 7.4 Thyristor-Switched Circuit Schemes for High Power Voltage Regulators and Phase Shifters .................................................................................................................................7-1 7.5 Voltage-Sourced AC-DC Converter Scheme .................................................................7-2 7.6 Comparison Between the Thyristor Controlled/Switched and Converter-Based Voltage Regulator and Phase Shifter Schemes ................................................................................7-2

8 GENERALIZED AC TRANSMISSION CONTROLLERS: UNIFIED POWER FLOW CONTROLLER (UPFC) AND INTERLINE POWER FLOW CONTROLLER (IPFC) (DRAFT).8-1

8.1 The Unified Power Flow Controller (UPFC) ...................................................................8-1 8.2 Interline Power Flow Controller ....................................................................................8-29 8.3 Structural and Rating Considerations ..........................................................................8-41 8.4 Protection Considerations ............................................................................................8-54 8.5 Application Examples...................................................................................................8-59

9 CONVERTIBLE AND EXPANDABLE STATIC COMPENSATORS (OUTLINE)...................9-1 9.1 Need for Convertibility in Evolving Utility Environment ..................................................9-1 9.2 Concept of Convertible Static Compensator (CSC) .......................................................9-1 9.3 Application Example: NYPA CSC Installation at Marcy Substation ...............................9-1 9.4 Functional and Rating Expandability via Converter-Based Building Block Structure.....9-1

10 VOLTAGE-SOURCED CONVERTER-BASED BACK-TO-BACK (BTB) SYSTEM INTERTIES (OUTLINE) ...........................................................................................................10-1

10.1 Application Motivation (As Contrasted to Line-Commutated Approach) ....................10-1 10.2 Operating Principles and Characteristics ...................................................................10-1 10.3 Functional Characteristics..........................................................................................10-1 10.4 Control Principles and Structures...............................................................................10-1 10.5 Converter Power Circuit Topology and Output Control Options ................................10-2 10.6 Rating and Protection Considerations........................................................................10-2 10.7 Performance During and Following Line Faults .........................................................10-2 10.8 Installation and Insulation Coordination Considerations ............................................10-2

3

10.9 Installation and Performance Comparison to Conventional Line-Commutated BtB Power System Ties ............................................................................................................10-2 10.10 BtB Application Examples ........................................................................................10-2

11 VOLTAGE-SOURCED CONVERTER-BASED DC TRANSMISSION SYSTEMS (DRAFT)11-1 11.0 Introduction ................................................................................................................11-1 11.1 Differences in Features and Characteristics from Conventional HVDC.....................11-2 11.2 Application Areas .......................................................................................................11-6 11.3 Basic Operating Principles and Characteristics .......................................................11-11 11.4 Practical Implementation..........................................................................................11-16 11.5 Basic Operational Limits .........................................................................................11-37 11.6 Converter Rating and Protection Considerations.....................................................11-39 11.7 Practical Operating Characteristics..........................................................................11-42 11.8 Control Principles and Structures.............................................................................11-46 11.9 Installation Considerations .......................................................................................11-56 11.10 Modularity...............................................................................................................11-57 11.11 Reliability, Availability and Maintainability ..............................................................11-59 11.12 VSC Transmission Application Examples ..............................................................11-61

12 AC TO DC CIRCUIT CONVERSION (DRAFT) .................................................................12-1 12.1 Background................................................................................................................12-1 12.2 Characteristics of Tripole DC .....................................................................................12-2 12.3 Configuration Options for ac to dc Conversion ..........................................................12-9 12.4 Power Capability of ac Circuits Converted to dc ......................................................12-10 12.5 The Effect of System Context on Total Path Flow ...................................................12-15 12.6 Dynamic Support Capability.....................................................................................12-20 12.7 Limits of AC System Optimization............................................................................12-22 12.8 AC vs. DC Losses....................................................................................................12-23 12.9 Fire and Ice Issues...................................................................................................12-31 12.10 Logistics of Non-Disruptive Conversion from ac to dc ...........................................12-33 12.11 Potential Conversion of Cable Circuits...................................................................12-35

13 HARMONICS (OUTLINE)...................................................................................................13-1 13.1 Harmonic Sources......................................................................................................13-1 13.2 Possible Effects of Harmonics ...................................................................................13-1 13.3 Typical Utility Specified Harmonic Limits ...................................................................13-1 13.4 Harmonic Reduction Techniques for Power Electronics Equipment..........................13-1 13.5 Passive Filter Design Consideration ..........................................................................13-1 13.6 Active/Hybrid Filtering and Damping Possibilities ......................................................13-2

14 STUDIES FOR TRANSMISSION CONTROLLER SELECTION AND SPECIFICATION PREPARATION (OUTLINE)....................................................................................................14-1

14.1 Main Objectives of Studies.........................................................................................14-1 14.2 Technical Information Needed for Controller Specification ........................................14-1

14.3 Main System Studies ................................................................................................14-2

15 ARCHIVE REFERENCE FOR SELECTED INSTALLATIONS (OUTLINE) 15.1 Introduction ................................................................................................................15-1 15.2 Controller Groups.......................................................................................................15-1 15.3 General Description Structure....................................................................................15-1 15.4 Functional Structure and Operating Features of the Control System ........................15-2 15.5 Summary of Commissioning and other Performance Tests.......................................15-2 15.6 Results and Benefits of the Installation ......................................................................15-3

A CONTRIBUTORS ................................................................................................................. A-1

B AUTHORS GUIDE ................................................................................................................ B-1

1-1

1 ALTERNATING CURRENT (AC) TRANSMISSION SYSTEMS (DRAFT)

Introduction

This chapter provides a systematic review of the basic theory of AC transmission systems—with historical perspective, the fundamental relationships of AC power transmission, the role of power electronics and FACTS, the basic types of compensators and controllers, and issues related to the future of transmission control.

1.1 Historical Perspective

The electric power industry began to evolve in the 1880s. Almost from the very beginning two competitive systems started to emerge: direct current (dc) power generation and transmission strongly pursued by Thomas Edison, and alternating current (ac) power generation and transmission initiated in Europe and transformed into a practical scheme with Nikola Tesla’s inventions. This scheme, implemented in the U.S.A. by industrialist George Westinghouse, decisively won the early competition in 1896 when the famous Niagara hydro power generation project convincingly demonstrated viable “long distance” ac power transmission over a 20 mile, 11 kV “high voltage” line from Niagara Falls to the city of Buffalo, N.Y. The success of the prestigious Niagara project fueled the universal acceptance and rapid development of ac power systems. The key to this acceptance was the technical feasibility to step up the alternating generator voltage by highly efficient magnetic transformers for transmission in order to minimize conductor size and losses, and step it down for the consumer to meet domestic and industrial load requirements. The ac power systems with high voltage transmission capability allowed the use of remote power generation and, ultimately, also the intertying separate power systems into a large area power grid characterizing modern supply systems today. This approach is in contrast to Edison’s original concept of dc power system, which, due to transmission limitations, envisioned a large number of distributed and independent dc central (generation) stations, each supplying no more than a few square miles of distribution network for local loads.

The U.S. electric power industry was one of the 20th Century’s most phenomenal growth industries. From the very beginning, the unparalleled expansion of industrial activity, rapidly growing population, and available energy sources contributed to this growth. The number of utilities reached 3,620 by 1902, and more than 6000 by the early 1920s. Privately owned utilities dominated the production of electric power, but municipal organizations also participated in retail power distribution. Originally, the utilities typically were small and locally owned, with very limited geographical service areas. However, by the late 1920s, a few large utility holding companies controlled over 80% of the industry’s power generation capacity.

The largely unregulated (and often abused) holding company era was brought to an end by the Public Utility Holding Company Act of 1935, and the creation of the Federal Power Commission. The industry concentration was significantly decreased and regulation of interstate electric rates and other legislative measures were introduced. In addition, the economic

1-2

depression of the 1930s prompted the government to participate in the production of electric power for regional economic development. This led to the establishment of a number of public utility entities and cooperatives, most of which were primarily distribution companies. Still by 1950, the total power supplied by the private sector decreased to about 80%.

The unprecedented technological developments after the Second World War with rapid industrial growth resulted in dramatic increase in the demand of electric power and the industry capacity expanded nearly tenfold from the 1950s to the early 1970s. This huge increase of power demand was answered by major expansion of generation and transmission facilities, and by the formation of regional power pools and increasing interconnection of individual power systems.

The socioeconomic conditions had unexpectedly begun to change during the 1970s with the utility industry facing a set of difficult economic, environmental, and social problems. The oil embargo in the mid 1970s, public opposition to nuclear power, social focus on clean air and other environmental issues led to considerable increase in operating cost and governmental intervention. The national energy legislation, various environmental initiatives, and other restrictive regulations went into effect. Alternate energy development plans for solar, geothermal, oil shale and others were initiated. At the same time the U.S. manufacturing industry went through major restructuring: large, concentrated manufacturing facilities were closed down and production was distributed to smaller facilities at different geographic locations. This, combined with pronounced demographic changes (people moved from cold to warm climates), resulted in a considerable geographical shift in power demand.

All these would have required the relocation or construction of new generation facilities and transmission lines relatively quickly to match the geographically different power demand profile and accommodate a volatile fuel cost structure. Neither the strong economic base, nor the previous freedom of action existed for utilities to adopt these conventional solutions. Indeed, the increasing public concern about environment and health, the cost and regulatory difficulties in securing the necessary “rights-of-way” for new projects, have often prevented or excessively delayed the construction of many generation facilities and transmission lines needed by the utilities.

The problems imposed by the new socioeconomic conditions fueled the further growth of interconnection among neighboring utility systems to share power with other regional pools and be part of a growing national grid. The underlying reason for this integration has been to take advantage of the diversity of loads, changes in peak demand due to weather and time differences, the availability of different generation reserves in various geographic regions, shifts in fuel prices, regulatory changes and other factors which may manifest themselves differently in other time and geographic zones.

The U.S. power system, evolving from originally isolated utility suppliers to regional power pool groups, did not have a flexible enough transmission grid to cope with the rapidly changing requirements under rapid economical and environmental changes. In the interconnected system “contracted” power was to be delivered some times from a distant generation site, often by “wheeling” it through the transmission systems of several utilities, to the designated load area. These arrangements inevitably led to uncontracted and undesired parallel- and loop-flows of power (since part of the line current from the sending-end flowed through each available parallel path in proportion to its admittance), which often overloaded some lines causing thermal and voltage variation problems. The receiving-end was also exposed to difficulties caused by the contingency loss of imported power and the consequent heavy overload condition on the local

1-3

system, leading to severe voltage depression with the danger of possible voltage collapse. Also, with the growing interconnected system it was increasingly difficult to maintain the traditional (conservative) stability margins without sufficient transmission reinforcement.

The voltage support and transient stability requirements of the expanding interconnected network, and the prevailing restrictions for new line construction, as well as economic considerations, led to the increasing applications of the first generation of power electronics-based the Static Var Compensators, in the late 1970s and more advanced transmission controllers in the 1990s.

The Notice of Proposed Ruling (NOPR) by the Federal Energy Regulatory Commission (FERC) in 1995 for a more competitive electric industry presented a new challenge, with implied structural change, for the utility industry. The main objective of NOPR was to facilitate the development of a competitive market by ensuring that wholesale buyers and sellers can reach each other through non-discriminatory, open access transmission services. The implication of “open access” is that power generation and transmission must be functionally “unbundled.

It is evident that the unbundled power system structure considerably increases the demand on the transmission network. The main economic emphasis of deregulation is to reduce the cost of electricity, that is, to minimize the cost of electric power generation by free (non-discriminatory) competition. This, as an opportunity, increases the number of independent power producers. Also, it may keep moving the geographical locations of lowest cost power generation, and vary the local generation levels, according to the relative cost of different fuels and other changing factors affecting the cost of energy production (e.g., environmental protection). The compulsory accommodation of the contracted (usually the least expensive) power by the transmission network can aggravate the parallel- and loop-flow problems, causing unpredictable line loading (thermal limits), voltage variation and the potential decrease of transient stability. The unbundling of power generation from transmission may also worsen these problems by eliminating the incentive for equipping the generators with often-costly functional capabilities, such as the effective control of reactive power generation and absorption (high-response excitation systems, power system stabilizers, etc.) to aid power transmission. The consequent possible decrease of coordinated reactive power support provided traditionally by the generators can further aggravate the transmission problems in the areas of voltage variation, transient and dynamic stability. The potential effect of all these on the reliability and security of the overall power system, without effective counter measures, could result in economically damaging forced load shedding and blackouts. In the last few years, both the U.S.A. and several European countries have been exposed to these problems.

The traditional solution to the above problems would be a massive reinforcement of the transmission network with new lines to reestablish, by the method of “brute force,” the conventional voltage limit, stability, and thermal margins under greatly expanded contingency scenarios. Such a major undertaking probably would not be possible due to the prevailing environmental and regulatory constraints, and likely opposition by affected public, nor would it be economically savvy. Although transmission systems do need new lines to handle the increasing loads and flexibility of energy transportation routs, these can be combined with the increasing application of power electronics-based, and real time computer-controlled, compensators and controllers to provide cost effective, “high tech” solutions to the prevailing problems and growing flexibility requirements with economically desired utilization of transmission assets.

1-4

In the late 1980s, the Electric Power Research Institute initiated a major program for the development of various power electronics-based Controllers for the practical realization of the Flexible AC Transmission Systems (FACTS), in which these Controllers regulate transmission voltage and power flow and, through rapid control action, mitigate dynamic disturbances. As a result of this program the Thyristor-Controlled Series Capacitor (TCSC), and a new family of converter-based Controllers, the Static Synchronous Compensator (STATCOM), Static Synchronous Series Compensator (SSSC), the Unified Power Flow Controller (UPFC), the Interline Power Flow Controller (IPFC), and Back-to-Back Ties (BtB) (BtB) with four-quadrant operation were developed and successfully commissioned through the 1990s.

1.2 Fundamental Relationships of AC Power Transmission

It is recognized that the majority of readers are familiar with the physical laws characterizing ac transmission lines and the various mathematical techniques may be applied in their analysis. It is not the intention here to dwell into this theory, but to establish, with a systematic review and physical explanation, those limiting factors of ac power transmission which can be effectively influenced by power electronics based Controllers to increase the transmittable power, system security, and asset utilization.

The main constituents of an ac power system are: generators, transmission (sub-transmission), and distribution lines, and loads, with their related auxiliary support and protection equipment. The generators are rotating synchronous machines. The transmission, sub-transmission, and distribution lines are essentially distributed parameter, dominantly reactive networks designed to operate at high, medium, and low alternating voltages, respectively. The loads may be synchronous, non-synchronous and passive, consuming in general both active and reactive power.

The modern transmission system is a complex network of transmission lines interconnecting all the generator stations and all the major loading points in the power system. These lines carry large blocks of power which generally can be routed in any desired direction on the various links of the transmission system to achieve the desired economic and performance objectives. Separate ac systems are often synchronously intertied with ac transmission lines to form a power pool in which energy can be transported between the individual systems. In this arrangement, at a given time some system may be importing and others exporting power, while some systems just providing the service of “wheeling” power through their transmission network to facilitate particular transactions. The power pools with individual systems can further be tied together with synchronous ac, or asynchronous dc links to form a large power grid to provide highly reliable electric power supply at minimum cost. The main characteristic of today’s transmission system is an overall loop structure, as illustrated with a simple power system schematic in Figure 1-1, which provides a number of path combinations to achieve the functional versatility desired. This is in contrast to early day transmission (and present day sub-transmission and distribution systems), which were (are) mostly radial, supplying power from generator to defined loads.

1-5

3

neighboringTie-line to

System 2

3L9

L4

L3

L5

4Load

To bus 6

1 Bus coding

L2L1

2

6

To bus 10

L8Load

Circuit breaker

Generator

Transformer

L7L6

Bus

neighboringTie-line to

System 1

Load Load

Figure 1-1 Electric power system structure

1.2.1 Analytical Characterization

In spite of the generally complex nature of an actual power system, the basic analytical relationships of power transmission can be derived by a simple so-called two machine model, in which a sending-end generator is interconnected by a transmission line with a receiving-end generator (which is sometimes considered as an infinite power voltage-bus). For the sake of generality, the sending-end and receiving-end generators in the model may also represent two independent ac systems (each represented by a hypothetical resultant machine and impedance) which are inter-tied by a transmission link for power exchange.

1.2.1.1 Ideal (Lossless) Line

An ac transmission line is characterized by its per-mile distributed circuit parameters: the series resistance and inductance, and the shunt conductance and capacitance. The characteristic behavior of the line is primarily determined by the reactive circuit elements, the series inductance l and shunt capacitance c. With a customary lumped-element representation of the ac transmission line, the two machine transmission model is shown in Figure 1-2. (Bold-faced letters represent voltage and current phasors [rotating vectors].)

1-6

Is

sV

line capacitanceline inductance

I

Vr

r

Figure 1-2 Representation of a lossless transmission line.

The transmittable electric power of the system shown in Figure 1-2 is defined by the following equation1:

PV V

Zs r

=0 sin

sinθ

δ (1.1)

in which Vs is the magnitude of the sending-end (generator) voltage phasor, Vs Vr is the magnitude of the receiving-end (generator) voltage phasor, Vr δ is the phase angle between Vs and Vr (transmission or load angle), Z0 is the surge or characteristic impedance given by

0Z =

lc (1.2)

θ is the electrical length of the line expressed in radians by

θ

πλ

β= =2

a a (1.3)

where λ is the wavelength and β is the number of complete waves per unit line length, i.e.,

β

πλ

ω π= = =2

2lc f lc (1.4)

and a is the length of the line.

The lossless line considered exhibits an ideal power transmission characteristic at the surge impedance or natural loading, at which the transmitted power is:

1-7

0

20

0P =

VZ

(1.5)

where V0 (= Vs = Vr) is the nominal or rated voltage of the line.

Conclusions on power transmission by lossless line:

1. At natural (surge impedance) loading the amplitude of the voltage remains constant and the voltage and current stay in phase with each other (but rotated together in phase) along the transmission line. Consequently, the transmission power, P0, is independent of the length of the line. At surge impedance loading the reactive power exchange within the line is in perfect balance, and the line provides its own shunt compensation. That is, the reactive power demand of the series line reactance is precisely matched by the reactive power generation of the shunt line capacitance.

2. Economic considerations and system operation requirements rarely allow surge impedance loading. At lighter loads the transmission line is capacitively over compensated. The voltage increase across the series line reactance, due to the charging current of the shunt line capacitance, is greater than the voltage drop caused by the load current. As a result, the transmission line voltage increases along the line, reaching its maximum at the mid-point. Of course, this “surplus” charging current also flows through the sending-end and receiving end generators (or ac systems) forcing them to absorb the corresponding (capacitive) reactive power. At greater than surge impedance loading the transmission line is under compensated. That is, the voltage increase resulting from the shunt line capacitance is insufficient to cancel the voltage drop across the series line reactance due to the load current. Therefore, the voltage along the line decreases, reaching the minimum at the mid-point. In this case, the net reactive power demand of the line (inductive) must be supplied by the sending-end and receiving-end generators.

1.2.1.2 Lossless inductive line

Equation (1.1) provides a generalized expression characterizing the power transmission over a lossless, but otherwise accurately represented line. For the explanation of the major transmission issues, and for the introduction of relevant power electronics based Controller concepts, it is convenient to use an approximate form of Equation (1.1) characterizing electrically short transmission lines, for which sin θ ≅ θ = βa = ω a lc . Then Z0θ = ω a lc l c/ = ω al = ωL

= X, the series inductance of the line, and the transmitted power becomes:

PV V

Xs r

≅ sinδ or PVX

≅2

sinδ (1.6)

This simplified equation neglects the shunt capacitance of the line. The effect of shunt capacitance on the transmission for lines shorter than 100 miles is indeed negligibly small. Moreover, although the line capacitance, as explained above, can cause over-voltage problems for under loaded lines, but, since sinθ ≤ θ and thus Z0 sinθ ≤ X, it always tends to increase the transmittable power. Thus, the neglect of shunt line capacitance represents a worse than actual case from the standpoint of maximizing the (steady-state or transient) transmittable power, but does not falsify the main considerations governing power flow control (which also determine the

1-8

possible utilization of transmission assets and the application of power electronics technology). The voltage related problems of open circuited and underloaded lines are usually handled satisfactorily by permanently connected or switched shunt reactors in combination with the excitation control of generators.

δ/2

(c)

π/20

P

2

(b)

maxPP,Q

Vs

I

2

π δ

P = XV

δsin

2=Q X

2V

δ/2

1-cos( )δ

= j IV Xx

Vm Vr

Vs

(a)

X/2

Vm Vr

X/2I

Figure 1-3 Model of a two-machine power system with inductive line (a), corresponding phasor diagram (b) and power transmission vs. angle characteristic (c).

Equation (1.6) can also be derived from the elementary techniques of ac circuit analysis, using complex phasors, an approach which allows a very effective treatment and illustration of different compensation and control concepts employed in power electronics based Controllers. Considering again the simple two machine model, but assuming a purely inductive transmission line with zero shunt capacitance, as shown in Figure 1-3a with the corresponding phasor diagram in Figure 1-3b. If the sending- and receiving-end voltages are defined by

Vs

jVe V j= = +

δ δ δ/(cos sin )

2

2 2 (1.7)

and

Vr

jVe V j= = −

− δ δ δ/(cos sin )

2

2 2 (1.8)

then the midpoint voltage is

1-9

V

V Vs rm m

jV e V=+

= =2

0

2cos

δ

(1.9)

and the current through the line is given by

I

V V=

−=

s r

jXVX

22

sinδ

(1.10)

In the case of lossless line assumed, power is the same at both ends (and at the midpoint), i.e.,

P V I

VX

m= =2

sinδ (1.11)

which is, of course, identical to that given by equation (1.6). The reactive power provided for the line at each end is

Q Q VI sin

VXs r= − = = −

δδ

2

2

1( cos ) (1.12)

The relationships between real power P, reactive power Q, and angle δ are shown plotted in Figure 1-3c. As seen, at a constant voltage (Vs=Vr=V) and fixed transmission system (X=const) the transmitted power is exclusively controlled by angle δ. Note also that real power, P, cannot be controlled without also changing the reactive power demand on the sending- and receiving-ends.

A different illustration using voltage and current phasors to provide a physical visualization for the coupled variation of real power, reactive power, and of voltage along the transmission line, as a function of angle δ, is presented in Figure 1-4a and 1-4b. Figure 1-4a shows that the voltage phasor Vx across the series line reactance increases, together with line current phasor I, as angle δ is increased. At the same time, the (line to neutral) voltage along the line decreases from the ends towards the middle, reaching a minimum at the actual midpoint. It can also be observed that the relative angular position of the voltage along the line from the two ends is continuously changing (in the opposite direction) until at the midpoint it falls precisely in phase with the line current I (which is at a fixed 90º angle with respect to the voltage Vx). Thus, the product of the midpoint voltage and the line current, VmI, yields the transmitted power P. It is evident from this relationship that the progressive increase of angle δ will not progressively increase the power P. This is because for δ>90°, the midpoint voltage will decrease more rapidly than the line current increases and, consequently, their product will decrease from its maximal value and ultimately reach zero at δ=180°.

1-10

Mid-point

δδ22

VsmV

SV I

X/2 X/2

X X 2decreases with

=

=

m

Q

Vr V

P

s

VmV

δV sinI 2

δ:

V

V V

cosI δ2

= cosV δ2

increases with

=I

I

Vr

sinV 2Vx =

δ: δ

xV

I

P = ImV

= =r

xV

Figure 1-4a Variation of mid-point voltage and line current with transmission angle δ.

Figure 1-4b illustrates the variation of the in-phase and quadrature components of the line current with respect to the (receiving-) end voltage, as angle δ is varied from zero to 180° (i.e., δ/2 varied from zero to 90°). The quadrature component of the line current with respect to the sending-end, or the receiving-end, voltage progressively increases with δ until it becomes the total line current at δ=180°. By contrast, the in-phase component of the line current with respect to the end voltage increases with δ in the 0<δ<90º interval, and then it decreases to zero as δ reaches 180°.

1-11

15

Angle δ2

Line current:

V

V o

Locus of P Qmax

Pmax

60

V90o

oV75

Receiving-end voltage:V30o

V45o

o

0

2δV ( )

15I o

30I o

45I o

60I o

75I o90I o

X 2I sin2V=δ

Locus of Q =

=

Q

P

δV sinI 2

V cosI δ2

Figure 1-4b Variation of the transmitted real power P and the (receiving-end) reactive power Q with transmission angle δ.

Conclusions on Power Transmission by a Lossless Inductive Line

1. At fixed end voltages and line impedance, the transmission angle controls the power flow in the line by changing the voltage across the line (measured from the sending- to the receiving-end).

2. The voltage along the transmission line decreases from the ends towards the midpoint of the line with increasing transmission angle. The midpoint voltage is maximum at zero degree, and zero at 180 degrees.

3. The line current increases with increasing transmission angle. The product of the midpoint voltage and the line current determines the transmitted active power at any given transmission angle. The active power is zero at 0 and 180 degrees; maximum at 90 degrees.

4. The variation of the transmission angle changes not only the line current, but also the angles between the end voltages and the line current. Thus, with increasing transmission angle the reactive power flow in the line starts to increase rapidly: it is zero at zero transmission angle, reaches the value of maximum transmittable active power at 90 degrees, and then increases to the double of that at 180 degrees.

5. The adjustment of the transmitted active power by angle control is always associated with a mathematically defined reactive power flow change in the line.

6. The reactive power in the line is supplied by the generators.

7. The transmittable electric power at a given system voltage is a function of the electrical length of the line, i.e., the value of the effective series line reactance X. Once the theoretical limit of

1-12

steady-state power transmission is reached at 90 degrees, the transmitted power would decrease with increasing line length (increasing X) unless either the line voltage is increased or the effective line impedance is decreased.

1.2.2 Transmission Via Multiple Paths in Intertied Power Systems

In a modern power system a load area usually receives power through two or more paths, some of which may be corridors of several lines. Although in this case the power in each line is still characterized by the previously derived analytical formulae with the associated observations, nevertheless some other factors important from the standpoint of system imposed transmission limits and line utilization, should also be considered. Indeed, these considerations will further expand the area of the applications for power electronics based Controllers to achieve optimized system operation with effective utilization of system assets.

A simple case of power flow through three paths is illustrated in Figure 1-5a. In a general case, the transmission lines representing the three paths could differ both in length and character (transmission voltage and unit-length impedance); consequently, the total impedance (inductive and resistive) determining the power transmission (Figure 1-5b) over the three independent paths between the two buses at a given transmission angle would be different. If the impedance difference is not correlated with the power transmission capacity of the line, then, evidently, one line could become overloaded, and another one underloaded. For example, one path could be a long, high impedance line with high transmission capacity and another one a short, low impedance line with much lower transmission capacity. Thus, although the three paths may represent the necessary transmission capacity between the two buses, the actual transmittable power would be limited by the low impedance line.

j

2

Vs

I

X

2jX

1

2

R

R

1Vr1I

3I 3jX 3R

Vs P1 Qr1VrQs1

P2 Qr2Qs2

P3 Qr3Qs3

System"s"

System“r"

System"s"

System“r"

(a)

(b)

1-13

Figure 1-5 Multi-path power transmission (a), and corresponding circuit elements determining power flow distribution (b).

In addition to the main potential problem of improper (inductive) impedance ratio between the three transmission paths, disproportionate load sharing could also be caused by the different values of the reactive to real impedance ratios (X/R) in the three lines. That is, with different X/R ratios the currents flowing in the lines have different angles with respect to the bus voltages, resulting in circulating currents through the loops formed by the lines.

In an actual power system, many buses are interconnected by transmission lines of different operating voltage, impedance and length to provide an arrangement in which power can freely flow from areas of surplus generation capacity to areas of power deficit. Such an arrangement practically always results in a number of major and many minor power loop flows, due to the differences in reactive and real impedances of the lines providing the various interconnecting paths. Uncontrolled loop power flow may become a limiting factor between certain points of the overall power system; it can hinder the proper utilization of system assets, and increase transmission losses.

Conclusions for Multiple-Paths Transmission

1. Power will flow according to the (normalized) reactive impedance of the paths, and not according to the power transmission capacity of the lines.

2. The “natural” utilization of the transmission lines is generally uneven; some may be fully loaded and some others grossly underloaded.

3. Internal circulating loop power flow may be caused by different X/R ratios of the lines forming the loop.

4. In a meshed power system with many interconnected buses the power flow between certain areas may take place over a number of major and many minor power loops, some of which could stretch out over long distances.

5. Loop power flows may limit transmittable power, hinder asset utilization, and increase transmission losses.

1.2.3 Inherent Steady-State Limits of AC Power Transmission

The relationship between the transmitted power and the transmission parameters given by Equation (1.11) indicates that the maximum power, Pmax=V2/X, transmittable over a lossless line at a given transmission voltage, is totally determined by the line reactance X and thus sets the theoretical limit for steady-state power transmission. A practical limit for an actual line with resistance R may be imposed by the I2R loss that heats the conductor. At a certain temperature the physical characteristics of the conductor would irreversibly change (e.g., it could get deformed with permanent sag). This sets a thermal limit for the maximal transmittable power. Generally, for long lines X, and for short lines R would provide the main transmission limitation.

AC loads are generally sensitive to the magnitude, and may be sensitive to the frequency of the applied alternating voltage. AC power systems are generally operated at a substantially constant (typically 50 or 60 Hz) frequency. The voltage levels in ac systems may moderately vary, but are

1-14

not allowed to exceed well defined limits (typically +5 and -10%). This tight voltage tolerance may impose the primary transmission limitation for long radial lines (no generation at the receiving end) and for the so-called tapped-lines, which feed a number of (relatively small) loads along the transmission line.

In a system with multiple transmission paths the steady-state transmission limits may be reached or exceeded as a result of parallel and loop power flows. These flows, as previously discussed, often occur in a multi-line, interconnected power system, as a consequence of basic circuit laws which define current flows by the impedance rather than the current capacity of the lines. They can result in overloaded lines with thermal and voltage level problems.

Conclusions for Steady-State Transmission Limits

1. The theoretically maximum transmittable power of a “natural” (uncompensated) line at a given transmission voltage is determined by the reactive line impedance (reactance), ultimately the length of the economically acceptable transmission line.

2. The line resistance determines the thermal limit for line. This limit is absolute, determined by line design (primarily conductor size). Whereas the effective line reactance can be changed by appropriate compensation to increase the limit for the transmittable power, or the distance of transmission, similar compensation could not change the actual line losses although it could decrease the resistive voltage drop over the total transmission path.

3. Allowed bus voltage variation of +5 and -10 per cent, also imposes a limit on transmittable power, particularly for radial lines.

1.2.4 Extension of Power Transmission Limits by Line Compensation and Power Flow Control

It follows from the reactive character of ac transmission lines (and the corresponding analytical expressions) that the steady-state transmittable power can be increased and the voltage profile along the line controlled by appropriate reactive compensation. The purpose of this reactive compensation is to change the natural electrical characteristics of the transmission line to make it more compatible with the prevailing load demand. Thus, shunt-connected, fixed or mechanically switched reactors can be applied to minimize line overvoltage under light load conditions, and shunt connected, fixed or mechanically switched capacitors can be used to maintain voltage levels under heavy load conditions. In the case of long transmission lines, series capacitive compensation is often employed to establish a virtual short line by reducing the inductive line impedance and thereby the electrical length of the line. In some multi-line system configurations, it can happen that the transmission angle imposed “naturally” on a particular line is inappropriate for the power transfer planned for that line. In this case, a phase angle regulator (or phase shifter) may be employed to control the angle of this line independent of the prevailing overall transmission angle.

In the following sub-sections, basic concepts for increasing the transmittable power by ideal shunt-connected reactive compensation, series compensation, and phase angle regulation, as well as for controlling active and reactive power flow in the line, will be reviewed. These basic approaches will provide the foundation for power electronics-based compensation and control

1-15

techniques capable not only to increase steady-state power flow, system imposed transmission limits but also to improve the stability and overall dynamic behavior of the system.

1.2.4.1 Shunt Compensation

Figure 1-6a shows the simple two machine (two-bus) transmission model in which an ideal var compensator is shunt-connected at the midpoint of the transmission line. This compensator is represented by a sinusoidal ac voltage source (of the fundamental frequency), in-phase with the midpoint voltage, Vm, with an amplitude identical to that of the sending- and receiving-end voltages (Vm = Vs = Vr = V) The midpoint compensator in effect segments the transmission line into two independent parts: the first segment, with an impedance of X/2, carries power from the sending end to the midpoint, and the second segment, also with an impedance of X/2, carries power from the midpoint to the receiving end. Note that the mid-point var compensator exchanges only reactive power with the transmission line in this process. The relationship between voltages, Vs, Vr,, Vm, (together with Vsm, Vrm,), and line segment currents Ism and Imr is shown by the phasor diagram in Figure 1-6b.

mr

I

P

P

P

(c)

2

4

(b)

π/2

max

max

π δ

sm

P,Q

max

δ/2 δ/2

(a)IdealComp.(P=0)

sVV m

smV

V

X/2

sV

smI mrI

Vm

V== V=V Vrs m

)

2V2

sin=PX

=Pp

V

X2

δ

sinδ/2

2=Qp X

4V δ/2(1-cos

rV

X/2

rV

Imr

Imr

22

0

Xj 2Ism

Xj 2

Figure 1-6 Two machine power system with an ideal mid-point reactive shunt compensator (a), corresponding phasor diagram (b), and transmitted power vs. transmission angle characteristic (c).

For the lossless system assumed, the real power is the same at each terminal (sending-end, mid-point, and receiving-end) of the line, and it can be derived readily from the phasor diagram of

1-16

Figure 1-6b using a similar computational process demonstrated in the revious section [see equations (1.7) through (1.11)].

With

V V V I I IVXsm mr sm mr= = = = =cos ; sin

δ δ4

44 (1.13)

the transmitted power is

P V I V I V I VIsm sm mr mr m sm= = = =cos cosδ δ4 4 (1.14)

or

PVX

= 22

2

sinδ

(1.15)

Similarly

Q VI sinVX

= = −δ δ4

21

2

2

( cos ) (1.16)

The relationship between real power P, reactive power Q, and angle δ for the case of ideal shunt compensation is shown plotted in Figure 1-6c. It can be observed that the mid-point shunt compensation can significantly increase the transmittable power (doubling its maximum value) at the expense of a rapidly increasing reactive power demand on the mid-point compensator (and also on the end-generators).

The concept of transmission line segmentation can be expanded to the use of multiple compensators, located at equal segments of the transmission line, as illustrated for four line segments in Figure 1-7. Theoretically, the transmittable power would double with each doubling of the segments for the same overall line length. Furthermore, with the increase of the number of segments, the voltage variation along the line would rapidly decrease, approaching the ideal case of constant voltage profile. Ultimately, with a sufficiently large number of line segments, an ideal distributed compensation system could theoretically be established, which would have the characteristics of conventional surge impedance loading, but would have no power transmission limitations, and would maintain a flat voltage profile at any load.

1-17

X/4

lmI

Isl

sV

Vs Vl

X/4 slI

Ilm

nr

I

I

mn

VnmVlV

rV

nr

mV nV

X/4lmI IX/4mnI

rV

Xj4Isl

Xj4

ImnXj4 Inr

Xj4

Figure 1-7 Two machine system with ideal reactive shunt compensators providing multiple line segmentation, and associated phasor diagram.

It is to be appreciated that such a distributed compensation hinges on the instantaneous response and unlimited var generation and absorption capability of the shunt compensators employed, which would have to stay in synchronism with the prevailing phase of the segment voltages and maintain the predefined amplitude of the transmission voltage, independently of load variation. To visualize the operation and control coordination complexity of a generalized compensation scheme exhibiting ideal transmission characteristics, consider the lossless but otherwise correctly represented system of Figure 1-2. Assume that the line is provided with a sufficient number of shunt connected ideal var compensators. At no load (zero transmission), the voltages of all compensators would be in phase and they would be absorbing the capacitive vars generated by the distributed line capacitance. With increasing load (increasing δ), the relative phase angle between the voltages of adjacent compensators would increase, but their var absorption would continuously decrease up to the natural (surge impedance) loading, where it would become zero. With further increasing load, beyond the surge impedance loading, the compensators would have to generate increasing amount of capacitive vars to maintain the flat voltage profile. However, at sufficiently heavy loads, the relative phase angle between two adjacent compensators could become too large, resulting in a large voltage sag, at which the power transmission could not be maintained, regardless of the var generation capacity of the compensators, unless additional compensators would be employed to increase further the segmentation of the line.

From the above discussion it is evident that the controlled shunt compensation scheme approximating an ideal line, whose surge impedance is continuously variable so as to maintain a flat voltage profile over a load range stretching from zero to several times the actual surge impedance characterizing that line, would be too complex, and far too expensive, to be practical, particularly if stability and reliability requirements under appropriate contingency conditions are

1-18

also considered. However, it is interesting to note historically that the practicability of limited line segmentation, using 12, 300 Mvar thyristor-controlled static var compensators, has been demonstrated by the major, 600 mile long, 735 kV transmission line of the Hydro-Quebec power system built to transmit up to 1200 MW power from the James Bay hydro-complex to the City of Montreal and to neighboring US utilities. More importantly, the transmission benefits of voltage support by controlled shunt compensation at strategic locations of the transmission system have been demonstrated by numerous installations in the world.

Of course, there are many applications in which mechanically-switched capacitors are applied to control transmission line voltage where there are slow, daily and seasonal load variations.

Although, reactive shunt compensation is discussed above in relation to a two-machine system, this treatment can easily extended to the more special case of radial transmission. Indeed, if a passive load, consuming power P at voltage V, is connected to the midpoint in place of the receiving-end part of the system (which comprises the receiving-end generator and transmission link X/2), the sending-end generator with the X/2 impedance and load would represent a simple radial system. Clearly, without compensation the voltage at the mid-point (which is now the receiving-end) would vary with the load (and load power factor). It is also evident that with controlled reactive compensation the voltage could be kept constant independent of the load. Shunt compensation in practical applications is often used to regulate the voltage at a given bus against load variations, or to provide voltage support for the load when, due to generation or line outages, the capacity of the sending-end system becomes impaired.

Conclusions on shunt compensation

1. Reactive shunt compensation can maintain constant midpoint voltage at increasing (varying) transmission angle, thereby in effect segmenting the transmission line into two independent lines of half length and thus doubling the theoretically transmittable power.

2. With multiple shunt compensation (multiple segmentation) the theoretical, steady-state transmittable power can be multiplied.

3. Shunt compensators are also effective in maintaining (regulating) the voltage at the end of a radial line.

4. Under steady-state conditions, shunt compensators used for voltage regulation or control do not exchange active power with the ac system (neglecting the possible internal losses of practical compensators).

1.2.4.2 Series Compensation

The basic idea behind series capacitive compensation is to decrease the overall effective series transmission impedance from the sending-end to the receiving-end. The conventional view is that the impedance of the series connected compensating capacitor cancels a portion of the actual line reactance and thereby the effective transmission impedance is reduced as if the line was physically shortened. An equally valid physical view, helpful to the understanding of power flow controllers, is that, in order to increase the current (and thereby the transmitted power) through the series impedance of a physical line (which clearly cannot be changed), the voltage across this

1-19

impedance needs to be increased. This can be accomplished by a series-connected (passive or active) circuit element that produces a voltage opposite to the prevailing voltage across the series line reactance. The simplest such element is a capacitor, but, as will be seen later, controlled voltage sources can also accomplish this function in a much more generalized manner that can ultimately facilitate full control of real and reactive power flow in the line.

The familiar two-machine model with a series capacitor compensated line, composed of two identical segments for the clarity of illustration, shown at Figure 1-8a. The corresponding voltage and current phasors are shown at Figure 1-8b. Note that the magnitude of the total voltage across the series line reactance, Vx = 2Vx/2, increased by the magnitude of the opposite voltage, VC, developed across the series capacitor.

x

Vs

(b)

I

Vr

sV

(a)

V

Vm Vr

I

2P=P sinδV

π/2

maxP

maxs

π δ

k = 0.2k = 0

X (1-k)

(1-cosk = 0.4

C

QP,

sQ(1-k)

=2V2

Xk

2

k CX=

2

X

)δ

2XC

2X

2XC

2X

Vm

I-j 2XC

0

I-j 2XC

(c)

Figure 1-8 Two machine power system with series capacitive compensation (a), corresponding phasor diagram (b), and transmitted power vs. transmission angle characteristic (c).

The effective transmission impedance Xeff with the series capacitive compensation is given by

Xeff = X - XC (1.17)

or

Xeff = (1 - k) X (1.18)

where k is the degree of series compensation, i.e.,

k = XC/X 0 ≤ k <1 (1.19)

The current in the compensated line and the real power transmitted, per Equations (1.10) and (1.11) are:

1-20

I =−2

1 2Vk X( )

sinδ

(1.20)

P V IV

k Xm= =

−

2

1( )sinδ (1.21)

The reactive power supplied by the series capacitor can be expressed by

( )

Q IVX

kkC = =

−−2

2

2

21

1X C ( cos )δ (1.22)

The relationship between the real power P, series capacitor reactive power QC, and angle δ is shown plotted at various values of the degree of series compensation k in Figure 1.8c. It can be observed that, as expected, the transmittable power rapidly increases with the degree of series compensation k. Similarly, the reactive power supplied by the series capacitor also increases sharply with k and varies with angle δ a similar manner as the line reactive power.

Series capacitors have been used extensively in the last 50 years throughout the world for the compensation of long transmission lines.

It should be noted that series reactive compensation can be an effective method for balancing power flow (and thereby avoiding line overloading), when power transmission involves multiple paths with lines of different length and character, by adjusting the impedance of the individual lines according to their loading capacity.

Conclusions on Series Compensation

1. Reactive series compensation reduces the effective reactive impedance and thereby the electric length of the line.

2. Series compensation provides the most effective method for economical long distance ac power transmission.

3. Series compensation is an effective technique to balance power flow in multi-path power transmission.

4. Under steady-state conditions, reactive series compensators do not exchange active power with the ac system (neglecting the possible internal losses of practical compensators).

1.2.4.3 Transmission Angle Control

In practical power systems it occasionally happens that the transmission angle required for the optimal use of a particular line would be incompatible with the proper operation of the overall transmission system. Such cases would occur, for example, when power between two buses is transmitted over parallel lines of different electrical length or when two buses are intertied whose

1-21

prevailing angle difference is insufficient to establish the desired power flow. In these cases a phase shifter or phase angle regulator is frequently applied.

The basic concept is explained again in connection with the two machine model in which a phase shifter is inserted between the sending-end generator (bus) and the transmission line, as illustrated in Figure 1-9a. The phase shifter can be considered as a sinusoidal (fundamental frequency) ac voltage source with controllable amplitude and phase angle. In other words, the effective sending-end voltage Vseff becomes the sum of the original sending-end bus voltage Vs and the voltage Vσ provided by the phase shifter, as the phasor diagram shown in Figure 1-9b illustrates. The basic idea behind the phase shifter is to keep the transmitted power at the desired level, independent of the prevailing transmission angle δ, in a predetermined operating range. Thus, for example, the power can theoretically be kept at its peak value after angle δ exceeds π/2 (the peak power angle) by controlling the amplitude of quadrature voltage Vσ so that the effective phase angle (δ-σ) between the sending- and receiving-end voltages stays at π/2. In this way, the actual transmitted power may be increased significantly, even though the phase-shifter per se does not increase the steady-state power transmission limit.

shifter

V

Phase-

0

P

maxP

-Vσ

Vs

Vσ

δδ+σ

(

π/2

+σ−σ π+σπ δ

+σ−σ δ−σ

2Pa = X

V sin ) δ−σ

V

Vx σ= 0( )

(seffseff

s

V )+σ

−σ(xV )

r

X

seffV

I

Vr

(seffseffV )−σ+Vσ

+σ(xV )

(a)

(b)

(c)

Figure 1-9 Two machine power system with a phase shifter (a), corresponding phasor diagram (b), and transmitted power vs. transmission angle characteristic (c).

With the above phase angle control arrangement the effective phase angle between the sending- and receiving-end voltages becomes (δ-σ), and the transmitted power P can therefore be expressed as:

1-22

( )PVX

= −2

sin δ σ (1.23)

The relationship between real power P and angles δ and σ is shown plotted in Figure 1-9c. It can be observed that, although the phase-shifter does not increase the transmittable power of the uncompensated line, theoretically it makes it possible to keep the power at its maximum value at any angle δ in the range π/2 < δ < π/2+σ by; in effect, shifting the P versus δ curve to the right. It should be noted that the P versus δ curve can also be shifted to the left by inserting the voltage of the phase shifter with an opposite polarity. In this way, the power transfer can be increased and the maximum power reached at a generator angle less than π/2 (that is, at δ = π/2-σ).

In contrast to the previously-investigated shunt and series compensation schemes, the phase shifter generally has to handle both active and reactive powers. The VA throughput of the phase shifter (viewed as a voltage source) is

VA = Vs - VseffI =V σI = VσI (1.24)

Assuming that the magnitude of the original sending-end voltage, Vs, and that of the effective sending-end voltage, Vseff, are equal, and using the expression given for the line current in Equation (1.10), VA can be written in the following form:

VAVX

=−42 2

2

sin sinδ σ σ

(1.25)

In Equation (1.25), the multiplier sin[(δ-σ)/2] defines the current, at a given system voltage and line impedance, flowing through the phase shifter, and the multiplier sin(σ/2) determines the magnitude of the voltage injected by the phase shifter.

Phase shifters employing a shunt connected excitation transformer with a mechanical tap-changer and a series connected insertion transformer to provide adjustable series voltage injection for phase angle control are often employed in transmission system to control steady-state power flow and prevent undesired parallel and loop power flows.

Conclusions on Transmission Angle Control

1. Transmission angle control can adjust the prevailing power flow, but cannot increase the maximum transmittable power characterizing the controlled line.

2. Transmission angle control changes both the active and reactive line power in a mathematically defined manner.

3. In the case of conventional angle regulators (tap-changing transformers), both the active and reactive power exchanged in the process of angle control is supplied by the ac system.

1.2.4.4 Independent Control of Active and Reactive Power Flow

Independent control of active and reactive power flow in the line is attainable by the generalized power flow controller, which can inject in series with the line the compensating ac voltage with a phase angle fully controllable from zero to 360 degrees with respect to a chosen reference (bus

1-23

voltage or line current). The two-machine system with the generalized power flow controller is shown in Figure 1-10a and the corresponding phasor diagram in Figure 1-10b. Similarly to the previously discussed phase shifter, the power flow controller adds voltage phasor Vpq, with controllable magnitude, Vpq, and phase angle, ρ, to the original sending-end bus voltage phasor Vs and thereby producing the effective sending-end voltage phasor Vseff for the line. Since in addition to the adjustable magnitude Vpq, angle ρ is freely variable from 0° to 360°, the voltage across the transmission line and, with that, the line current, is controllable both in magnitude and phase angle with respect to the receiving-end bus voltage (or other selected voltage reference). The capability of adjusting both the magnitude and phase angle of the line current enables the generalized power flow controller to control independently the active and reactive power in the line.

V

2Ppq =X

V sin ) δ

VVx

seffseffs

V

xV

r

Controller

PowerVs

V X

seffV

I

VrFlow

pq

= V s+V pq

δ

oδ

ρ

o

-1.5

Controllable region1.0

P ( )

-0.5

min-1.0

0

δ (V =0)pq

=30oδ

o=0

rQ

δ0.25

0.5

maxP ( )

=90oδ

δ

1.5

Pδ

-0.5

0

1.0

0.5

1.5

π/2 π δ

(V =0)pq

max δP ( )

P ( ) min δ

Vpq ( Vsin δ+ X

ρ2 +Ppq

(b) (c) (d

(a)

Figure 1-10 Two machine power system with a generalized power flow controller (a), corresponding phasor diagram (b), control region of attainable active and reactive line power (c), and transmitted power vs. transmission angle characteristic (d).

The active and reactive control capability of the generalized power flow can be conveniently illustrated in the Q,P plane where the normalized reactive power Q (p.u.) is plotted against the normalized active power P (p.u.) of the simple two-machine model shown in Figure 1-10a. For the uncompensated system, with V2/X=1 (p.u.) stipulation, the function Q=f(P) can be readily written from Equations (1.11) and (1.12):

Q P= − − −1 1 2 (1.26)

or

( )Q P+ + =1 12 2 (1.27)

1-24

Equation (1.27) shows that the Q=f(P) function for the uncompensated system describes a circular arc drawn with a radius of 1.0 around the center defined by coordinates Po=0, Qo=-1.0, shown by dashed-line in Figure 1-10c. Each point defined by its P and Q coordinates on this arc corresponds to a particular transmission angleδ; e.g., coordinates P=0, Q=0 corresponds to δ=0; P=1.0, Q=-1.0 corresponds to δ=90°, etc., which identifies the starting control point of the power flow controller. Assuming, as illustrated, that the prevailing transmission angle, δ, is 30°, then, the full (360°) rotation of the compensating voltage phasor Vpq, with its maximum magnitude Vpq(=0.5 p.u.), will describe a circle in the Q,P plane with a radius of 0.5 around its center defined by coordinates P30° =0.5 and Q30° =-0.134, which characterize the uncompensated system at δ=30°. The area within this circle defines all P and Q values obtainable by controlling the magnitude and angle of compensating voltage phasor Vpq. In other words, this circle in the Q,P plane defines all P and Q transmission values attainable with the given rating of the power flow controller for the power system considered. With changing transmission angle, the control area with the same circular boundary would simply move up or down along the arc characterizing the uncompensated system. The transmittable power against the transmission angle δ between the boundaries determined by the magnitude Vpq of the injected voltage is shown in Figure 1-10d.

Similarly to the phase shifter, and in contrast to the series reactive compensator, the generalized power flow controller has to handle both active and reactive powers. The VA throughput of the phase shifter (viewed as a voltage source) is

VA = Vs - VseffI =V pqI = VpqI (1.28)

The conventional, phase shifter type implementation of the generalized power flow controller (e.g., using a mechanical tap-changer with in-phase and quadrature voltage steps) would necessitate the power system (sending-end bus) to supply both the active and reactive power, i.e., the total VA, exchanged in the course of the desired control action. However, as will be seen later, the converter-based implementation would internally generate all the reactive power exchanged, and thus the system would supply only the active power exchanged (which actually transmitted to the receiving-end bus).

The basic concept of independent active and reactive power flow control, that is, in effect the insertion of a voltage source with controllable magnitude and phase angle in series with the line, can also be applied for the minimization of active and reactive loop power flows in multi-path and meshed systems.

As illustrated in Figure 1-11a, if the angle of the inserted voltage is in quadrature with the current in a substantially inductive line, then the magnitude of the inserted voltage will control only the magnitude of the line current, and thus primarily the active power flow in the line. This means essentially that series reactive compensation primarily changes active power flow.

1-25

Reactive power compensation changes the magnitudeof the line current and thus primarily controls real power

Real power compensation changes the angle of theline current and thus primarily controls reactive power

VVs

)ζV (p+-Vp=

Vp

-RpV-

+RIpV+_I

I

pP

I

XL

LVpQ

V

(-P)Ir

V-Vs

p

V+ p

IoI

/2/2δ δ

(+P)

Lo

VL(-P)

L(+P)

V

rV

)q qV ζ= V+j (- I_I (-90 )

sV

qV

Lmax

(+90 ) qV

VLoo

V

max

min

δ δ /2/2

I

IoVr

LminVo

I

(+90 )qV (-90 )o VI

qo VVs

VqqP

I

XL

LVqQ

r(a)

(b)

Figure 1-11 Illustration for effect of line compensation: (a) reactive compensation primarily changes active power, and (b) active power compensation primarily changes reactive power flow.

Similarly, Figure 1-11b illustrates that if the inserted series voltage is in phase with the line current, then it will primarily control the angle of this current, and thereby the reactive power flow in the line. This means that series active power compensation primarily changes reactive compensation in the line.

From the above it follows that active loop power flow can be minimized by quadrature voltage injection providing essentially series reactive compensation; reactive loop flow can be minimized by in-phase voltage injection providing essentially series real power compensation.

Conclusions on Independent Active and Reactive Power Flow Control

1. The active and reactive power flow in the line at a given transmission angle can be controlled independently by a compensating voltage of adjustable magnitude and fully controllable phase angle (0°-360°) injected in series with the line. The phase angle of the injected voltage of appropriate magnitude is adjusted so as to force the line current to assume the angular position with respect to the selected (receiving-end) voltage that results in the desired active and reactive power flow.

2. The injected voltage in series with the line generally results in both active and reactive power exchange between the power flow controller and the ac system. With possible conventional implementation (in-phase and quadrature tap-changing transformers) both the active and reactive power exchanged would be supplied by the ac system; with converter-based implementation, only the active power would be supplied by the ac system, the reactive power would be internally generated.

1-26

3. Injected series voltage with controllable angle can be used for the minimization of either, or both, the active and reactive loop power flows in the line by providing controllable reactive and/or active power compensation.

1.2.5 Dynamic Limits of AC Power Systems

AC power systems employ rotating synchronous machines for electric power generation. (They may also employ rotating synchronous compensators (condensers) for reactive power compensation.) It is a fundamental requirement of useful power exchange that all synchronous machines in the system operate in synchronism with each other maintaining a common system frequency. However, power systems are exposed to various dynamic disturbances (such as line faults, equipment failures, various switching operations), which may cause a sudden change in the real power balance of the system and consequent acceleration and deceleration of certain machines. The ability of the system to recover from disturbances and regain the steady-state synchronism under stipulated contingency conditions becomes a major design and operating criterion for transmission capacity. This ability is usually characterized by the transient and dynamic stability of the system. A transmission system is said to be transiently stable if it can recover normal operation following a specified major disturbance. Similarly, the system is said to be dynamically stable if it recovers normal operation following a minor disturbance. The dynamic stability indicates the damping characterizes the system. A dynamic (or “oscillatory”) instability means that a minor disturbance may lead to increasing power oscillation and the eventual loss of synchronism.

During and after major disturbances the transmission angle and transmitted power may significantly change from, and oscillate around their steady-state values. Consequently, a power system cannot be operated at, or even too close to its steady-state power transmission limit. An adequate margin is needed to accommodate the dynamic power “swings” while the disturbed machines regain their synchronism in the system.

The phenomenon of escalating decrease and an eventual collapse of the terminal voltage as a result of an incremental load increase is referred to as voltage instability. Voltage collapse is the result of a complex interaction between induction motor type loads and certain voltage regulators, such as tap-changing transformers, which may take several seconds to minutes. The essence of this process is that decreasing terminal voltage results in increasing load current and poorer load power factor (induction motors) which tend to further decrease the terminal voltage. The voltage regulators (tap-changing transformers) are not able to change the character of this process and under sufficiently severe conditions (low system voltage and heavy loading) it degenerates (in a positive-feedback manner) into a voltage collapse. The voltage stability limit identifies for a given system the specific V and P condition at which the next increment of load causes a voltage collapse.

Summary on Dynamic Transmission Limits

1. The main synchronous machines of the power system must operate in synchronism to maintain a substantially constant, common system frequency.

2. Transient stability means that the system can recover normal operation following a major disturbance.

1-27

3. Dynamic stability means that the system can recover normal operation following a minor disturbance. Dynamic stability in effect means that the system has sufficient damping.

4. Voltage instability means an escalating process which results in the progressive decrease and the eventual collapse of the terminal bus-voltage after an incremental load increase. Voltage stability limit specifies the voltage and load limits for the terminal bus beyond which the process of voltage collapse would begin.

1.2.6 Dynamic Compensation for Stability Enhancement

As shown in Section 1.2.4, shunt and series line compensation, as well as transmission angle and power flow control can significantly change the prevailing transmitted power. Thus, it is reasonable to expect that, with suitable and fast controls, these techniques would become able to vary the power flow in the system so as to counteract disturbances and thereby increase the transient stability limit and provide effective power oscillation damping. It is also predictable that appropriate shunt and series compensation would be effective for increasing the voltage stability limit.

In the following two sub-sections, the potential effectiveness of shunt and series compensation, angle and power flow control for transient stability improvement and power oscillation damping are explored and compared. In the subsequent third sub-section, the use of shunt and series capacitive compensation for the increase of voltage instability limit for a radial transmission line is discussed.

1.2.6.1 Transient Stability Improvement

The potential effectiveness of shunt and series compensation, angle and power flow control on transient stability improvement can be conveniently evaluated by the equal area criterion. The meaning of the equal area criterion is explained with the aid of the simple two machine (the receiving-end is an infinite bus), two line system shown in Figure 1-12a and the corresponding P versus δ curves shown in Figure 1-12b. Assume that the complete system is characterized by the P versus δ curve “a” and is operating at angle δ1 to transmit power P1 when a fault occurs at line segment “1”. During the fault the system is characterized by P versus δ curve “b” and thus, over this period, the transmitted electric power decreases significantly while mechanical input power to the sending-end generator remains substantially constant. As a result, the generator accelerates and the transmission angle increases from δ1 to δ2 at which the protective breakers disconnect the faulted line segment “1” and the sending-end generator absorbs accelerating energy, represented by area “A1”. After fault clearing, without line segment “1” the degraded system is characterized by P versus δ curve “c”. At angle δ2 on curve “c” the transmitted power exceeds the mechanical input power and the sending end generator starts to decelerate; however, angle δ further increases due to the kinetic energy stored in the machine. The maximum angle reached at δ3, where decelerating energy, represented by area “A2”, becomes equal to the accelerating energy represented by area “A1”. The limit of transient stability is reached at δ3 = δcrit, beyond which the decelerating energy would not balance the accelerating energy and synchronism between the sending-end and receiving-end could not be restored. The area “Amargin”, between δ3 and δcrit, represent the transient stability margin of the system.

1-28

δ210

P1

δ δ π/2

sV

1 2

Pmax

XT 3

LX

4

Fault

LX mV

c (post-fault)

3 crit

b (during-fault)

δδ π

Amargin

a (pre-fault)

XXL

T Vr

XL

P

A1

A2

(a)

(b)

Figure 1-12 Illustration of the equal area criterion for transient stability: (a) power system model, and (b) transmitted power vs. transmission angle characteristic under fault condition.

From the above general discussion it is evident that the transient stability, at a given power transmission level and fault clearing time, is determined by the P versus δ characteristic of the post fault system. Since, as previously shown, shunt and series compensation and angle control improve the natural transmission characteristic of the system, it can be expected that the judicious employment of these techniques would be highly effective in increasing the transmission capability of the post-fault system and thereby enhancing transient stability.

For comparison, consider the four basic two-machine (sending-end generator, receiving-end infinite bus) systems, with no compensation, mid-point shunt compensation, series compensation, phase angle control, and generalized, active and reactive power flow control, as shown in Figures 1-3, 1-6, 1-8, 1-9 and 1-10. For clarity, the above introduced equal-area criterion is applied here in a greatly simplified manner, with the assumption that the original single line model represents both the pre-fault and post-fault systems. (The impracticality of the single line system and the questionable validity of this assumption have no effect on this qualitative comparison.) Suppose that in the uncompensated and all four compensated systems the steady-state power transmitted is the same. Assume that all four systems are subjected to the same fault for the same period of time. The dynamic behavior of the four systems is illustrated in Figures 1-13a through 1-13e. Prior to the fault, each of the four systems transmits power Pm (subscript m stands for “mechanical”) at angles δ1, δp1, δs1, δa1, and δpq1, respectively. (Subscripts p, s, a, and pq stand for “parallel”, “series”, “angle” and “active and reactive power flow”.) During the fault, the transmitted electric power (of the single line system considered) becomes zero while the mechanical input power to the generators remains constant (Pm). Therefore, the sending-end generator accelerates from the steady-state angles δ1, δp1, δs1, δa1, and δpq1 to angles δ2,

1-29

δp2, δs2, δa2, and δpq2 at which the faults clears. The accelerating energies in the four systems are represented by areas A1, Ap1, As1, Aa1, and Apq1. After fault clearing, the transmitted electric power exceeds the mechanical input power and the sending-end machine decelerates, but the accumulated kinetic energy further increases until a balance between the accelerating and decelerating energies, represented by areas A1, Ap1, As1, Aa1, Apq1 and A2, Ap2, As2, Aa2, Apq2, respectively, is reached at δ3, δp3, δs3, δa3, and δpq3. The difference between angles δ3, δp3, δs3, δa3, and δpq3, representing the maximum angular swings, and the critical angles δcrit, δpcrit, δscrit, δacrit, and δpqcrit determines the margin of transient stability, that is, the “unused” and still available decelerating energy, represented by areas Amargin, Apmargin, Asmargin, Aamargin, and Apqmargin.

mP

PP

max

crit

A

A

1δ 32 δδ δ

1

(a)

2

=P

A

π δ

margin

sinLX

V 2δ

max

P

P2V

XA

3

A

1pδ δ δpp2

p1

(b)

maxp2A

pcritδ π= δ

pmargin

P

2pP

2

L

δsin2

Psin

σ

A

δ0 a1δ δ

a1

a2

(d)a3 π +acritδ π

maxP

σ

max

Pm

A a2

V 2P

amarginALX

δ

σ )−δ(σ

2Ppq = X

V sin ) δVpq ( V

sin δ+ X

ρ2 +

0

maxP

1.5 Pmax

P

Apq1

pqcritpq1δ δpq2 pq3δδ

(e)δ

pq2A

A pqmargin

Pm

π=

sin

0

maxP

1.5 Pmax

P

As1

scrit1sδ δ3s2 sδδ

(c)δ

P

k

s2A

A smargin

s X1/3=

(1-k)L

2V δ

Pm

π

Figure 1-13 Equal-area criterion to illustrate the transient stability margin for a simple two machine system, (a) without compensation, (b) with an ideal mid-point compensator, (c) with a series capacitor, (d) with a phase shifter, and (e) with a generalized power flow controller.

Comparison of Figures 1-13a through 1-13e clearly shows a substantial increase in the transient stability margin that the various compensation approaches can provide through the control of different system parameters. The shunt-connected (“parallel”) var compensation method provides the improvement by segmenting the transmission line and regulating the midpoint voltage. The series capacitive compensation approach reduces the effective transmission impedance and minimizes the transmission angle. Phase angle control keeps the transmission angle, δ-σ, at π/2 for maintaining maximum power transmission while the generator angle δ swings beyond this value. The active and reactive power flow controller forces the desired power flow by the injecting the necessary voltage in series with the line.

The use of any one of the four compensation approaches obviously can increase the transient stability margin significantly over that of the uncompensated system. However, it should be noted that the diagrams illustrate ideal cases, the potential of various compensation techniques

1-30

for improving transient stability. The theoretically attainable (and practically unnecessarily large) increase in the transient stability margin represented by the yellow-colored areas in the figures could be obtained only at economically unacceptable MVA ratings, particularly for the shunt and series capacitive compensators. Still, the improvements in transient stability usually required in practice can be achieved most of the time at reasonable MVA rating and at competitive cost.

If the uncompensated system has a sufficient transient stability margin, the compensation techniques described can considerably increase the transmittable power while maintaining the necessary stability margin.

In the explanation of the equal area criterion at the beginning of this section, a clear distinction was made between the “pre-fault” and “post-fault” power system. It is important to note that from the standpoint of transient stability, and thus of overall system security, the post-fault system is the one that counts. That is, power systems are normally designed to be transiently stable, with defined pre-fault contingency scenarios and post-fault system degradation, when subjected to a major disturbance (fault). Because of this (sound) design philosophy, the actual capacity of transmission systems is considerably higher than that at which they are normally used. Thus, it may seem technically plausible (and economically savvy) to employ fast acting compensation techniques, instead of overall network compensation, specifically to handle dynamic events and increase the transmission capability of the degraded system under the contingencies encountered.

Conclusions on transient stability

1. Equal area criterion is an effective tool for the visual understanding and comparative evaluation of methods for transient stability improvement.

2. All shunt and series compensation methods, as well as angle and power flow control are effective for transient stability improvement with sufficiently fast control action.

3. Because of the powerful effect of reactive series compensation and power flow control with directly injected forcing voltage, these can be expected to give greater increase in transient stability at a given equipment rating that reactive shunt compensators and phase angle regulators.

1.2.6.2 Power Oscillation Damping

In an under-damped power system any minor disturbance can cause the machine angle to oscillate around its steady state value at the natural frequency of the total electromechanical system. The angle oscillation, of course, results in a corresponding power oscillation around the steady-state power transmitted. The lack of sufficient damping can be a major problem in some power systems and, in some cases it may be the limiting factor for the transmittable power.

Until the late 1970s, the excitation control of rotating synchronous machines was the available active means for power oscillation damping. Later technological developments made it possible to realize rapidly controllable reactive shunt and series compensators, transmission angle and power flow controllers to provide highly effective power oscillation damping. (Since compensations can be used for power flow control, and angle or power flow control can be considered as compensation of specific system parameters, the general term of “compensator” or “Controller” is often used for brevity to refer to the group of these equipments and “compensation” for their action.)

1-31

Since power oscillation is a sustained dynamic event, it is necessary to vary the applied compensation so as to achieve consistent and rapid damping. The control action required is in principle the same for all of compensation approaches. That is, when the rotationally oscillating generator accelerates and angle δ increases (dδ/dt > 0), the electric power transmitted must be increased to compensate for the excess mechanical input power. Conversely, when the generator decelerates and angle δ decreases (dδ/dt < 0), the electric power must be decreased to balance the insufficient mechanical input power. (The mechanical input power is assumed to be essentially constant in the time frame of an oscillation cycle.)

The requirements of output control, and the process of power oscillation damping, by the four compensation approaches, are illustrated in Figure 1-14a through 1-14f. Waveforms at a show the undamped and damped oscillations of angle δ around the steady-state value δ0. Waveforms at b show the undamped and damped oscillations of the electric power P around the steady-state value Po. (The momentary drop in power shown in the figure represents an assumed disturbance that initiated the oscillation.)

(d) 0

0(e)

σ

k

Q(c) 0

p

t

t

t

t

P(b) o

P

δ

(a) δo

Undampedt

Undamped

0(f)t

Vpq

Figure 1-14 Waveforms illustrating power oscillation damping by shunt compensator, series compensation, phase shifter, and generalized power flow controller: (a) generator angle, (b) transmitted power, (c) var output of a shunt compensator, (d) degree of series compensation, (e) angle variation of a phase shifter, and (f) active power variation by generalized flow controller.

Waveform c shows the reactive power output Qp of a shunt-connected var compensator. The capacitive (positive) output of the compensator increases the midpoint voltage and the transmitted power when dδ/dt > 0, and it decreases those when dδ/dt < 0.

Waveform d shows the required variation of k=XC/X for series capacitive compensation. When dδ/dt > 0, k is increased and thus the line impedance is decreased. This results in the increase of

1-32

the transmitted power. When dδ/dt < 0, k is decreased (in the illustration it becomes zero) and the power transmitted is decreased to that of the uncompensated system.

Waveform e shows the variation of angle σ produced by the phase shifter. (For the illustration it is assumed that σ has an operating range of -σmax ≤ σ ≤ σmax, and δ is in the range of 0 < δ < π/2.) Again, when dδ/dt > 0, angle σ is negative making the power versus δ curve (refer to Figure 1-9c) to shift to the left, which increases the angle between the end terminals of the line and, consequently, also the real power transmitted. When dδ/dt < 0, angle σ is made positive, which shifts the power versus angle curve to the right and thus decreases the overall transmission angle and transmitted power.

Waveform f shows the injected forcing Vpq voltage to increase active power flow during machine acceleration (dδ/dt > 0) and decrease it during deceleration (dδ/dt < 0). During damping, the angle ρ of the injected voltage is kept at ±π/4, at which the forcing voltage controls exclusively the active power flow.

As the illustrations show, a “bang-bang” type control (output is varied between minimum and maximum values) is assumed for all four compensation approaches. This type of control is generally considered the most effective, particularly if large oscillations are encountered. However, for damping relatively small power oscillations, a strategy that varies the controlled output of the compensator continuously, in sympathy with the generator angle or power, may be preferred.

To detect accurately machine acceleration and deceleration to control the output of the compensators is usually location dependent and the selection of the useful signals for control may require studies and experimentations.

Conclusions on power oscillation damping

1. Power oscillation damping can be achieved with all compensators and power flow controllers by varying their output so as to increase active power transmission when the involved machines are accelerating, and decreasing it when they are decelerating.

2. The derivation of useful control signal indicating machine acceleration and deceleration may require studies and experimentations in practice.

1.2.6.3 Increase of Voltage Stability Limit

Voltage collapse can be a potential problem primarily for buses supplying load areas with heavy motor loads. Radial lines are particularly prone to this problem. The solutions for all cases are similar and illustrated here for the simple radial system with feeder line reactance of X and load impedance Z, shown in Figure 1-15a, together with the normalized terminal voltage Vr versus power P plot at various load power factors, ranging from 0.8 lag and 0.9 lead. The “nose-point” at each plot given for a specific power factor represents the voltage instability corresponding to that system condition. It should be noted that the voltage stability limit decreases with inductive loads and increases with capacitive loads.

1-33

s r(a) V V

X

Z 0.5

Vr (pu)

0.95 lag

0.50

1.0

0.8 lag

P (pu)1.5

unity pf

0.9 lead

1.0

0.97 lead

I

X

+- Q

Z(b) s rV V

I

Varcomp. 0.5

1.0

0.95 lag

0.50

1.0

0.8 lag

P (pu) 1.5

unity pf

0.9 lead 0.97 lead

Vr (pu)

Z=R

X

(c) s rV V

I1.0

0.5

1.00

2.0 3.0 P (pu)

CX =0X =0.75 X

CX =0.5 XC

Vr (pu)XC

Figure 1-15 Variation of voltage stability limit with load and load power factor (a), extension of voltage stability limit by reactive shunt compensation (b), and extension of voltage stability limit by series capacitive compensation (c).

The inherent circuit characteristics of the simple radial structure, and the Vr versus P plots shown, clearly indicate that both shunt and series capacitive compensation can effectively increase the voltage stability limit. Shunt compensation does so by supplying the reactive load and regulating the terminal voltage (V-Vr=0) as illustrated in Figure 1-15b. Series capacitive compensation does so by canceling a portion of the line reactance X and thereby in effect providing a “stiff” voltage source for the load, as illustrated for a unity power factor load in Figure 1-15c.

Conclusions on voltage stability

1. The problem of voltage instability and collapse occur usually at buses supplying loads with heavy motor loads.

2. Both shunt and series compensation ban be used for voltage stability increase. The shunt compensation provides closer regulation, the series compensation a smoother, more manageable transition from stable to unstable state.

1-34

1.3 Transmission Problems and Needs: The Role of Power Electronics

1.3.1 Problems and needs

The physical laws of ac power systems impose basic limitations on the transmittable power, apart from the current conducting capacity of the line. These limitations are due partly to the electro-mechanical characteristic of the ac machines, and partly to the electrical LC lattice network type characteristic of the ac transmission lines. The nature of generator imposed limitation is maintainable rotational stability, to prevent either a run-away speed and frequency change (transient instability) or periodic frequency variation with corresponding oscillation of the transmitted power between the generators and other synchronous elements of the system (dynamic instability). This necessitates the limitation of the transmitted power to a level at which the generators can recover normal synchronous operation following a defined (worst case) major and minor disturbances in the system. The transmission line imposed limitation is acceptable voltage variation along the line since the series line inductance causes increasing voltage droop with increasing load (and/or increasing lagging load power factor), and the line capacitance causes growing voltage increase with decreasing load. The voltage variation, primarily voltage decrease, sets the limit for the line-length dependent transmittable power.

As shown in the previous sections, the natural limits of ac power transmission can be increased, theoretically to any extent by line compensation and power flow control techniques, which change the natural impedance characteristic of the line and control the otherwise “free”, impedance-defined power flow. The line impedance caused steady-state voltage variation can be minimized by fixed and mechanically switched series and shunt reactive compensation, however, the management of the dynamic rotational and voltage variation problems requires fast control actions. The application of fast controls in the execution of shunt- and series-compensation, as well as angle and power flow control, makes it possible to reduce dynamic voltage and frequency (machine speed) variation and thereby increase the transmittable power without decreasing the necessary limits established for secure system operation. In other words, fast controls capable of counteracting dynamic system variations allow higher utilization of transmission assets.

Conclusions on Transmission Problems and System Needs

1. Transmission limitations are caused by the rotational stability problems of generators due to disturbances, and to voltage variations with changing load due to the reactive line impedance.

2. Steady-state transmission limits can be increased by fixed, or mechanically switched reactive compensation; the improvement of dynamic rotational and voltage stability requires fast control actions.

1.3.2 The Role of Power Electronics

Dynamic voltage variation and rotational stability problems of ac power transmission, which had necessitated the under-utilization of lines and other system assets, provided the incentives in the late 1970s to introduce power electronics-based control for reactive compensation. This normal evolutionary process has been greatly accelerated by subsequent developments in the utility industry in the subsequent decades, which have aggravated the early problems and highlighted the structural limitations of power systems in a greatly changed socio-economical environment

1-35

(see Section 1.1 Historical Perspective). The desire to offer economical, and rapidly executable solutions to these problems by means of increased utilization of system assets, led to focused technological developments under the Flexible AC Transmission System (FACTS) initiative of the Electric Power Research Institute (EPRI) to provide power electronics-based Controllers for real time control of voltage and power flow in ac transmission systems.

It is to be noted that the aim of the power electronics-based Controllers is higher utilization of transmission assets and greater freedom in managing power transmission, particularly under heavy power demand or emergency conditions. The application of power electronics is, in general, not a substitute for new lines, but a powerful means to achieve higher utilization of all lines, old and new, in the system.

Conclusion on Role of Power Electronics

Power of electronics by means of providing rapidly controllable compensation and power flow control of the line can increase the transmittable power and the utilization of the transmission assets.

1.3.3 The Objectives of FACTS

Modern transmission networks are increasingly expected to facilitate the transfer of electric power from any supplier to any consumer over a large geographic area under market forces (supply vs. demand) controlled, and thus continuously varying, pattern of contractual arrangements. The realization of such a network is significantly constrained by construction cost, right-of-way, and environmental restrictions. Thus, the optimization and full utilization of existing transmission assets, together with systematically planned and executed construction of new facilities, appears to be the only realistic approach for meeting the complex transmission requirements of modern socio-economic system.

The Electric Power Research Institute, after years of supporting the development of high power electronics for reactive compensation, formalized the broad concept of Flexible AC Transmission System (FACTS) in the late 1980s. The acronym FACTS identifies alternating current transmission systems incorporating power electronics-based controllers to enhance the controllability and increase power transfer capability. The FACTS initiative was originally launched to solve the emerging system problems in the late 1980s due to restrictions on transmission line construction, and to facilitate the growing power export/import and wheeling transactions among utilities, with two main objectives:

1. To increase the power transfer capability of transmission systems, and

2. to keep power flow over designated routes.

The first objective implies that power flow in a given line could be increased up to the thermal limit by forcing the necessary current through the series line impedance if, at the same time, stability of the system is maintained via appropriate real-time control of power flow during and following system faults. This objective of course does not mean to imply that the lines would normally be operated at their thermal limit loading (the transmission losses would be unacceptable), but this option would be available, if needed, to handle severe system contingencies. However, by providing the necessary rotational and voltage stability via FACTS Controllers, instead of large steady-state margins, the normal power transfer over the

1-36

transmission lines is estimated to increase significantly (in some cases up to about 50%, according to some studies conducted).

The second objective implies that, by being able to control the current in a line (by, for example, changing the effective line impedance), the power flow can be restricted to selected (contracted) transmission corridors (which have been contracted and have the capacity) while parallel and loop-flows can be mitigated. It is also implicit in this objective that the primary power flow path must be rapidly changeable to an available secondary path under contingency conditions to maintain the desired overall power transmission in the system.

It is easy to see that the achievement of the two basic objectives would significantly increase the utilization of existing (and new) transmission assets, and could play a major role in facilitating deregulation with minimal requirements for new transmission lines.

The implementation of the above two basic objective requires the development of high power compensators and controllers. The technology needed for this is high power (multi-hundred Mva) electronics with real-time operating control. However, once a sufficient number of these fast compensators and controllers are deployed over the system, the coordination and overall control to provide maximum system benefits and prevent undesirable interactions with different system configurations and objectives, under normal and contingency conditions, present a different technological challenge. This challenge is to develop appropriate system optimization control strategies, communication links, and security protocols. The realization of such an overall system optimization control can be considered as the third, and still a future, objective of the FACTS initiative.

Conclusions on the Objectives of FACTS

The overall objective of FACTS is to economically solve transmission problems, increase transmission capacity by higher utilization of system assets, and provide improved flexibility for power export and import among utilities, using state-of-the-art power electronics-based compensators and controllers. To this end, FACTS is to

1. increase power transmission capability, up to thermal limit if needed, with the necessary transient, dynamic and voltage stability, and

2. restrict power flow along defined transmission corridors, and eliminate (minimize) capacity limiting loop flows.

1.4 Transmission Compensators and Controllers

1.4.1 Theoretical concepts

Shunt compensators are primarily used to regulate the transmission voltage at critical system locations under varying load and system conditions. Thus, an ideal shunt compensator is functionally a controllable ac current source that can force the necessary current through the effective line impedance to maintain the desired bus voltage by negating the voltage drop caused by the prevailing line (load) current, as illustrated in Figure 1-16a. Since the line impedance is substantially reactive, the voltage regulation can be achieved by purely reactive (capacitive or inductive) compensating current. There are two approaches to the realization of controllable

1-37

(reactive) current sources which can be implemented by practical power electronic circuits: one is controllable reactive admittance (Figure 1-16b) and the other is a controllable voltage source in series with a tie reactance (Figure 1-16c). The adjustment of the admittance changes the compensating current drawn by the compensator from the bus; and similarly, the magnitude of the voltage source with respect to the prevailing bus voltage determines the voltage across its tie reactance and thereby the compensating current drawn from the bus.

sreg

(a) VV

X1I

(b)

(c)

X2

compI

Controllable reactivecurrent source

sreg

VV

X1I X2

compI

Controllablereactive admittance

sreg

VV

X1I X2

compI

Controllable reactivevoltage source

Xcomp

±Ycomp

Figure 1-16 Representation of an ideal shunt compensator by a controllable reactive current source (a) and its realization by a controllable reactive admittance (b), and by a controllable voltage source with reactive tie impedance (c).

Series compensators are primarily used to control transmission line current and thereby the transmitted power. Thus, an ideal series compensator is functionally a controllable ac voltage source, as illustrated in Figure 1-17a, that can inject the compensating voltage in-phase, or anti-phase, with the prevailing voltage drop across the line impedance to increase or decrease the line current, as if the effective line impedance was changed. Since the series line impedance is inductive, the voltage across is in quadrature with the line current. Therefore, the compensating voltage injected to increase or decrease the line current has also to be in quadrature with the line current. Therefore, the series compensator does not exchange active power with the line and thus it can be realized by a controllable reactive impedance in series with the line, or by a controllable voltage source whose phase angle is kept rigidly at ±90° with respect to the line current, as shown in Figures 1-17b and 1-17c, respectively.

1-38

s(a) V

X1I

(b)

(c)

X2

compV

sV

X1I X2

Controllablereactive impedance


compV

±Xcomp


sV

X1I X2compV

Figure 1-17 Representation of an ideal series compensator by a controllable reactive voltage source (a) and its realization by a controllable reactive impedance (b), and by a controllable reactive voltage source with transformer coupling (c).

Phase Shifters (Angle Regulators) and general active and reactive power flow controllers insert the compensating voltage with respect to a selected bus voltage to control the corresponding effective transmission angle. For phase shifting, the injected voltage is adjusted so as to establish the desired transmission angle between the sending- and receiving-end voltages, regardless of the phase angle it may have with respect to the line current. For the independent control of active and reactive power flow, the injected voltage is controlled so as to establish the necessary magnitude and phase angle for the line current with respect to the bus voltage of interest (typically the receiving-end), at which the desired active and reactive line power is defined. It is evident that the source of the injected voltage used for phase shifting, as well as that employed for power flow control, must be able to exchange both reactive and active power with the ac system. Thus, the realization of these by controllable impedance, similar to the series reactive compensator) would require both a variable reactive impedance (capacitance or inductance) and variable positive or negative resistance (power sink and source) in series with the line. Of course, in a practical implementation only the reactive power exchange can be supported by external elements (capacitor, reactor, or reactive power generator like a synchronous compensator), the active compensating power exchanged through series compensation must be supplied or absorbed by the ac system itself. Two conceptual implementations functionally suitable for both phase shifting and power flow control are shown in Figures 1-18a and 1-18.b. A tap-changing transformer arrangement with (±) in-phase and quadrature voltage injection capability (Figure 1-18a), can, with sufficient number of taps, rotate the inserted voltage from zero to 360° and change its magnitude from zero to a maximum value. In this case, however, both the active and reactive power exchanged through series compensation is supplied or absorbed by the sending-

1-39

(or receiving-) end bus via the tap-changing transformers. An idealized motor-generator set with four-quadrant operation (Figure 1-18b) is an equivalent conceptual solution in which, however, only the active power exchanged in the compensation is provided by the ac system, the reactive power is supplied by the generator itself.

(b) Ideal synchronous motor-generator set

sV

X1I X2pqV

Voltagecontrol

Voltagecontrol

(a)In-phase and quadratureTap-changer

sV

X1I X2pqV

+

+

--

Figure 1-18 Conceptual realizations of an ideal phase shifter or generalized power flow controller by in-phase and quadrature tap-changers (a), and by an ideal synchronous motor-generator set (b).

It should be noted that the shown realizations of the general power flow controller can be restricted to those of specific types of phase shifters. In particular, if the magnitude of the phase-shifted voltage is not required to stay constant, then the in-phase voltage insertion can be omitted, which leads to considerable circuit simplification for the tap-changing transformer type of realization. (However, it does not change the conceptual motor-generator arrangement.) Using only quadrature voltage insertion, the phase-shifted voltage increases with the angle of shift, σ, i.e., Vshifted =V/cosσ. This type of phase-shifter is often referred to quadrature booster. It should also be noted that the tap-changing transformer with in-phase voltage insertion is often used separately in practical applications for voltage regulation.

The conceptual representations of the basic transmission compensator and controllers illustrated in Figures 1-16 through 1-18 can, as will be seen, be realized functionally faithfully by power electronic circuits.

Conclusions on compensator and power flow concepts

1. Reactive shunt compensation increases transmittable power by maintaining transmission voltage, series compensation by decreasing line impedance, and phase shift by controlling transmission angle.

1-40

2. Ideal shunt compensator is a controllable reactive current source that can be realized by a controllable reactive admittance (susceptance), or by a controllable reactive voltage source in series with a fixed reactive tie impedance (reactance).

3. Ideal series compensator is a controllable reactive voltage source with respect to the line current that can be realized by a controllable series reactive impedance, or by a controllable reactive voltage source.

4. Ideal phase shifter or power flow controller is functionally a voltage source with controllable magnitude and phase angle that can exchange both active and reactive power with the ac system. It can be realized by combination of in-phase and quadrature tap-changing transformers, or by a coupled arrangement of two voltage sources functioning as an ideal, four-quadrant motor-generator set.

5. A quadrature booster is a phase-shifter, whose injected voltage is maintained in quadrature with respect to the controlled bus voltage.

1.4.2 Power electronics-based realization of basic transmission compensators and controllers

The development of power electronics-based compensators and controllers has followed two distinctly different technical approaches, as the conceptual representations in the previous section suggest, both resulting in a comprehensive group of controllers able to address targeted transmission problems. The first group employs reactive impedances or tap-changing transformers with conventional thyristor switches (can be turned on by gate control, but turns off only if the current becomes zero) as controlled-elements; the second group uses static switching-converters with gate turn-off semiconductors (GTO, GCT, IGBT, etc.), as controlled voltage sources.

The thyristor-controlled compensators/controllers have a common characteristic in that the necessary reactive power required for the compensation is generated or absorbed by traditional capacitor or reactor banks (shunt and series reactive compensators), or drawn from the ac system (phase-shifter), and the thyristor switches (valves) are used only to set the combined reactive impedance these banks present to ac the system or connect the necessary voltage steps for phase shifting.

By contrast, the compensator/controller approach based on static switching-converters, operated as synchronous voltage sources, is able to generate internally, without capacitors and inductors, the reactive power needed for reactive compensation. In addition, they can also exchange active power with ac system and thus also facilitate real power compensation and asynchronous power transfer.

1.4.2.1 Shunt Compensators: SVC and STATCOM

Static Var Compensator (SVC). A typical shunt-connected static var compensator (SVC), composed of thyristor-switched capacitors (TSCs) and thyristor-controlled reactors (TCRs), is shown in Figure 1-19a. With proper coordination of the capacitor switching and reactor control, the var output can be varied continuously between the capacitive and inductive ratings of the equipment.

1-41

Reactorbanks

Couplingtransformer

L

Thyristorvalves

Capacitorbanks

C C

Auxiliaryinputs

Parametersetting

L

Transmission line

Control

PT

refV

(a)

CmaxCapacitive

C

II

0

VT

InductiveLmaxI

LI

Transientrating

(b)

Figure 1-19 Static Var Compensator employing thyristor-switched capacitors and thyristor-controlled reactors (a), and its V-I regulating characteristic (b).

The compensator is normally operated to regulate the voltage of the transmission system at a selected bus. The V-I characteristic of the SVC, shown in Figure 1-19b, indicates that regulation with a given slope around the nominal voltage can be achieved in the normal operating range defined by the maximum capacitive and inductive currents of the SVC. However, the maximum obtainable capacitive current decreases linearly (and the generated reactive power in quadrature) with the system voltage since the SVC becomes a fixed capacitor when the maximum capacitive output is reached. Therefore, the voltage support capability of the conventional thyristor-controlled static var compensator rapidly decreases with decreasing system voltage.

Static Synchronous Compensator (STATCOM). The STATCOM is analogous to an ideal, rotating synchronous compensator (condenser), which is, as known, a synchronous machine generating a balanced set of (three) sinusoidal voltages at the fundamental frequency, with controllable amplitude and phase angle. This ideal machine has no inertia, its response is practically instantaneous and it can internally generate reactive (both capacitive and inductive) power by excitation control.

The basic switching converter equivalent of the rotating machine is a matrix of solid-state switches which connect a dc voltage source sequentially to the three output terminals so as to generate a set of three balanced voltages. This type of converter is called voltage-sourced converter (VSC). The simplest realization of VSC, as a synchronous voltage source, is shown in Figure 1-20a in the functional role of synchronous compensator (condenser). (It is to be noted that this simple switch matrix would only produce a quasi-square wave output. In practice more complex switch arrangements and/or special modulation techniques are used to produce outputs that can sufficiently approximate sine-waves.) Since in this role the converter-based voltage

1-42

source is used strictly for reactive shunt compensation, like the rotating synchronous compensator or the conventional static var compensator, thus it exchanges no active power (except for losses) and the dc voltage can be sustained by a relatively small dc capacitor. The converter itself keeps the capacitor charged to the required voltage level by making the ac output voltage of the converter to slightly lag the system voltage, and thereby absorb a small amount of active power from the ac system to replenish the internal operating losses. This method can also be used to increase or decrease the capacitor voltage, and thereby the amplitude of the output voltage of the converter, for the purpose of controlling the var generation or absorption in the same way as done by the excitation control of a rotating synchronous machine, as illustrated in Figure 1-20b.

Energy

storage

+

Switchingconverter

-

C

gcV V Vga gb

source or

Idc

(a)

AC system

I

CouplingTransformer

q

TV

dc

(b)

AT DC TERMINAL

dcI 0 +V

0

Absorbs

T

Supplies

V

< VT

> TV +Iq

AT AC TERMINAL

"Q"

"Q"

Vdc DC terminal

Generatorac terminalVg

(+Idc

Auxiliaryinputs

Parametersetting

Control

PT

refV

(c)

CmaxCapacitive

C

II

0

VT

InductiveLmaxI

LI

Transientrating

1

Transientrating

+-Vg

Vg

Vg Vg

VgVg-∆ +∆-Iq

Vg∆

)~

Figure 1-20 STATCOM employing a voltage-sourced converter (a), var output control by output voltage variation (b), and V-I regulating characteristic (c).

The V-I characteristic of the STATCOM is shown in Figure 1-20c. As illustrated, the STATCOM can provide both capacitive and inductive compensation and is able to control its output current over the rated maximum capacitive or inductive range independently of the ac system voltage. That is, the STATCOM can provide full capacitive output current at any system voltage, close to zero. By contrast, the SVC can supply only diminishing output current with decreasing system voltage, as Figure 1-19b illustrates.

1.4.2.2 Series Compensators: TCSC and SSSC

Thyristor-Controlled Series Capacitor (TCSC). There are two basic schemes of thyristor-controlled series capacitors: one uses thyristor-switched capacitors, as shown schematically in Figure 1-21a, and the other employs a fixed capacitor in parallel with a thyristor-controlled

1-43

reactor, as shown Figure 1-21b. (Note that the former is often referred to as Thyristor-Switched Capacitor.)

(b)

PT

(a)CT FC

C

Thyristor-ControlledReactor

Transmissionline

Tyristor valve

Parameter

Controlsetting

inputs

Control

C 1 C n

PT

CT FC

Parameter

Controlsetting

inputs

Control

Transmissionline

Tyristor valve

k

2.0

k = 0.2

k = 0

ππ/2

X L

CX=

δ

k = 0.33

(1-k)P =s X L

Vsin δ

2

(C)

0

P (pu)

1.0

1.5

Figure 1-21 Controllable series compensator scheme using thyristor-switched series capacitors (a), or a thyristor-controlled reactor parallel with a fixed capacitor (b), and transmitted power vs. transmission angle characteristic with variable series capacitive compensation (c).

In the thyristor-switched capacitor scheme of Figure 1-21a, the degree of series compensation is controlled by increasing or decreasing the number of capacitor banks in series. To accomplish this, each capacitor bank is inserted or bypassed by a thyristor valve (switch). To minimize switching transients and utilize “natural” commutation, the operation of the thyristor valves is coordinated with voltage and current zero crossings. Since, the voltage across the series capacitor is a direct function of the line current, the prevention of damaging overvoltage during faults and other surge current conditions usually necessitates the use of a ZnO type voltage clamping device or other by-pass arrangement in parallel with the thyristor-switched capacitor banks.

In the fixed-capacitor, thyristor-controlled reactor scheme of Figure 1-21b, the degree of series compensation in the capacitive operating region is increased (or decreased) by increasing (or decreasing) the thyristor conduction period, and thereby the current in the TCR. Minimum series compensation is reached when the TCR is off. The TCR may be designed to have the capability to limit the voltage across the capacitor during faults and other system contingencies of similar effect.

The transmitted power versus transmission angle as a function of series capacitive compensation is shown in Figure 1- 21c.

The two schemes may be combined by connecting a number of basic TCR with fixed parallel capacitor modules in series in order to achieve greater control range and flexibility.

1-44

Static Synchronous Series Compensator (SSSC). The concept of the SSSC is based on the recognition that the function of the series capacitor is simply to produce an appropriate voltage at the fundamental ac system frequency to increase the voltage across the inductive line impedance, and thereby the fundamental line current and the transmitted power. (This of course has the same electrical effect as if the series line inductance was reduced to that of a shorter line.) Therefore, an ac voltage source of fundamental frequency, similar to that employed in the STATCOM, inserted in series with the line with a locked quadrature (lagging) phase relationship to the line current, and whose amplitude is made proportional to that of the line current, i.e., Vc=kXI (where, as before, Vc is the amplitude of the inserted voltage, X is the line impedance, k is the degree of series compensation, and I is the amplitude of the line current), provides series reactive compensation equivalent to that obtained by a series capacitor, as illustrated in Figure 1-22a. However, in contrast to the series capacitor, the SSSC is able to maintain a constant compensating voltage in face of variable line current, or control the amplitude of the injected compensating voltage independent of the amplitude of the line current. (This is analogous to the STATCOM, which is able to maintain a constant compensating current in face of grossly decreased bus voltage.)

Energy

storage

+Switchingconverter

-

C

gcV V Vga gb

source or

Idc

(a)

Transmission line

I

CouplingTransformer

line

qV

(b)

VdcDC terminal

Gen. acterminalVg

Auxiliaryinputs

Parametersetting

Control

PT

q,refV

Iline

Iline

q+V

q-V

CT

X

qV = -jkX IlineFor the emulation of series capacitor

-0.5

0

0.5

1.0

P (pu)

1.5

q

V = -0.707

V = -0.353

V = 0

q

q

q

V = 0.353

V = 0.707

π/2

q

π δ

qVsinVP =

X

2 V+δ

Xδcos2

= IjVqI

VqFor line currentIndependent compensation

Figure 1-22 SSSC employing a voltage-sourced converter (a), and its transmitted power vs. transmission angle characteristic (b

For normal capacitive compensation, the output voltage lags the line current by 90 degrees. However, the output voltage of the SVS can be reversed by simple control action to make it lead the line current by 90 degrees. In this case, the injected voltage decreases the voltage across the inductive line impedance and thus the series compensation has the same effect as if the reactive line impedance was increased.

1-45

The transmitted power P versus transmission angle characterizing the SSSC is illustrated in Figure 1-22b. Comparison of these plots to those shown in Figure 1-21c for the series capacitor (of comparable compensation range) clearly shows that the series capacitor increases the transmitted power by a fixed percentage of that transmitted by the uncompensated line at a given transmission angle and, by contrast, the SSSC increases it by a fixed fraction of the maximum power transmittable by the uncompensated line, independent of the transmission angle, in the important operating range of 0 through π/2. From the standpoint of practical applications, steady-state power flow control or stability improvements, the SSSC clearly has considerably wider control range than the controlled series capacitor of the same Mva rating.

1.4.2.3 Phase Shifters and Power Flow Controllers: TCPAR and UPFC, IPFC, BtB

Thyristor-Controlled Phase Angle Regulator (TCPAR). Although there is no high power, non-mechanical phase-shifter in service, the principles for using a phase-shifting transformer with a thyristor tap-changer are well established. To avoid excessive circuit complexity, the Thyristor-Controlled Phase Angle Regulator, just as its commonly used conventional counterpart with a mechanical tap-changer, provides quadrature voltage injection, that is, it is in effect a quadrature booster.

A Thyristor-Controlled Phase Angle Regulator is shown in Figure 1-23a. It consists of a shunt-connected excitation transformer with appropriate taps, a series insertion transformer and a thyristor valve arrangement connecting a selected combination of tap voltages to the secondary of the insertion transformer. The excitation transformer has three non-identical secondary windings, in proportions of 1:3:9. It can produce a total of 27 steps using only 12 thyristor valves (of three different voltage ratings, i.e., the voltage rating of the valves is also in the 1:3:9 ratio) per phase with a switching arrangement that can bypass a winding or reverse its polarity.

V

Thyristortap-changer

Transmission

transformerExcitation

line

9

3 Control

Parametersetting

Referenceinput

Measured variables

Series transformer

V

1

σ

V*

(a) (b)

sin

0 /2π

P2

XV (=P

-0.66-0.33

-1.0

δπ

σ =Vσ =V

σ =V

δcosδ +VVσ )

0.660.33

1.0

σ =Vσ =V

σ =VVσ

σV

V*

2(VV* 2Vσ+= )½ σ = 0V σ = 0V

Figure 1-23 Thyristor-controlled “quadrature booster” type of phase shifter (a), and its transmitted power vs. transmission angle characteristic (b).

1-46

The phase angle requirements for power flow control can be determined from angle measurements, if available, or from power measurements. The TCPAR could be applied to regulate the transmission angle to maintain balanced power flow in multiple transmission paths, or to control it so as to increase the transient and dynamic stability of the power system. The transmitted power versus transmission angle characteristic of the TCPAR, with the described practical realization of the quadrature booster, is shown in Figure 1-23b.

It is to be noted that the phase angle between the voltage injected by the TCPAR (which is, by design, in quadrature with the line to neutral terminal voltage) and the line current is arbitrary, determined by the pertinent parameters of the overall power system. This means that, in general, the TCPAR must exchange, via the series insertion transformer, both active and reactive power with the ac system. Since this type angle regulator cannot generate, or absorb, either real or reactive power, it follows that both the active and reactive power that it supplies to, or absorbs from the line when it injects quadrature voltage must be absorbed from it, or supplied to it, by the ac system. As a consequence, the Mva ratings of the excitation and insertion transformer are substantially the same.

The fact that the tap-changing transformer type angle regulator cannot generate or absorb reactive power is a disadvantage in those applications in which the reactive power exchanged has to be transmitted through a line of appreciable length, due to the resulting voltage drop. In these cases, to avoid large voltage drops across the line, the TCPAR needs to be complemented with some type of controllable var supply, or located close to the power generator.

Converter-Based Angle Regulators and Power Flow Controllers. Converter-based phase-shifters and more general power flow controllers employ the general functional concept of the ideal synchronous motor/generator set, shown in Figure 1-18b, using two voltage-sourced converters in back-to-back configuration with a common dc link as illustrated in Figure 1-24. Such an arrangement is a close approximation of the ideal motor-generator set capable of full four-quadrant operation: Active power flow is fully controllable between the ac terminals in either direction, and each ac terminal can independently supply or absorb reactive power in the defined operating range, independent of the active power transfer within Mva rating of the equipment. With this general arrangement either ac terminal can be coupled in shunt or in series with the transmission line and thus four different power flow controller arrangements, shunt-series, series-series, and shunt-shunt, which are referred to as Unified Power Flow Controller (UPFC), Interline Power Flow Controller (IPFC), and Back-to-Back Tie (BtB) can be realized.

1-47

Switchingconverter

gcV V Vga gb

CouplingTransformer

Q

AT AC TERMINAL 2

DC terminal

Generator 1ac terminal

Vg

Parametersetting

Switchingconverter

gcV V Vga gb

Idc

CouplingTransformer

2V

DC terminal

Generator 2ac terminal

Vg

References

System variables

1V

Control P

2

2

>P2>Q2

P2 >Q2 <

00

00

>0P20Q2

<

0P20Q2

<<

Q

AT AC TERMINAL 1

P

1

1

>P1>Q1

P1 >Q1 <

00

00

>0P10Q1

<

0P10Q1

<<

C

Vdc

C

Vdc

+ +

P1 = -P2MVA limitMVA limit

Figure 1-24 Realization of ideal synchronous motor-generator set by two voltage-sourced converters in back-to-back connection.

The Unified Power Flow Controller scheme is shown with the conventionally used converter symbols in Figure 1-25a. In this back-to-back configuration of "Converter 1" and "Converter 2", as indicated before, the active power can freely flow in either direction between the ac terminals of the two converters, and each converter can independently generate (or absorb) reactive power at its own ac output terminal.

1-48

dc

ParameterSettings

MeasuredVariables

Control

Supplytransformer

AC+ V

Transmission Line

Converter 1

pq

VpqV+V

AC

SeriesTransformerConverter 2

I

V

V+VpqV Vpq

VRef Z Ref σ Ref or Q RefPRef

(b)

0

QP,

Q90ο 180ο

P

360270ο ρο

Po

oQ

V = 0( )pq

V = 0( )pq

ρ

(a)

Ppqmax

Qpqmax

Figure 1-25 UPFC scheme using two back-to-back voltage-sourced converters (a), variation of active and reactive line power as a function of control angle ρ (b), attainable transmitted active power and reactive line power vs. transmission angle characteristics by control (c), and circular control region for P and Q at a given transmission angle (d).

Converter 2 provides the main function of the UPFC by injecting a voltage vpq (represented by phasor Vpq) with controllable magnitude Vpq and phase angle ρ (0 ≤ ρ ≤ 2π) in series with the line via an insertion transformer. This injected voltage acts as a synchronous ac voltage source exchanging active and reactive power with the ac system as the line current flows through it. The reactive power exchanged at the ac terminal (i.e., at the terminal of the series insertion transformer) is generated internally by the converter. The active power exchanged at the ac terminal is converted into dc power which appears at the dc link as a positive or negative real power demand.

The basic function of Converter 1 is to supply or absorb the real power demanded by Converter 2 at the common dc link to support the active power exchange resulting from the series voltage injection. This dc link power demand of Converter 2 is converted back to ac by Converter 1 and coupled to the transmission line bus via a shunt-connected transformer. In addition to the real power need of Converter 2, Converter 1 can also generate or absorb controllable reactive power and thereby provide independent shunt reactive compensation for the line. It is important to note that, in contrast to the Thyristor-Controlled Phase Angle Regulator, the reactive power exchanged through the series and shunt transformers is self-sufficiently supplied or absorbed by the corresponding converters themselves, and therefore it does not have to be transmitted by the line.

It follows from the unrestricted capability of the UPFC to control the angle of the injected voltage without any restriction (0 ≤ ρ ≤ 2π) that it can emulate the operation of the conventional power flow controllers (voltage regulation, series reactive compensation and phase shifting),

1-49

simply by controlling angle ρ in specific ways. For terminal voltage regulation the injected voltage is kept in phase or anti-phase with the bus voltage, that is, angle ρ is kept 0 or π. For series reactive compensation the injected voltage is kept in quadrature with the prevailing line current by the appropriate control of angle ρ. For angle regulation using the quadrature booster scheme, angle ρ is kept at ±90o with respect to the bus voltage. For ideal phase shift, the magnitude of the injected voltage Vpq is varied with angle ρ so as to keep the magnitude of the phase-shifted voltage constant independent of the amount of phase-shift. In addition to these specific operation modes, the UPFC is also able to provide the combined control action of these specific controllers, as if each of them was individually coupled in series with the line. This could be visualized as deriving the combined compensation by determining the resultant voltage vpq, with its amplitude Vpq angle ρ, by summing all the constituent voltages providing the selected individual compensations. In phasor form, this means a simple vectorial summation, i.e., Vpq=∆Vm(voltage regulation)+Vq(series compensation)+Vσ(phase-shifting).

The multi-functional compensation capability of the UPFC is based on the independent magnitude and unrestricted angle control of the injected compensating voltage, which is also the basis for executing independent active and reactive power flow control in the line. Since this is generally the real objective of power transmission control, the conventional terms of series compensation and phase shifting become irrelevant, and the UPFC can be viewed simply with the functional objective of controlling the magnitude and phase angle of the injected voltage so as to force the magnitude and angle of the line current, with respect to a selected voltage (e.g., the receiving-end), to such values which yield the desired active and reactive power flow in the transmission line.

The variation of the active and reactive line power, P and Q with angle ρ around the steady-state transmitted power Po and reactive power Qo at a given transmission angleδo is shown in Figure 1-25b, the complete active and reactive power versus transmission angle characteristics in Figure 1-25c, and the attainable active and reactive power by the control of the injected voltage Vpq at the given steady-state transmission angle δo in Figure 1-25d. As seen in Figure 1-25b, as angle ρ is varied between 0 and 2π, Ppq(ρ) and Qpq(ρ) varies sinusoidally (with 90o phase difference) around the steady-state values of transmitted power Po and reactive power Qo obtained at the given transmission angle δo. The attainable maximum of Ppq and, similarly, that of Qpq is determined by the magnitude of the injected voltage Vpq , that is, Ppqmax= VVpq/X and Qpq= VVpq/X. Consequently, as Figure 1-25c illustrates, the transmittable active power P(δ) is controllable at any transmission angle δ (0 ≤ δ ≤ π) between

1-50

0.5

π/2

-0.5

0.0 δπ

P ( )

1.0

1.5o

P (=0.5)

Xpq

X

VV (=-0.5)

pqVV

VV

1.0

0.0

-0.5

π/2

(=0.5)

δπ

pq

X

-1.0

-1.5

-2.0

- Q

(=-0.5) X

pqVV

(c)

+≤≤− δδδ +≤≤− δδδ

Q

ρ

P

= 0δ o

P = max

δ o= 30

δ

VVpq

X

P ( ) o 30o

Q ( ) o 30o

o

P + o pqmaxP π/6

30o

π/6

Q ( )o 30o

-0.12

Boundary: Control area

V = 0pq

(d)X

pq(VVpqP 2

pqQ 2+ = )2

2

(d)

XVV

PPX

VVP pq

opq

o +≤≤− )()()( δδδ

and the reactive power Q(δ) is between

XVV

QQX

VVQ pq

opq

o +≤≤− )()()( δδδ

The variation of Ppq(ρ) and Qpq(ρ) with angle ρ shown in Figure 1-25b defines a circular region whose center is defined by the coordinates of the steady-state active and reactive power Po(δo ) and Qo(δo), that is,

( ) ( ) ( ) ( )

2max22

⎭⎬⎫

⎩⎨⎧

=−+−X

VVQQPP pq

oooo δρδρ

This Ppq,Qpq circular control region is shown around the steady-state Po(δo ) and Qo(δo) coordinate values at the assumed transmission angle of δo=30o in Figure 1-25d. The control of P and Q within the region can be executed independently within the restriction of Ppq

2 + Qpq

2 = (VVpq/X)2. Note that the Ppq,Qpq control region remains the same at any value of the transmission angle δ.

1-51

The unique capabilities of the UPFC summarized above provide powerful, hitherto unavailable meansto maintain or vary the real and reactive power flow in the line to satisfy load demand and improve system stability with optimized asset utilization. The Interline Power Flow Controller represents a novel concept with the objective of providing a flexible power flow control scheme for a multi-line power system, in which two (or more) lines would require series reactive compensation. The IPFC scheme provides, together with independent SSSC type controllable series reactive compensation for each line, a capability to transfer active power between the compensated lines. This capability makes it possible to equalize both real and reactive power flow between the lines, to transfer power demand from overloaded to underloaded lines, to compensate against resistive line voltage drops and the corresponding reactive line power, and to increase the effectiveness of the compensating system for dynamic disturbances (transient stability and power oscillation damping). In general, the IPFC provides a highly effective scheme for power transmission management at a multi-line substation.

The basic Interline Power Flow Controller scheme, also using of two back-to-back converters, is shown in Figure 1-26a. Each converter is coupled to a different transmission line via its own series insertion transformer and is able to provide independent series reactive compensation to its own line. (Note that the two lines can be totally different in voltage, transmission capacity and loading.)

dc

ParameterSettings

MeasuredVariables 1 Control

AC+ V

Transmission line 2

Converter 1

AC

SeriesTransformer 2Converter 2

I V +Vpq2V Vpq2

QRefP2Ref

(a)

Transmission line 1

2

I1

2 2

Vpq1V1 V + Vpq11

SeriesTransformer 1

P1RefPrimary line

selector

MeasuredVariables 2

Figure 1-26a Basic IPFC scheme for two lines (a), and active and reactive power vs. transmission angle characteristics attainable for the two lines (b).

Each converter produces controllable ac output voltage at the fundamental frequency, which is synchronized to the voltage of the transmission line which that converter controls. The injected voltage in each line generally has one component that is in quadrature and another that is in

1-52

phase with the relevant line current. The quadrature component provides series reactive compensation for the line to control active power flow, and the in-phase component defines the active power exchanged with the line to control the reactive flow. Since each converter is self-sufficient in generating or absorbing reactive power, the quadrature voltage components in each line can be independently varied to control the active power transmitted over the line. However, since the active power exchanged by a converter at its ac terminal has to be supplied to, or absorbed from its dc terminal, the in-phase output voltage component of each of the two converters must be controlled so as to ensure a net zero active power balance at their common dc terminals. In other words, the active power compensation demand of one line must be fully supplied (or absorbed) by the other line. This restriction means that changing the reactive power in one direction in one line (e.g., decreasing it), it will have to be changed in the other direction in the other (e.g., increasing it). In other words, the reactive power flow can be optimized in one (the weaker or overloaded) line by transferring the burden of reactive power flow to the other (stronger or underloaded) line. From the optimized, prime line viewpoint, the IPFC provides the same full, active and reactive power flow control that characterizes the UPFC. The active power flow in the other line can be controlled by reactive compensation in the same way that characterizes the SSSC. However, the reactive power flow in this line, in addition to the prevailing line load, will be determined by the active power compensation demand of the other line due to the fundamental operating condition that the active power exchanged by the two converters must sum to zero, that is, P1pq = - P2pq. The practical effect of this constraint is illustrated by the active and reactive power versus transmission angle characteristic of the IPFC at its two ac terminals in Figure 1-26b. For this illustration it is assumed that the two lines are identical and thus the IPFC with two identically rated converters has the same control range of power flow control for each line. It can be observed in the figure that the active power flow control range is the same at both terminals, i.e., the reactive compensation can be carried out on each line independently either to increase or decrease the active power flow P(δ). However, the reactive power flow control ranges are mirror images of each other for the two lines. That is, if active power is supplied to one line in order to decrease reactive power flow, then the same amount of active power must be absorb from the other line which will increase the reactive power flow in that line. This requirement is illustrated in the Q(δ) versus δ characteristic plots by showing the attainable control areas of Q1pq(δ) and Q2pq(δ) above and below the Qo(δ) curve in different colors as mirror images of each other at the two ac terminals of the IPFC.

1-53

(b)

(=-0.5)

0.5

π/2

-0.5

0.0 δπ

P ( )

1.0

1.5 δo1

P(=0.5)X(=-0.5)

pq1V V1

Xpq1V V1

Xpq1(V V

pq1P 2pq1Q 2+ = )21

1

1

1

1.0

0.0

-0.5

π/2

(=0.5)

δπ

Q ( )o1

-1.0

-1.5δ

-2.0

-Q

Xpq1V V1

Xpq1V V1

1

1

1

1.0

0.0

-0.5

π/2

(=0.5)

δπ

Q ( )o2

-1.0

-1.5δ

-2.0

-Q

Xpq2V V2

Xpq2V V2

2

2

2

0.5

π/2

-0.5

0.0 δπ

P ( )

1.0

1.5 δo2

P(=0.5)X(=-0.5)

pq2V V2

Xpq2V V2

Xpq2(V V

pq2P 2pq2Q 2+ = )2

2

2

2

2

For Line 1 For Line 2

(=-0.5)

2 2

Figure 1-26b Basic IPFC scheme for two lines (a), and active and reactive power vs. transmission angle characteristics attainable for the two lines (b).

Figure 1-26b clearly show the added flexibility the IPFC configuration provides for series compensation: it is able to control not only the active power flow in a two- or multi-line transmission system, but also can equalize or control the reactive power flow in the lines. The IPFC provides an excellent tool to solve economically power flow problems in a multi-line transmission system in which the actual power flows are not proportional to the capacities of the corresponding lines or to their desired power transmissions, or in which the desired real power transmission in some lines is hindered by relatively high reactive power flow.

It is evident from the above discussion that the IPFC concept, characterized so far only for two lines, can be extended to multiple (n) lines as illustrated in Figure 1-27a. The underlying idea of this generalized IPFC approach is that the strong or under-loaded lines are forced to help the weaker or over-loaded lines in order to optimize the utilization of the whole transmission system. The operation of a multi-line IPFC requires that the sum of the active power exchanged by the total number of converters must be zero. However, the distribution of the positive and negative power exchanged with the individual lines by the individual converters within this overall constraint is arbitrary (as long as the Mva rating of the individual converters is not exceeded). This arrangement would provide a UPFC-type two-dimensional compensation for some lines and an SSSC-type reactive compensation for the others.

1-54

Line n

Line 2

Line 1

DC bus

+ - + - + - + - + - + -

DC bus

+ -

Line n

Line 2

Line 1

(b)(a)

SSSC 1 SSSC 2 SSSC n SSSC 1 SSSC 2 SSSC nSTATCOM

Figure 1-27 Multi-line line IPFC scheme (a), and generalized multi-line IPFC with line independent active power support and bus voltage regulation (b).

The constraint for keeping the sum of the active power exchanged with the n-lines zero can be circumvented by adding a shunt-connected converter to the multi-converter IPFC as illustrated in Figure 1-27. This arrangement is particularly attractive in those cases in which the active power compensation requirement of the "weak" lines exceeds the real power that can be absorb from the "strong" lines without appreciably impacting their own power transmission or when shunt reactive compensation at the substation is required anyway for voltage support.

The Back-to-Back Tie scheme is shown with the conventionally used converter symbols in Figure 1-28. Through the DC link, the tie between AC System 1 and AC System 2 can be asynchronous with large and a variable angle difference or even with different system frequencies. It can also be synchronous, in which case the tie is able to provide synchronizing torque in either direction. Generally, from the viewpoint of either system, the tie behaves like a perfect synchronous generator, providing rapid control for active power with effective current limitations, and also executing bus-voltage regulation by fast reactive compensation. This generator-like characteristic is illustrated at both terminals of the tie in Figure 1-28 by two perfect four-quadrant circular Q versus P plots, within which any combination of simultaneous active and reactive power output is achievable at each side up to the limit determined by the rating of the equipment: P2+Q2≤MVArating. Of course, the fundamental operating requirement, P1+ P2=0, i.e., P1= -P2, must be maintained at any MVA output. In actual operation, generally the required active power transmittal setting would have the priority, and with the maintenance of this, the control would regulate the bus voltages at the two sides independently by the adjustment of the reactive output according to the V-I regulation characteristics shown in Figure 1-28.

1-55

ParameterSettings

MeasuredVariables Control

CouplingTransformer 1

AC SYSTEM 1

Converter 1 Converter 2

V2

V1Ref PRef

dcAC

+ VAC

Q

At AC System 2 bus

P

2

2

>P2>Q2

P2 >Q2 <

00

00

>0P20Q2

<

0P20Q2

<<

MVA limit

CouplingTransformer 2Q

At AC System 1bus

P

1

1

>P2>Q2

P2 >Q2 <

00

00

>0P20Q2

<

0P20Q2

<<

MVA limit

V1 AC SYSTEM 2

P1 = -P2

V2Ref

C2max

C2

I

I

0

V2

L2maxIL2I1

C1max

C1

I

I

0

V1

1LmaxIL1I1

Figure 1-28 Asynchronous/synchronous tie scheme using two back-to-back voltage-sourced converters, and the corresponding power transmission and voltage regulation terminal characteristics.

Summary on Power Electronics-Based Transmission Compensators and Controllers

1. There are two approaches to the realization of power electronics-based compensators and controllers: One employs thyristor-switched capacitors and reactors, and tap-changing transformers, the other voltage-sourced converters as synchronous voltage sources. The first approach has resulted in the Static Var Compensator (SVC), the Thyristor-Controlled Series Capacitor (tcsc), and the Thyristor-Controlled Phase Shifter. The second approach has produced the Static Synchronous Compensator (STATCOM), the Static Synchronous Series Compensator (SSSC), the Unified Power Flow Controller (UPFC), the Interline Power Flow Controller (IPFC), and the Back-to-Back Tie (BtB).

2. The two groups of Controllers operate in distinctly different manners:

• The thyristor-controlled group employs capacitor and reactor banks with thyristor valves in traditional shunt or series circuit arrangements. The thyristor valves control the on and off periods of the fixed capacitor and reactor banks realizing a variable reactive (shunt or series) impedance.

• The voltage-sourced converter-based group, using power semiconductors with gate turn-off capability (GTO, GCT, IGBT, etc.), can emulate a synchronous voltage source that internally generate reactive power for transmission line compensation, without the use of ac capacitors or reactors. The converter, if supported by a dc power supply or energy storage device, can also exchange real power with the ac system.

1-56

3. The different operations result in different application characteristics:

• Shunt compensators. The maximum attainable maximum capacitive compensating current of the SVC, operating as a variable reactive admittance, decreases linearly with the system voltage. The STATCOM, functioning as a synchronous voltage source, exhibits a superior V-I characteristic by being able to maintain the maximum capacitive output current at any system voltage down to zero.

• Series compensators. The compensating voltage provided by the Thyristor-Switched/Controlled Series Capacitor is a function of the line current. The compensating voltage of the SSSC, as a reactive voltage source, can be maintained at any value independent of the line current, or can be varied with the line current to emulate a compensating capacitor. Moreover, the SSSC can inherently provide capacitive as well as inductive compensation, which makes it highly effective in power flow control, as well as in power oscillation damping.

• Phase shifters. Thyristor-Controlled Phase-Shifters based-on tap-changer transformers draw both the active and reactive power from the ac system they exchange through series voltage injection in executing angle control. The back-to-back converter arrangement, emulating the operation of an ideal motor-generator set, internally generates or absorbs all the reactive power the phase shifting requires. This minimizes line voltage variation and losses.

4. The back-to-back arrangement of two voltage-sourced converters emulates the operating characteristics of an ideal motor-generator set with four-quadrant operation in the reactive versus active power plane. The use of this versatile arrangement with shunt and series, series and series, and shunt and shunt coupling results in three different and extremely powerful transmission controllers: the UPFC, IPFC, and BtB Tie.

• The UPFC with a shunt and a series coupled converter can provide concurrent or individual voltage, impedance, and angle regulation or, alternatively, independent real and reactive power flow control. With this operating flexibility it can readily adapt to particular short term contingencies or future system modifications.

• The IPFC with two or more back-to-back connected converters, each coupled in series with a different line, provides means to transfer real power among lines, and execute independently controllable reactive compensation for each line. This capability makes it possible to equalize power flow among lines or manage power flow according to demand and line capacity. In general, the IPFC provides a highly effective scheme for the optimized use of a multi-line transmission system.

• The two voltage-sourced converters in back-to-back connection with shunt coupling at each side provides a perfect arrangement for tying two power system asynchronously or synchronously with fully controllable power transfer and terminal voltage regulation at both sides.

1.5 Issues and Options of Future Transmission Control

The majority of installed power electronics-based (FACTS) Controllers are conventional thyristor-controlled Static Var Compensators (SVCs) and a much smaller number of Thyristor-Controlled Series Capacitors (TCSCs). Since the 1990s, there has also been installed an appreciable number of STATCOMs and also some prototype installations of SSSC and multi-

1-57

function power flow controllers, UPFC, IPFC, and BtB. In contrast to these newly developed converter-based Controllers, Thyristor-Controlled Var Compensator, Series Capacitors and Phase Shifters, do not have multi-functional capability. They can execute only the single function (i.e., shunt compensation, series compensation, or angle control) for which they were installed. They may, however, have the limited capability to change their operating mode, e.g., the SVC may be set up to regulate voltage or reactive power, the TCSC to control line impedance or power flow, etc. The single function operating capability is also true for the STATCOM and SSSC with the significant difference that these can be expanded to multi-function transmission control capability by the addition of one or more converters in back-to-back configuration. The UPFC and IPFC are generally characterized by multi-functional transmission control capability, e.g., simultaneous voltage regulation, active and reactive power flow control, as well as by operational convertibility, e.g., to be operated as independent STATCOM and SSSC, IPFC or UPFC or even as BtB Tie. Moreover, within any selected operating mode, they offer a number of selectable control functions, e.g., transmission angle and impedance or direct active and reactive power control.

The customary and still prevailing utility practice has been to install a dedicated transmission Controller and operate it in a fixed mode with substantially steady-state reference signals controlling local system variables (voltage, current, impedance etc.). In other words, the Controller usually receives no other indication of system problems and contingencies than the change in some locally sensed variables. Its response to the received indication is fast and automatic, trying to correct the detected system discrepancy. However, this may not be the needed, or even a compatible, response to the problem under a possibly different post fault conditions and system parameters (e.g., system impedance). Moreover, although some SVC installations can initiate the operation of mechanically switched capacitor banks for steady–state compensation, the Controllers would generally have no direct information of the actions of other controlled power system devices. In fact, it could preclude the operation of those with slower response, or interact with those of comparable speed. There are possible control strategies measures and protection schemes to provide an acceptable steady-state solution to these problems. However, in order to get predictable and proper dynamic response, as well as maximal steady-state benefits, the direct coordination of fast Controllers and other devices, located close enough to interact, in addition to real time information on prevailing system conditions, will not only be operationally advantageous, but absolutely necessary with the ever growing number, and increasingly powerful transmission Controllers deployed.

Beyond the interaction prompted coordination and real time system information, the anticipated deployment of ever increasing number transmission Controllers, particularly the converter-types with multiple functional capabilities, provides a new basis for re-conceptualizing the control of the bulk power transmission system as a dynamic entity. Within this entity, power electronics-based Controllers could be employed to establish optimal power flow patterns for economic benefits consistent with network security requirements under prevailing system conditions and handle dynamic disturbances by concerted control action.

From the technical standpoint the most plausible system control structure to manage the transmission grid as a dynamic entity is hierarchical, built up with layers of controls from the level of each individual Controller ultimately to the top level central control that strategically coordinates the operation of the overall network. This hierarchical control, illustrated in Figure 1-29, could be envisioned in several ways. One extreme would be to mandate that the top level central controller direct the operation of each operational Controller in real time. The opposite

1-58

extreme would be to operate the Controllers independently from local system data and providing coordination from local system operators in the form of steady-state reference signals (i.e., the method utilities presently use). The first extreme would clearly raise serious questions regarding the security of the total system since errors in the data collection and processing, malfunctions in the central control or in the communication system, could result in major and potentially system wide outages. The second extreme would not completely utilize the inherent control and operating flexibility of the power electronics-based Controllers of the system for full economic or security benefits. A reasonable solution may be similar to that used in organizational managerial structures in which the overall strategic decisions are made at the top level, the tactical decisions at mid-level and the implementation details are executed at local levels. In this scenario one could visualize that the central control would determine the main transmission paths for the overall system according to demand and economic objectives, taking into account equipment availability and other constraining factors, and would coordinate the operation of neighboring transmission areas. The area controls would optimize the defined power flow through their system by setting, via appropriate communication links, their Controllers into the desired operating mode and by providing reference signals for them for the selected system variables controlled. The area control would supervise the operation of the area transmission system by collecting system status data as well as critical operating and availability information from each Controller. Under area contingency or system flow change demand, the area control would re-optimize its system operation by appropriately changing the functional operating mode of the Controllers and their reference inputs. In the case of an area disturbance, the area control would dynamically change the reference inputs to the appropriate Controllers in order to achieve a coordinated counter action to reestablish rapidly rotational and voltage stability. Within this scheme, the individual Controller would operate self-sufficiently from locally measured system variables and its output would be controlled, in the operating mode defined, by the reference input the area control system provides. In the case of an area control outage or malfunction, the local Controllers would automatically fall back to independent operation determined by the prevailing system steady-state references. Of course, under all conditions, the individual Controllers would accept overriding instruction from authorized operating personnel.

1-59

Intertie

ControlCentral

Control 1Area

Control 2Area

Control nArea

Area System 1with FACTSControllers Area System 2

with FACTSControllers

Area System nwith FACTSControllers

Figure 1-29 Conceptual hierarchical control scheme for interconnected power systems using power electronics-based transmission controllers.

The establishment of a viable hierarchical system control for a large transmission grid would be a large undertaking. It would require fundamental changes in control strategies, development of new system security procedures and wide-area measurement capabilities together with highly reliable communication links and protocol. It would also require the development of analytical (software) tools to process real time information for system security, voltage and rotational stability assessments.

In addition to the many challenging technical issues, the hierarchical control of power electronics-based power transmission also poses ownership and responsibility problems for the still evolving utility system structure. There could be many questions posed and many different answers offered, based on different transmission network and ultimate organizational structures assumed. It would be speculative to pursue these here. However, the incentives are clearly established: the broad application of power electronics-based Controllers with other appropriate advanced technologies promises to facilitate competitive electric power system with modern, highly flexile transmission network that fully utilizes system assets while maintaining the required security and reliability.

References

[1] Miller, T.J., Editor, “Reactive Power Control In Electric Systems,” Chapter 2, John Wiley & Sons, 1982.

[2] Gyugyi, L., “Power Electronics in Electric Utilities: Static Var Compensators,” Proceedings of the IEEE, Vol. 76, No. 4, April 1988.

1-60

[3] Hingorani, N.G., ”High Power Electronics and Flexible AC Transmission System,” IEEE Power Engineering Review, pp. 3-4,July, 1988.

[4] Gyugyi, L., “Solid-State Control of Electric Power in AC Transmission Systems,” International Symposium on “Electric Energy Conversion in Power Systems,” Invited paper, No. T-IP. 4, Capri, Italy, 1989.

[5] Gyugyi, L., “A Unified Power Flow Control Concept for Flexible AC Transmission Systems,” IEE Fifth International Conference on AC and DC Power transmission, London, Conference Publication No. 345, pp 19-26, 1991, also IEE PROCEEDINGS-C, Vol. 139, No. , July 1992.

[6] Gyugyi, L., “Dynamic Compensation of AC Transmission Lines by Solid-State Synchronous Voltage Sources,” IEEE Transactions on Power Delivery, Vol. 9, No. 2, April 1994.

[7] Gyugyi, L., et al., “Static Synchronous Series Compensator: A Solid-State Approach to the Series Compensation of Transmission Lines,” IEEE Transactions on Power Delivery, Vol. 12, No. 1. January 1997.

[8] Edris, A., et al., “Controlling the Flow of Real and Reactive Power,” IEEE Computer Applications in Power, January 1998.

[9] Gyugyi, L., et al., “The Interline Power flow Controller Concept: A New Approach to Power Flow Management in Transmission Systems,” IEEE/PES Summer Meeting, Paper No. PE-316-PWRD-0-07-1998, San Diego, July 1998.

[10] Song, Y.H., and Johns, A.T., Editors, “Flexible ac transmission systems (FACTS),” Chapter 1, The Institution of Electrical Engineers, 1999.

[11] Hingorani, N.G., and Gyugyi, L., “Understanding FACTS,” The Institute of Electrical and Electronics Engineers, Inc., 2000.

2-1

2 POWER SEMICONDUCTORS AND VALVES (DRAFT)

2.1 Basics of Semiconductors

2.1.1 Simplified Physics

To facilitate understanding of the various semiconductors, it is advantageous to look at some physics. While the theory of semiconductor physics including energy bands and quantum mechanics is rather complex, for the purpose of this reference book it is advantageous to describe the mechanism of conduction in semiconductors in a simplified manner. For more detailed information reference is made to [1].

The precondition of current flow in any material is the existence of charge carriers, for example electrons, which are free to move. If a voltage is applied to the material, these charge carriers move in the direction of the electrical field (or in opposite direction depending on the polarity of the charge).

Today, all commercially available power semiconductors used for high power applications are made from pure mono-crystalline silicon. Silicon is in the fourth column of the periodic table of elements, i.e. the silicon atom has four electrons. In a silicon crystal the atoms are bonded in a regular structure; each atom has four neighbors and two adjacent atoms share two electrons. The three-dimensional setup can be illustrated in a two-dimensional sketch as in Figure 2-1.

Silicon atom

Electron

+

+ Free electrons

Ionized silicon atom

a) T = 0K b) T > 0K

Figure 2-1 Two-dimensional Illustration of a Silicon "Lattice" Structure

At absolute zero temperature, T = 0 K, no charge carriers are available for current flow since all electrons are bonded, all atoms are neutral (Figure 2a). In that case the semiconductor behaves like an insulator. Thermal energy or light can break the bonds so that electrons leave the structure and are free to move, leaving holes. The remaining atoms then have a positive charge (Figure 2-1b).

2-2

Any free electron can return into a hole so that the associated atom becomes neutral again. When an electron leaves the structure (generation of charge carriers) and at the same time another electron falls back into a hole (recombination) of a different atom, this appears as movement of the positive charge or can be interpreted as a movement of a hole with positive charge. It has become customary in semiconductor terminology to think of holes as positive charges. When an electrical field is applied, two types of charge carriers are available for current flow: electrons with negative charge and holes with positive charge. Since the density of charge carriers increases with temperature the conductivity increases also. This kind of conductivity is called intrinsic conductivity.

Generally, power semiconductors can be in one of two states: the blocking or off state and the on state. In the off state the main current flow is blocked (equivalent to an open switch), only a small so-called off state current flows through the device as given by its internal resistance. In the on state (equivalent to a closed switch), the semiconductor is conducting the current; a small voltage drop of a few volts appears across the main connectors of the semiconductor. In the two states described above, losses (on state losses and off state losses) are produced because a current flows through the device and a voltage across the device occurs at the same time. Losses are also produced in the transition between on state and off state; these losses are so-called switching losses. They are caused by an overlap of high voltages across and high currents through the device. The heat that is produced by losses has to be dissipated by heat sinks and a cooling system.

2.1.2 Processes in Manufacturing

The various processes in the manufacturing of semiconductors aim at modifying the conductivity of the base material to create specific characteristics of the semiconductor. In a process called doping, some atoms within the lattice structure of the pure silicon crystal can be replaced with different atoms, for example with atoms of the third or fifth column of the periodic table of elements.

In case of doping with atoms of the fifth column, like phosphorus, an excess electron is available at the associated atom, which can easily leave the structure and thus becomes available for current flow. This kind of doping is called n-doping. On the other hand, p-doping can be achieved with elements of the third column of the periodic table of elements. In that case, an electron is missing in the structure, thus a hole is available for current flow. An element used for p-doping, for example, is boron.

Doping can be achieved by diffusing foreign atoms into the high purity silicon wafer (chip) in a high temperature process, creating doped layers of defined shape and depth. When a p-doped layer abuts on a n-layer, a so-called pn junction is created. If a voltage is applied to the pn junction, the behavior of that pn junction depends on the voltage polarity: a positive potential at the p-layer makes the pn junction conducting, a positive potential at the n-layer results in blocking of the pn junction for current flow.

Ultra-pure silicon is the precondition for the production of modern high power semiconductors. To achieve the desired behavior of the devices, defined structures of n-doped and p-doped zones are realized within the silicon wafer.

2-3

Starting with homogenous n-doped silicon, different methods are available for doping, e.g. diffusion, epitaxial growth, and ion implantation. Electron irradiation and in some cases implantation of ions like helium ions are used to fine tune the behavior of the devices with respect to switching performance and on state characteristic.

So-called back-end processes include beveling, alloying, soldering, packaging, and final testing to complete the manufacturing processes.

2.2 Basic Power Semiconductor Switch Types and Their Characteristics

2.2.1 No Turn-on or Turn-off Control: Diodes

Diodes are the simplest semiconductor devices. The silicon wafer has two layers, one p-doped and the other n-doped, figure 2-2. It has an electrical contact at the p-doped layer termed anode (A) and a contact at the n-doped layer termed cathode (C). When a positive voltage is applied at the anode (forward bias), both the holes in the p-doped layer and the electrons in the n-doped layer move towards the pn-junction, flooding this area with free charges. As a result, a forward current IF is flowing through the device, the diode is conducting.

Figure 2-2 Device Symbol and Structure of Diode

When a negative voltage is applied to the anode, however, both the holes and the electrons move away from the junction, depleting the junction from charge carriers. A space charge region is created without sufficient charge carriers for current conduction and the diode blocks the voltage. Only a small reverse current IR is flowing that is defined by the high ohmic resistance of the semiconductor material. The resulting voltage/current characteristic of the device is illustrated in figure 2-3 a).

Since diodes cannot be controlled, their operation mode is given at any instant by the surrounding circuit. When an ac voltage is applied between anode and cathode, the diode will be conducting during the positive half wave and will block during the negative half wave.

p - doped layer

n - doped layerspace charge

reverse bias

IR

A

C

p - doped layer

n - doped layer

forward bias

IF

C

A

pn-junction

A

C

iD

VAC

2-4

Figure 2-3 Characteristics of Diodes

The transition from on state to the blocking state is not ideal. When diodes are turned off, the pn-junction is still flooded with carriers. The reverse bias moves those carriers away from the junction so that the current becomes negative until a peak is reached and then the reverse current decays as shown in figure 2-3b. The diode gets a reverse blocking capability when the peak reverse current is reached. The negative current is called reverse recovery current; the peak value is named peak reverse recovery current, and the area between the t-axis and the negative current is called the recovery charge, which depends mainly on the rate-of-rise of current and the junction temperature. Special attention has to be paid to that turn-off behavior since it creates losses (turn-off losses) and results in over-voltage peaks. Rectifier diodes are usually turned off with rates-of-rise of current of some amperes per microsecond whereas freewheeling diodes as used in voltage sourced converters are turned off with up to several thousand amperes per microsecond.

As per 2006, high power rectifier diodes are available with blocking voltages up to approximately 10 kV and freewheeling diodes up to 6.5 kV. The current carrying capability of discrete devices is in the range of some hundred amperes to several thousand amperes, depending on the blocking voltage. While rectifier diodes have not found application in power electronics based transmission controllers, freewheeling diodes are an essential component in voltage sourced converters.

2.2.2 Turn-on Control only: Thyristors

A thyristor is a four-layer device with three contacts: Anode (A), Cathode (C), and Gate (G), figure 2-4. When the thyristor is reverse biased, i.e. plus is connected to the cathode and minus to the anode, the two pn-junctions J1 and J3 are reverse biased so that no injection of carriers occurs and no current flow is established. There is only a very small reverse current as long as the voltage remains below the avalanche and punch-through limits. When the thyristor is forward biased (i.e. plus at the anode) and the gate connection is open, the two outer pn-junctions are forward biased while the central pn-junction J2 is reverse biased and a space charge region is developed. The thyristor is said to be in the off-state. Only a small reverse current flows through junction J2, which can be viewed as the off-state leakage current at low forward bias.

vAC

iD

a) Static Characteristic

t

iD

b) Dynamic Characteristic

2-5

Figure 2-4 Device Symbol and Structure of Thyristor

From the off state, the thyristor can be turned on by injecting a positive current pulse into the gate, i.e. the p-base. This results in flooding the forward blocking junction J2 with charge carriers and current flow from anode to cathode is then established. Once turned on, if the current remains sufficiently large as defined by the external circuit, the thyristor will stay on even without a gate current. It is thus a latching device which remains in conduction as long as the current through it exceeds a critical value termed the holding current. A thyristor that is turned on by an electrical gate pulse is called an Electrically Triggered Thyristor (ETT).

Alternatively, a light pulse may be applied into the forward blocking junction J2 while a positive anode-to-cathode voltage is applied to the device. Carriers are then generated, which – when multiplied in cascading amplifying structures integrated in the silicon wafer – can turn the device on. A thyristor that is turned on by an optical gate pulse is called a Light Triggered Thyristor (LTT).

Once a thyristor has turned on, it behaves like a diode, i.e. it remains in the on state until the load current is brought to zero in the surrounding circuit. Usually ETTs and LTTs are different only in the gate design, the rest of the design can be identical. To turn on an ETT, an electrical gate current of some amperes has to be applied; to turn on a LTT, an infrared light pulse of some ten milliwatts has to be applied to the gate of the thyristor through a light guide. Thyristors in the high power range have one or more amplifying gate stages, which are integrated in the silicon wafer. With the amplifying gate stages and a distributed gate design, a fast spread of the on state current across the whole cathode area can be realized after triggering, which leads to a safe turn-on behavior, i.e. the capability to sustain the required rates of rise of on state current (usually some hundred amperes per microsecond). Figure 2-5 illustrates the turn-on process and shows a thyristor pellet with a typical interdigitated gate structure.

A

C

iT

vAC

GiG

C G

A

High doped nLow doped pLow doped nHigh doped p

pn+

n-

p+J1

J2

J3

2-6

Figure 2-5 Turn-on Process and Pellet of Thyristor

Thyristors in line commutated converters have a symmetric blocking capability, i.e. they can block the voltage in both directions across anode and cathode. The turn-off behavior of thyristors is quite similar to that of diodes, i.e. a reverse current flows through the device until the peak reverse recovery current is reached and the reverse current decays. After turn-off, the thyristor is not capable to block a forward voltage for a time, which is called turn-off time. The turn-off time is usually in the range of some hundred microseconds, see also 2.3.2.

2.2.3 Turn-on and Turn-off Control: Gate Turn-off (GTO) Thyristors and Various Transistors

Semiconductors with turn-on and turn-off control are the essential electronic switches for voltage sourced FACTS controllers. While the GTO has quite a reputation as a power device with turn-off capability, transistors are often associated with the microchip and computer industry (logic level transistors). However, there are high power transistors that have revolutionized the variable speed drives technology and have also been successfully used in power transmission applications. As GTOs and transistors are based on different control principles, they will be discussed in separate subsections.

tiG

vAC

iT

t

a) Anode-Cathode Voltage,Thyristor Current, and Gate Pulse

b) Thyristor Pellet With TypicalInterdigitated Gate Structure

2-7

Gate Turn-off Thyristors (GTO)

The GTO is a thyristor that permits the extraction of positive charges from the p-layer by injecting a high negative current into the gate. As a result, at the forward blocking junction J2 (see figure 2-4) a space charge region develops and the device switches to the off state. During this process, lateral potential differences within junction J3 have to be avoided. Therefore the cathode is configured as an array of small segments (typically 3mm by 0.2mm) that are arranged in concentric rings around the device center and that are each surrounded by the metallized surface of the p-layer serving as the gate, see figure 2-6.

Figure 2-6 Device Symbol and Structure of a GTO

Like any thyristor, the GTO is turned on by feeding a positive triggering current pulse of some Amperes into the gate. It latches when the holding current is exceeded, so the duration of the pulse need only be in the order of 10µs. However, optimization of the turn-off properties results in relatively high forward voltage drop during conduction. This may be counteracted by feeding a positive gate current continuously during the on state. The turn-on time typically is a maximum of 10µs.

The negative current pulse for turn-off is defined by the turn-off current gain (ratio of load current over gate current), which typically is three to five. The turn-off time is in the order of 30 to 40µs. During that time the rate-of-rise of the recovery voltage needs to be controlled, which requires the use of snubber circuits.

The performance of the GTO can be improved when the turn-off gain is reduced to one, i.e. the peak gate current equals the load current, and the peak is reached in less than one microsecond. In this case, manufacturers' data sheets state a peak gate power of 120kW for a 4000A device. Such devices have been developed successfully and have been available on the market for some years under the name GCT (Gate Commutated Thyristor). Considering the maximum permissible negative gate voltage of 20V, the steep pulse is achieved by designing an extremely low-

A

C

G

np

n

p

Gate (G) Cathode (C)

Anode (A) metallized contact

Cathode segment

Gate

a) Device Symbol b) Pellet Pattern c) Basic Segment Structure

2-8

inductance circuit, by a special design of the gate leads within the housing of the power semiconductor, and by integrating the GTO into the gate drive, figure 2-7. The combination of GCT and gate drive is referred to as IGCT (Integrated Gate Commutated Thyristor). The turn-off time of the GCT is around 10µs.

Photograph courtesy ABB Semiconductors

Figure 2-7 Integrated Gate Commutated Thyristor Showing the Power Semiconductor and Gate Drive (Cover Removed), Including an Array of Energy Storage Capacitors and MOSFET Switches

Both the GTO and the GCT require limitation of the rate-of-rise of current at turn-on, usually by a series reactor. However, the snubber circuit required by the GTO to limit rate-of-rise of voltage during turn-off is optional for the GCT and can be omitted in exchange for a somewhat reduced turn-off current capability. Both the GTO and the GCT for use in voltage sourced converters are asymmetric devices; they are able to block forward voltage only and have to be protected from negative voltage by a diode connected in antiparallel. In some designs, this diode is integrated onto the same wafer.

There are a number of other types of turn-off thyristors, e.g. the Static Induction Thyristor (SITh) – also known as Field Controlled Thyristor (FCT), the MOS Controlled Thyristor (MCT), or the Emitter Turn-Off Thyristor (ETO). While the SITh does not seem to have made it into production, the MCT was claimed to have become commercially available [3]. It is basically a thyristor with two MOSFETs integrated into the gate structure, one for turn-on and one for turn-off. The ETO is a hybrid device, combining a GTO with a number of discrete high current, low voltage MOSFETs that are arranged around the GTO but connected in series to the cathode and in parallel to the gate. They are reported to show similar turn-off performance as the IGCT [10]. As with the MCT, their advantage would be that they are voltage controlled as an IGBT and therefore require small control power only.

However, as per the year 2006 neither the MCT nor the ETO appear to have been used in any power electronics based transmission controller.

2-9

Power Transistors

There are many types of power transistors, all based on a three-layer structure of the silicon chip —either of npn type or pnp type. Their basic function can be thought of as that of a controllable resistor: if the resistance is zero, the transistor can conduct current; if it is infinite, the transistor is off. In contrast to thyristors, the transistor therefore does not normally latch into conduction and the control power to modulate the resistance is very low. For power applications, two categories are of interest: the Bipolar Junction Transistor (BJT) and the Field Effect Transistor (FET) which differ in the type of control to modulate the resistance and thus the device current.

The device symbol and a simplified cross section of the BJT in npn configuration are shown in figure 2-8. BJT terminology is different to that of thyristors: the three terminals are named Collector (C), Base (B), and Emitter (E).

Figure 2-8 Device Symbol and Simplified Schematic Cross Section of a Bipolar Junction Transistor in npn Configuration

The base is physically located between the emitter and the collector and is made from lightly doped, high resistivity material. It is typically very narrow. A small current injected through the junction between the base and the emitter changes the properties of the base-collector junction so that it can conduct current even though it is reverse biased. This creates a much larger current between the collector and emitter, controlled by the base-emitter current by way of the charges it injects into the junction..

The FET is a unipolar device and functions on the principle that semiconductor conductivity can be increased or decreased by the presence of an electric field. An electric field can increase the number of free electrons and holes in a semiconductor, thereby changing its conductivity. The most common type of FET is the Metal-Oxide-Semiconductor FET (MOSFET), figure 2-9.

In this device, when a positive voltage is applied between the gate and the body/substrate, the capacitor formed by the metal contact plate and the p-layer, with the silicon dioxide as a dielectric in between, results in an electric field underneath the oxide that converts the silicon surface into an n-type channel between the two n-layers. When a voltage is applied between source and drain, a current can flow that is controlled by the gate voltage influencing the width of the channel. As the device is voltage-controlled, its advantage is the very low control power along with higher switching speed. Disadvantages against bipolar transistors include a high resistance in the on state and larger chip area, especially for higher voltage ratings.

Collector

Emitter

iC

vCE

Basen

n

p

E B C

2-10

Figure 2-9 Device Symbol (a) and Simplified Schematic Cross Section (b) of a Metal-Oxide-Semiconductor Field Effect Transistor (MOSFET) in npn Configuration (n-Channel MOSFET)

A monolithical combination of the bipolar junction transistor and the field effect transistor on the same silicon chip is the Insulated Gate Bipolar Transistor (IGBT). It can be regarded as a MOSFET controlled bipolar transistor. Figure 2-10 shows the device symbol and the static i-v characteristic.

Figure 2-10 Device Symbol and Static Voltage –Current Characteristic of an IGBT

The gate design of the IGBT is comparable to that of a MOSFET: a channel is created in the active silicon area, if a gate-emitter voltage is applied. This channel enables conduction between emitter and collector. One advantage of that technology is the fact that the gate voltage controls the load current electrostatically, i.e. very fast and at any instant, even during switching transients.

n np

Source Gate Drain

Body / Substrate

Oxide Metal Contact

D

S

G

a) b)

Collector

Emitter

Gate

iC

vCE

VGE1

VGE2

VGE3

VGE4

iC

VCE

2-11

As is apparent from figure 2-10, above a certain level of collector-emitter voltage, the collector current remains constant irrespective of the collector-emitter voltage. However it is dependent on the gate-emitter voltage: the load current is limited by the device itself. This feature results in a very desirable effect: if an IGBT turns on into a short circuit situation, the current is limited and it usually can be turned off by the device within some microseconds when an appropriate gate signal is applied. Switching times of IGBTs are in the range of microseconds or less.

Furthermore, the switching slopes can be adjusted by the gate drive circuit to achieve the optimum waveforms with respect to over-voltage peaks and switching losses. Snubbers to keep the rate of rise of current and voltage to acceptable limits are not necessary. In comparison to GCTs, the gate drivers for IGBTs are quite simple, since they do not have to deliver high currents to the gate. These drivers are commercially available for nearly all types of IGBTs, equipped with over-voltage protection, short circuit detection and further functions, if desired.

As of 2006, the IGBT has become the power semiconductor of choice for most industrial applications of power electronics. It has also found its way into HVDC technology for low power applications (e.g. 300MW) and FACTS controllers (STATCOM), primarily for flicker mitigation.

2.3 Comparison of Basic Power Semiconductors

As with the computer and IT industries that continually reinvent themselves due to advances in logic level transistor technology, power semiconductors do not appear to have reached their ultimate capabilities. R&D is continuing with manufacturers and all discussion in subsequent paragraphs can only be a snapshot as per 2006, the time of this writing.

2.3.1 Voltage and Current Ratings

Thyristors

The voltage rating of thyristors is described by the repetitive peak blocking capability of the device at a defined junction temperature, typically 120°C. For power system applications, the industry has more or less standardized on 8,000V to 9,000V to reduce the number of series connected devices. One manufacturer is offering 12,000V elements. As thyristors are symmetrical devices, they can block the rated voltage in both forward and reverse directions.

2-12

Figure 2-11 Evolution of Thyristor Rating

The current rating of thyristors depends on the diameter of the silicon wafer, but also on the rated voltage of the device. The reason is that higher blocking voltage requires a thicker silicon wafer, which increases the on state losses. Current rating of thyristors essentially is a thermal issue. It is defined in the data sheets by a sinusoidal current pulse lasting one half cycle of power frequency and averaged over the full cycle (average on state current Itav) at a defined case (package) temperature. The latter varies in datasheets between manufacturers from 60°C to 88°C, complicating comparative evaluation. Typical numbers for an 8,000V device are 2,500A (100mm wafer material) and 3,500A (125mm wafer material), each for a case temperature of 60°C. For a 12,000V device made from 125mm wafer material the number is quoted as 1,500A for a case temperature of 88°C.

GTO and GCT

A typical peak repetitive voltage rating is 4,500V, the corresponding average on state current Itav is around 1,000A. An even more important parameter would be ITGQM, the maximum controllable turn-off current. Here, a typical number is 4,000A. Most GTOs and GCTs are made from 100mm silicon wafer material; the maximum peak voltage rating available is 6,000V at an ITGQM of 3000A. However, one manufacturer offers a device made from 150mm silicon wafer material that is rated at a peak blocking voltage of 6,000V and an ITGQM of 6,000A.

It should be noted that GTOs and GCTs for use in voltage sourced converters do not require reverse voltage blocking capability. They are therefore made as asymmetric devices, i.e. they can only block voltage in the forward direction. A diode is connected in antiparallel and protects them from any reverse voltage.

Thyristor blocking voltage

Peak thyristor current

199019801970 2000

2

4

6

8kVkA

2-13

IGBT

Like GTOs and GCTs, IGBTs are unsymmetrical devices that can only block voltage in forward direction. The ratings of IGBT devices available on the market reflect the needs of the variable speed drives industry. Here, the IGBT has almost completely replaced the GTO. As the variable speed drive industry does not normally use series connection of devices, the bulk of high power IGBTs is made in the form of modules (see 2.3.5) including the antiparallel diode. Their voltage rating is defined as the permissible collector-emitter voltage VCES. It ranges from low voltage devices (several hundred volts) to 6,500V for high voltage applications.

The current rating of an IGBT is defined by the d.c. collector current IC for a defined temperature of the case. Typical numbers are 2,400A, 1,800A, and 1,600A for a 1,700V device; 1,200A for a 3,300V device; 600A for a 6,500V device. Because of the complex MOS structure on the surface of the device, IGBTs can only be manufactured as chips with an area of about 1 cm2. All IGBT units therefore are made from parallel connected chips, refer to section 2.3.5.

Press pack IGBTs are available for high voltage applications like power electronics based transmission controllers. The maximum device ratings available are 2,500V and 2,000A.

2.3.2 Turn-on and Turn-off Characteristics

Figure 2-12 Traces of Current and Voltage Upon Turn-on of a Thyristor

Comparing the turn-on and turn-off characteristics of thyristors and IGBTs resembles comparing a tractor with an automobile. Though the mechanisms are similar (since they result from doped silicon structures), there are remarkable differences in speed and resulting secondary effects.

0.9 VDM

VDMp(t)

vT (t)

iT(t)

ITM

IGVT

0.1 VDM

t gt

tgrtgdt0 t

v, i, p

vDM: peak forward voltage; vT: on state voltage; iG: gate pulse currentiT: on state current; ITM: peak on state current; p(t): turn on loss;tgd: delay time; tgr: rise time; tgt: turn on time

2-14

Turn-On

Using the turn-on characteristic of the thyristor (figure 2-12) as an example, it is seen that turn-on is not instantaneous upon application of the gate pulse at t0. Instead, there is a delay time (typically 2…5µs, depending on the manufacturer) until voltage starts breaking down and current really starts building. This is followed by a period called rise time, during which conduction spreads out from the gate edges across the device face (at a velocity of approximately 0.1mm/µs). The highly interdigitated gate structure, figure 2-5, serves to minimize the distance of any point on the thyristor area to the gate, thus limiting the rise time to well below 10µs. The total turn-on time of a thyristor then can be taken to be in the order of 10µs, while one manufacturer quotes 15µs for a 12,000V device. Note that the numbers 0.1 and 0.9 in figure 2-12 are standardized and serve for a clear definition of the delay and rise times. During turn-on, the rate-of-rise of the current needs to be limited to avoid local overheating and destruction while the thyristor is not fully conducting; typical numbers are 300A/µs, or 100A/µs for a 12,000V device. It is on the valve designer to incorporate appropriate limiting circuits, if any, in the thyristor valve.

The turn-on characteristic of the GTO and the GCT is very similar to that of the thyristor. Due to its cellular structure with each segment being surrounded by the gate (see figure 2-6), the rise time is shorter and the turn-on time is typically quoted as 7.5µs. For the same reason, there are no problems with current spreading and di/dt is less critical. This seeming advantage is at the expense of a higher gate pulse: the many segments can be thought of as small parallel connected thyristors. To turn them on, each has to be provided with its gate current in parallel, and preferably at a high rate-of-rise of the gate pulse.

IGBT turn-on is also characterized by a delay time (until the conduction channel, see figure 2-9, is established) and a rise time (until the d.c. collector current is fully established). IGBTs have a reputation for very fast switching. In fact, for low voltage devices, e.g. 600V at 60A current rating, both the delay and rise times are measured in several ten nanoseconds. However, for high voltage devices that are of interest for power system controllers, total turn-on times quoted range from 1µs to 3µs, depending on voltage/current rating and manufacturer. The IGBT has no problem with slow current spreading like conventional thyristors.

Turn-Off

Typical voltage and current traces of the turn-off characteristic of a thyristor along with some commonly used parameters are illustrated in figure 2-13. The process is similar to that of a diode, figure 2-3. However, for practical thyristors because of their differently optimized doping profiles, reverse bias is of only limited help; clearing the carriers from the base zones is effected mostly through recombination.

2-15

Figure 2-13 Turn-Off Characteristic of a Thyristor a) Traces of Current And Voltage; b) Turn-Off Power Loss

A key thyristor parameter for the user is the circuit commutated turn-off time tq. It is the time from the current zero crossing to the point when the thyristor can take up forward voltage again. If forward voltage is applied before a time tq has elapsed, the thyristor turns on. This is called tq - triggering and has to be avoided. In order to keep tq small, the stored charge Qrr = Qs + Qf as well as the carrier lifetimes have to be small. tq increases with forward current and with temperature. For 8,000V class thyristors as used in controllers for transmission applications, tq is typically from 600 to 800µs, depending upon the manufacturer and the circuit conditions; for a 12,000V device, numbers between 1,000 and 2,500µs are quoted.

The GTO is turned off by feeding a large negative current with high di/dt into the gate (see section 2.2.3). This removes stored charges from the cathode segments, starting at their periphery. When sufficient stored charge has been removed (after the storage time ts), the anode current begins to fall rapidly, figure 2-14, and the differential between load current and anode current commutates into the turn-off snubber capacitor. This capacitor is necessary because the many segments forming the cathode (figure 2-6) do not turn off precisely simultaneously and have a tendency to filamentation, resulting in turn-off losses being generated in the small volume of the few last segments to turn off; hence the need to have commutated most of the load current into the snubber capacitor at the end of the fall time.

a)

b)

v(t) , i(t)

ITM

VT

VR

ts tf

trr

QsQf

VRM

t

t

PRQ

0

0

ITM: peak on-state current; VT: on-state voltage; VRM: peak reverse voltagets: storage time; tf: reverse current fall time; trr: reverse recovery time;tq: circuit commutated turn-off time; Qs+Qf=Qrr: reverse recovery charge;PRQ: turn-off power loss

tq

2-16

Figure 2-14 Turn-Off Characteristic of a GTO

As is apparent from figure 2-14, the GTO does not experience a reverse current as does the thyristor, due to the active removal of charges by the negative gate current. Storage time and fall time typically add up to the order of 30µs. However, at the end of the fall time, when cathode current has ceased, there is still some current flowing from the anode to the negatively biased gate that is termed tail current and sweeps out the remaining stored charge in the device. The tail time is relatively long; it can be influenced by the design of the device, however a compromise has to be found between on-state and turn-off characteristic.

With the GCT, because of the current gain of one and the extremely steep gate pulse, the storage time is reduced to 2µs and the fall time to about 1µs, with a few microseconds quoted for the tail time. The gate unit, which is triggered by an optical control signal, imposes an additional delay of 2µs. Manufacturers use different definitions in their data sheets to define the turn-off characteristic, but it would appear safe to assume that the total turn-off time is in the order of 10µs. ITGQM, the repetitive controllable on-state current , can be as high as 4,000A at a repetitive peak off-state voltage of 4,500V, or 3,000A at 6,000V. Because of the extremely short fall time, the GCT can be operated without a snubber circuit, albeit at reduced maximum controllable current.

iG: gate current; iA: anode current;vAK: anode-cathode voltage; VD: forward voltage;ts: storage time; tf: fall time; ttail: tail time

iGIG T 0

0

0

iAI0 ttail

tftsvAK

VD

time

time

time

2-17

Figure 2-15 Typical Turn-Off Characteristic of a High Voltage IGBT

The turn-off characteristic of an IGBT combines the characteristics of its two components (see section on Power Transistors in 2.2.3): the MOSFET and the BJT. The sequence of events is illustrated in figure 2-15. Turn-off is initiated by reducing the gate-emitter voltage to below the threshold level. After some delay time, the collector-emitter voltage rises to its blocking state value. Only then does the collector current start its steep descend. All this activity happens in the MOSFET section of the IGBT. As is indicated in figure 2-15, there are two clearly identifiable intervals in the collector current waveform: the initial steep drop is followed by a slower descent that reminds of the tail current in a GTO and that is due to the stored charge in the BJT section. As the MOSFET is already off and there is no negative voltage applied to the IGBT, those charges can only be removed by recombination.

While manufacturers' datasheets state the turn-off delay time and the (MOSFET section) fall time, there is no information on the tail time. As figure 2-15 suggests, the total turn-off time for high voltage IGBTs would be in the order of 6µs.

collector current iC (t)collector-emitter voltage vCE (t)gate-emitter voltage vGE (t)

1 3 5 7 9

2

4

6

kVkA

1.2

0.8

0.4

00 µs

MOSFET current

BJT current

10

-10

0

V

µs1 3 5 7 9

2-18

2.3.3 Conduction and Switching Losses

Losses result from the simultaneous presence of voltage across and current through the semiconductor. During conduction, all devices show a forward voltage drop. It is a function of the conducting cross section, the rated voltage (and thus the thickness of the silicon material), and the optimization in doping profiles chosen by the manufacturer. Also, there are marked differences between various types of devices that are due to their internal structure. Conduction losses can be calculated from pertinent parameters provided in manufacturers datasheets.

For illustration, consider a typical 100mm (4") thyristor with a peak blocking voltage of 8.000V that is operated at 2.000A. At conduction, it would have a forward voltage of 2V, resulting in instantaneous conduction losses of 4kW. If a 125mm (5") device were used, the forward voltage would be around 1.75V and the conduction losses would reduce to 3.5kW. A GTO made from 100mm material could have a rated voltage of 4.500V; at conduction it would exhibit a forward voltage drop of 3.4V. Conduction losses for this device would thus be 6.8kW at 2.000A. To make a valid comparison with the thyristor, this number has to be doubled since two GTOs would have to be connected in series to match the peak blocking capability of the thyristor. A press-pack IGBT rated 2.000A would have a rated voltage of 2.500V and a forward voltage drop of 1.9V when conducting. Conduction losses would be three times 3.8kW because three devices would have to be connected in series for equivalent blocking capability.

Switching losses are produced both during turn-on and turn-off. They exceed the conduction losses because the concurrent levels of voltage across and current through the device are higher during the switching process. Figures 2-12 and 2-13b show qualitative curves for the power loss over time, p(t), for a thyristor. Manufacturers datasheets of thyristors do not provide parameters for easy calculation of the switching losses as those are highly dependent on circuit conditions. Instead, digital simulation tools have to be used. This is also true for GTOs and IGBTs; however, datasheets for these devices do provide information on the energy loss per turn-on and turn-off process – the power loss curve integrated over several ten microseconds and for specified operating conditions and gate pulse characteristics.

All discussion above relates to the instantaneous losses incurred in the semiconductor. Loss numbers that are relevant for determining cooling requirements or for loss evaluation purposes are obtained by integrating the loss curve p(t) over one complete cycle to arrive at the RMS value. Thyristors turn on and off once every power frequency cycle; their turn-off losses in particular tend to be high because of the long turn-off time of several hundred microseconds. GTOs, GCTs, and IGBTs (in that order) exhibit substantially reduced turn-off losses because of their higher switching speed. However, they often perform multiple switching operations per power frequency cycle, up to one kilohertz. In that case the overall switching energy loss can be considerably higher than that of thyristors for comparable operating conditions.

2.3.4 Surge Capabilities

The capability of a power semiconductor to conduct and survive a surge current basically is a thermal issue. Conduction losses increase with the square of the instantaneous current value and heat up the blocking junction: there is only limited transient heat dissipation. Above approximately 120°C the breakover voltage decreases sharply, see figure 2-16, and the device

2-19

will not be able to block any voltage if the temperature at the end of a current surge reaches or exceeds about 180°C. This characteristic is reversible as long as the temperature does not get above approximately 270°C when permanent damage would occur.

Figure 2-16 Temperature Dependence of Breakover Voltage

With this in mind, the surge current capability of power semiconductors is specified in data sheets as a sinusoidal pulse of 8.3ms or 10ms duration, and applied at a time when the device is operated at the maximum blocking junction temperature, with no forward or reverse voltage applied after the current pulse. Typical numbers for an 8kV class, 100mm thyristor are 35kA to 55kA, depending on the manufacturer; a 125mmdevice of the same voltage rating could withstand 90kA. GTOs up to 6,000V are good for 25kA (100mm wafer), or 40kA for a 125mm wafer; GCTs made from 100mm silicon wafers are good for up to 28kA. IGBT manufacturers appear to follow different practices when stating a surge current capability in the published datasheets. Quoted numbers depend on the type of packaging (see section 2.3.5) and range from 6kA for a 6,500V, 600A module to 23kA for a 2,500V, 2000A press pack device.

2.3.5 Packaging

Semiconductor chips and wafers need an enclosure to protect them from the environment. Electrical contacts and means to dissipate the heat losses need to be provided. High power semiconductors may be packaged in two different housing types: press pack housings or modules.

Press Packs

Press pack housings are intended for double sided cooling. They are usually clamped between heat sinks; the paths for current and heat are copper blocks on the anode and cathode side of the wafer that are separated by a ring of insulating material, figure 2-17. For high voltage devices, this material is high strength porcelain in most cases, though fiber glass reinforced resin is also

0 50 100 150 200

temperature [ °C ]

blocking capability

2-20

used. Press pack housings with porcelain insulation are the standard design for high voltage thyristors, GTOs and GCTs. They are designed to be explosion proof; arcing inside the device due to faults will not shatter the material.

Direct series connection of press pack devices is quite simple since the devices can be arranged in one shared stack together with the heat sinks. As the silicon wafer upon failure always fails short and there is no wiring in the stack that could break and arc, the load current can continue flowing through a failed device in a serial connection. The required clamping force for a 125mm (5") thyristor, for example, is in the range of 100kN.

Figure 2-17 Packaging Of A Silicon Wafer In A Press Pack Housing

Press pack housings are also available for high voltage IGBTs. Because IGBTs are only manufactured as small chips of approximately 1cm2 size, several chips have to be connected in parallel to achieve the current ratings required for controllers in transmission systems. Two designs are in use, figure 2-18, both combining IGBT chips and antiparallel connected diode chips.

Both designs shown in figure 2-18 rely on a clamping force applied to the heat sinks placed on either side of the device; the principle of heat removal and current flow basically is the same as illustrated in figure 2-17 a). The special challenge with the mechanical design of the PPI housing is to assure even distribution of the pressure force to all chips. This is solved in different ways by the manufacturers.

2-21

Figure 2-18 Press Pack IGBT (PPI) Housing Designs

Modules

Modules are commonly used as a housing for semiconductors in industry and traction drives with and without series connection. They are designed for single sided cooling and are bolted on heat sinks by screws; spring loaded clamping is not necessary. For IGBT packaging, modules are the preferred technology.

Figure 2-19

Module Packaging of IGBTs

Individual chips are soldered to the metalized surface of a substrate that insulates the chip from the base plate while providing good thermal conductivity. In module packages the paths for current and heat are separated. As shown in figure 2-19, the electrical terminals are on the top side of the module, the heat flows through the base plate of the module to the heat sink. Since the electrical part of the module is insulated from the base plate, it is possible to mount modules having different potentials on a shared heat sink. High power modules are assembled from several semiconductor chips, which are connected in parallel, as for IGBTs and freewheeling diodes. For industrial applications they are often connected into complete converter circuits as

a) Array Of 21 Chips Arranged In APorcelain Housing (Top Cover Removed)

b) Array Of 6 Submodules Including 12 ChipsEach In A Fiber Glass Reinforced Mould

Photograph Toshiba Photograph ABB

SolderSubstrate

Semiconductorchip

Bond wire

Baseplate

a) Schematic Of Individual Chip Mounting b) Complete IGBT Module

2-22

for lower voltage diodes, thyristors and IGBTs. The electrical connection on the top side, i.e. parallel connection of chips and the connection to the terminals is realized by discrete wires. These wires are bonded to the chip surface by a process using a combination of heat, pressure, and ultrasonic energy to make a weld.

While modules can be assembled rather easily into complete power electronic equipment, they are not explosion proof as are press packs. When a fault occurs in a voltage sourced converter, the resulting short circuit currents can be sufficiently high for the bonding wires to melt and create in an open circuit. That can result in an arc within the module which leads to cracking of the module shell. For this reason, modules have not been used in power electronics based transmission controllers at the time of this writing.

2.4 Trends in Power Semiconductor Characteristics

Today, power semiconductors have reached a high reliability and high performance regarding voltage and current capability. The main goals of developments are:

• further increase of blocking voltage and / or current capability

• reduction of losses

• increase of ruggedness by integrating self-protection functions

As long as monocrystalline silicon remains the base material for power semiconductors, higher ratings and lower losses for thyristors, GTOs, and GCTs can be achieved by the use of larger diameter wafers, such as 150mm (6"). As power electronics based transmission controllers are practically the dominant market for the extra high power devices, the needs of projects will determine whether the trend is towards higher blocking voltage or higher current rating.

IGBT chips are mass produced in very large quantities and to specifications that are optimized for industrial applications. Given the small number of devices used in the transmission controller market and the high cost of development and of investment for manufacturing, it is not very likely that special chips will be developed, which can match the blocking voltage levels achievable with thyristors unless they are required by industrial or traction applications.

New materials have been proposed for use in power semiconductors, such as silicon carbide (SiC) or diamonds. They are expected to offer higher blocking voltages, higher current ratings out of smaller wafer sizes, higher operating temperatures (reduced cooling requirements), and lower losses. A 110kVA inverter based on SiC power devices has been demonstrated; however there are still substantial material issues that need to be overcome before this technology will be available for transmission controllers.

R&D programs are underway to develop diamond based semiconductors. They have a promise of offering even better characteristics than SiC. Though some experimental high voltage switches were built, the technology does not seem to be ready for commercial application, nor be competitive to currently available high power silicon based devices. It can thus be expected that in the near to mid term future power electronic devices used in transmission controllers will undergo an evolution, not a revolution.

2-23

2.5 Functional Switch Requirements in High Power Transmission Controllers

Power electronics based transmission controllers are based on either of two principles: switching of impedances or dc/ac conversion. They are connected to the high voltage ac system in shunt through a transformer or in series to a transmission line. Their operating voltage therefore typically is below 100kV with the exception of HVDC where 800kV is close to being introduced commercially.

Whether impedance switching or conversion, the switches used must operate at least once every cycle of power frequency for 8760 hours per year, and for a design life of 30 years. – a requirement that could never be met with mechanical switches. Switches must react instantaneously to control commands and they must be reliable: despite their often complex circuitry including easily hundreds of components, they are expected to be maintenance free outside an annual (biannual, triennial) servicing period.

Switches are required to perform their function in synchronism with the ac system voltage for all operating conditions, including disturbances. Should the ac voltage be lost temporarily, they must be able to regain operability immediately at voltage recovery.

Switches/converters in transmission controllers are usually installed in substations; they must function under all ambient conditions, therefore they must be appropriately shielded from the environment. They must be automatically remote controlled and include adequate protection against thermal overloading and dielectric overstressing.

2.6 The Need for, and Issues of, Semiconductor Valves Using Series Connected Devices

The power handling capability of single semiconductors, as expressed by the product of peak blocking voltage and maximum rms current, is high: around 40MVA for thyristors, 10MVA for GTOs and GCTs, and 5MVA for IGBTs. The rating of transmission controllers, however, typically is above 100MVA; a single semiconductor therefore cannot meet the requirements of a switch or valve (the two terms can be taken synonymously).

To achieve the rating required, devices are often connected in series. Alternatively, for voltage sourced converters (see Chapter 3), complete converters are sometimes connected in series or in parallel. Direct parallel connection of devices was discontinued with the availability of large devices with high current rating.

With series connection of devices, a number of issues have to be addressed in the design of the valves:

• Even distribution of power frequency and transient voltages in the series string has to be assured. This is typically achieved by a passive voltage grading network. For IGBTs, due to their fast switching characteristic, active voltage sharing control through the gate units is also possible.

2-24

• The series connected devices, each being at different potential, must be adequately insulated to ground. This requires a structure holding the devices that provides a staged insulation.

• All devices in the series string need to be turned on (and turned off) simultaneously to establish current flow (and current turn-off). If individual devices turn on late (turn off early), they would be overstressed by the imposed system voltage.

• To bridge the different potentials of individual devices, their respective gate units are triggered from ground potential through fiber optic links; the auxiliary power needed is extracted from the main circuit by appropriate circuitry.

• Heat losses have to be removed from the devices at their individual potentials. The method of choice is insulating pipe work in the valve structure to circulate deionized water to an external heat exchanger.

2.6.1 Voltage vs. Current Rating for Optimal Design

In most cases, shunt connected transmission controllers are connected to the high voltage grid through step-down transformers. In principle, therefore, there is much flexibility in selecting the rated voltage of the semiconductor switch. The optimum design approach then is to use the maximum current capability of the semiconductor device, considering any overload and fault current withstand requirements, and to connect a sufficient number of devices in series to achieve the voltage rating required to arrive at the desired power rating of the controller. In view of overall cost for a practical approach, consideration will be given to the availability and cost of ancillary equipment including switchgear, insulators, buswork, PTs and CTs, and an appropriate standard secondary voltage of the transformer will be selected. Typical voltage ratings for shunt connected controllers are in the 30kV range.

Switches in controllers that are connected in series to a high voltage line of necessity have to be rated for the line current including any transient fault current. Their voltage rating then is a straightforward function of the desired power rating of the controller.

It may be worth noting that in the design for power electronics based transmission controllers, it is general practice to utilize the semiconductor devices during normal operation up to about half of their peak blocking capability only. The other half is reserved for dynamic and transient overvoltages, margins for voltage sharing unbalance, protection margins, and test safety factors.

2.6.2 Redundancy and Protection Issues

Modern power semiconductors have evolved into extremely reliable devices as demonstrated by the failure rates of thyristors published biannually by CIGRE for HVDC systems, e.g. [5]. Half of the 30 systems reporting did not have any thyristor failure over the two-year reporting period; in most other cases the failure rate was below 0.1%. No equivalent information exists for high power GTOs and IGBTs; however, the CIGRE data can be applied to thyristor valves for reactor control and thyristor valves for capacitor switching. Given mature valve design practices and device manufacturing technology, power semiconductor failures can be classified as statistical events.

2-25

Other than in most industrial applications of power electronics, failure of a complete transmission controller in case of even a statistical failure of a single semiconductor is not acceptable. Semiconductors in the series string are required to fail short and not open circuit, and they must be able to conduct the load current until they can be replaced during the next maintenance activity. As long as the controller is in normal operation without overvoltage stresses, failure of a single semiconductor usually does not endanger the remaining devices in the switch because they are typically utilized to 50% of their capability. However, probability is high that voltage transients are experienced during the potentially long time to maintenance of one year (or more). Therefore, at least one redundant semiconductor is added in the series string of each switch. Degradation of voltage withstand capability then is only given when one more semiconductor fails than the number of redundant devices. Typically, the status of each semiconductor in the switch is monitored by an external circuit, and an alarm is issued if more than the number of redundant series levels have failed.

Power semiconductor switches for transmission controllers are often protected against transient overvoltages by direct parallel connected surge arresters. Their withstand capability is coordinated with the protective level of those arresters as determined in insulation coordination studies. In addition, individual devices in thyristor switches may include a forward overvoltage protection that initiates a gate pulse if a safe level of forward voltage stress is exceeded, gating the device safely into conduction.

Transmission controllers include some current control feature, acting on the gate pulses. As this is typically very fast, it is also used for overcurrent protection. Upon detection of excessive overcurrent by means of appropriate sensors, such as a CT or differential protection, further gate pulses may be inhibited to block the switch. For some controllers using thyristor switches a suitable strategy may also be to make the switch ride through the fault by continuously gating the thyristors, thus relieving them from any voltage stress. Special considerations apply to IGBT switches: due to their characteristic (see section 2.2.3) they are able to limit and turn off fault currents if an appropriate gate signal is provided within less than 10µs. An extremely fast detection system is needed to make use of this feature.

2.7 Design Considerations of Thyristor Valves for Reactor Control

Reactor control is used for shunt reactors, commonly known as Thyristor Controlled Reactors (TCR), which are typically connected to the high voltage grid through step-down transformers. Control of the reactive power is achieved by placing a bidirectional thyristor switch (commonly termed "valve") in series with the reactor and delaying the turn-on gate pulses relative to the natural current zero crossings, thus effectively varying the r.m.s. value of the reactor current, see chapter 5. Reactor control by thyristor valves is also used in Thyristor Controlled Series Capacitors (TCSC) where a thyristor controlled reactor is connected in shunt to the capacitor bank to vary the series impedance, see chapter 6. The same design considerations apply for both applications.

2-26

Figure 2-20 Structure Of A TCR And TSC Thyristor Valve

The basic element of the thyristor valve is the pair of antiparallel thyristors; each pair has an electronic unit for gating, monitoring, and individual overvoltage protection. It is also equipped with a snubber circuit for a.c. voltage grading and damping of transients and may include a resistor specifically for d.c. voltage grading: although the steady state voltage across a TCR valve has no d.c. component, if the valve is blocked there may develop internal d.c. voltages due to different leakage currents of the thyristors. The thyristor pair along with all its auxiliary components is commonly known as a thyristor level, and a number of levels is connected in series so as to meet the voltage withstand requirements of the valve.

Obviously, the thyristor valve needs to be designed to withstand all steady state, dynamic, and transient voltages that it may be subjected to in service (as determined in system studies), whether those are caused by faults in the system or produced internally within the valve. A specific case for the latter would be the recovery voltage overshoot at turn-off (figure 2-13). It may be limited by a surge arrester as shown in figure 2-20; however, as this is a periodically recurring phenomenon, the arrester would have to be specially rated for this duty.

Design of TCR thyristor valves with respect to current stresses is not normally a problem: given the large diameter thyristors available and the highly efficient water cooling, even severe conditions as a.c. system faults resulting in a sustained asymmetrical current without zero crossing can be accommodated until protection operates (see 2.7.4).

1 Gating and monitoring unit, local overvoltage protection; 2 Snubber circuit (voltage grading);3 DC voltage grading (optional); 4 Valve reactor (optional); 5 Valve arrester (optional);6 Valve supervision; 7 Firing signal generator; 8 Valve protection

4

5

2

3

1

2

3

1

2

3

1

2

3

1

2

3

1

86 7Firing signals

to the thyristors

Monitoring signalsfrom the thyristors

Valve electronics(ground potential)

Thyristor valve

2-27

2.7.1 Turn-on of Series-Connected Thyristors

The control range of thyristor valves for reactor control is defined by the delay of the turn-on gate pulse with respect to the zero crossing of the valve voltage, called the firing angle α. It ranges from 90° (full conduction) to 180° (no conduction). When in the full conduction mode, i.e. rated current of the reactor, thyristors are turned on into the crest of the system voltage. This results in the highest di/dt from discharging snubber circuit and stray capacitances. Thyristors have to be rated for this duty, otherwise a limiting reactor would have to be included in the valve (figure 2-20). Turn-on duty is reduced with increasing firing angle because the valve will then turn on from a lower momentary value of the system voltage.

2.7.2 Turn-off of Series-Connected Thyristors

The design needs to consider the turn-off characteristic of thyristors as described in section 2.3.2, see figure 2-13, in particular the stored charge Qrr for the individual devices. Qrr is not a fixed parameter; it depends on di/dt at current zero crossing and junction temperature, but it also exhibits manufacturing tolerances. The effect in the series connection is a variation in voltage sharing after turn-off, at voltage recovery, as illustrated in figure 2-21.

Figure 2-21 Voltage Unbalance Of Series Connected Thyristors As A Result Of Variation In Recovered Charge Qrr (Simplified, Not To Scale)

The situation depicted in figure 2-21 can be thought of as all but one thyristors in the series string having a Qrr B and one thyristor having a lower Qrr A. The latter thyristor recovers first and takes on the instantaneous system voltage at that point in time (curve VA). The reverse current in the valve that is still flowing bypasses the thyristor through the snubber circuit, charging the capacitor. The other thyristors, recovering later, take on their portion of the instantaneous system voltage prevailing at that point in time (curve VB). As a result, the thyristor with Qrr A gets a higher negative voltage but a lower positive voltage than all the other thyristors; also, the voltage turns positive at a later point in time. This unbalance is repeated in the same way from cycle to cycle; it is a major criterion for designing the snubber circuits of the thyristors and for determining the number of series connected devices in the valve. An optimum is sought in the design between the permitted spread in Qrr, the size of the snubber capacitor and the associated snubber losses, and the total number of devices in the series string.

Qrr A

Qrr B

V A

V B

I

t

t∆V = ∆Qrr/C

2-28

2.7.3 Gating Requirements

As a basic requirement, all thyristors in the series string must turn on simultaneously to establish current flow. For electrically triggered thyristors, the gate pulse with an amplitude of several amperes and a duration of about 10µs is provided by a gate unit located adjacent to the thyristor that in turn is initiated by a control signal from ground potential. As a general practice in the industry, control signals are transmitted through fiber optic links to cross the potential between ground and the individual thyristor level. Auxiliary energy is drawn from the main circuit by each gate unit. Due to component tolerances and slightly varying turn-on times of the individual thyristors, a jitter of typically one to two microseconds has to be expected. During this period, the snubber circuit in parallel to each thyristor forms a bypass for those devices that have not turned on yet.

For direct light triggered thyristors, the gate pulse of about 40mW is provided through a fiber optic link directly from the controls at ground potential. As all gate pulses originate from the same light source, the jitter is reduced to the variations in thyristor turn-on time. When direct light triggered thyristors are used, the function of the device labeled "1" in figure 2-20 is reduced to monitoring the thyristor performance.

2.7.4 Protection of the Valve

Provision has to be made for the case that one (or a few) thyristor in the series string does not receive a firing pulse. In this case the full prevailing system voltage would be imposed on that thyristor; depending on the operating conditions there is a high probability that it would be overstressed and fail. Each thyristor is therefore fitted with a protection circuit that initiates a gate pulse and safely turns the device into conduction if the voltage gets close to its peak blocking capability. This protection is commonly known as VBO (Voltage Break Over) protection. For direct light triggered thyristors, it is integrated in the thyristor wafer itself.

As a special case for thyristor controlled reactors, if the firing angle of the thyristors is very close to 180°, the system voltage at turn-on may be lower than the VBO level of a single thyristor. In that case a missing firing pulse to one thyristor would result in unsymmetrical blocking of the whole switch. To avoid a significant dc component of the valve current, the maximum firing angle of the healthy polarity must then be limited by control action.

If a thyristor valve for reactor control turns on at a firing angle below 90°, this is considered false firing. It causes a pronounced dc current and an unbalanced loading of the valve. If the firing occurs at α = 0, this represents a phase-to-phase short circuit of the transformer and results in the fully displaced short circuit current in one direction. For protection, continuous firing for both current directions may be applied, avoiding voltage stress on the valve while the initial dc component of the current dies down. Blocking of the affected valve after the first current zero will only be successful if the hold-off interval is sufficient for the thyristors to regain their blocking capability.

Overcurrent protection of TCR valves may be initiated by an instantaneous or a time-dependent overcurrent relay, with due consideration of the prospective overcurrents as determined from system studies. High current by itself does not damage thyristors; it results in extra heating due

2-29

to the increased conduction losses. If the rated junction temperature is exceeded substantially, the blocking capability of the thyristor decreases (figure 2-16), resulting in voltage withstand failure.

Depending on the design, protective action in case of overcurrent may include:

• Limit the valve current (provided the valve is still controllable)

• Apply continuous turn-on at α=90° to avoid voltage stress (provided the TCR reactor impedance is large enough to limit the circuit current to the rated continuous current of the valve)

• Trip as the ultimate action.

2.8 Design Considerations of Thyristor Valves for Capacitor Switching

The structure of thyristor valves for capacitor switching (TSC valve) is the same as for TCR valves, figure 2-20. Contrary to a TCR valve, continuous control of the capacitor current by phase angle control is not possible. The TSC valve therefore acts as a switch that is either ON or OFF. To be in the ON status, the thyristors of the appropriate current flow direction are gated each half cycle at the natural current zero crossing of the capacitor current; they then carry the full capacitor current. To turn the valve into the OFF status, the gate pulses are blocked and the current extinguishes at the next zero crossing.

For the design of the TSC valve, special consideration needs to be given to the effect of the series reactor (see chapter 5). The inductance of this reactor, besides setting the resonance frequency of the circuit including the capacitor and all inductances in the system, serves to limit the inrush current from the capacitor at the instant of valve turn-on. On the other hand, it has the effect that the capacitor voltage is higher than the system voltage by a factor that is a function of the TSC self resonant frequency and of the system frequency.

During steady state operation, because the capacitor current precedes the system voltage by 90°el, the actual turn-off happens at the crest of the system voltage. Upon current extinction, the capacitor remains charged to the same voltage, discharging through internal discharge resistors and external circuits such as the thyristor valve. That discharge process is slow (up to several minutes) and the thyristor valve, being connected to the capacitor at one terminal and (through the series reactor) to the system voltage at the other, is stressed by the differential of the capacitor and system voltages. This amounts to a power frequency oscillation between zero and twice the crest value of the system voltage, i.e., an offset (d.c. component) by the capacitor voltage. Compared to a TCR valve connected to the same bus, the TSC valve therefore requires approximately twice the number of thyristor levels connected in series. Also, to obtain even distribution of the d.c. component within the valve, each level normally is fitted with a grading resistor (item 3 in figure 2-20).

As with all power equipment, the TSC valve design needs to consider voltage and current stresses resulting from system faults and control malfunctions, as determined from system studies. A TSC in the ON status is a special piece of equipment in that respect as any sudden change in system voltage may excite high oscillatory currents between the capacitor and any inductances in the circuit, also stressing the thyristor valve. Such would be the case for a phase-

2-30

to-ground fault close to the TSC. Depending on the point-on-wave of the fault, the magnitude of the oscillating current may be up to five times the nominal current of the TSC and needs to be considered in the design of the valve. At fault clearing the current extinguishes at the natural current zero, which is coincident with the crest value of the system voltage including any temporary overvoltage and another oscillatory current may be initiated in the TSC, to be considered in the design of the valve.

2.8.1 Turn-on of Series-Connected Thyristors

When the TSC valve is ON in steady state operation, the thyristors for the two current flow directions are turned on alternatingly at the respective natural current zero crossing, i.e. at a delay angle α=270°. At this point the valve voltage (being the differential between the capacitor and system voltages) may be in the order of 5% of the system voltage only. This tends to slow down the turn-on of the thyristor; special care has to be taken in the design of the valve to assure simultaneous turn-on of all series connected thyristors, e.g. by selecting devices with a low latching current (see 2.2.2) and by providing a strong firing pulse.

2.8.2 Turn-off of Series-Connected Thyristors

The turn-off process of thyristors in a TSC valve is similar to that in a TCR valve (see 2.7.2): due to the recovery charge, an oscillatory transient voltage overshoot is superimposed on the system frequency recovery voltage. However, in steady state operation, the momentary value of the valve voltage at current extinction is of the order of 10% of the crest system voltage and therefore the peak value of the extinction overshoot is several times lower than the maximum voltage appearing across the valve half a cycle after extinction (around twice the system crest voltage). The oscillation is damped by the RC snubber circuit connected in parallel to each thyristor pair; the snubber circuit is also designed to reduce differences in voltage distribution along the series string that are caused by differences in the reverse recovery charge of individual devices. Here, the same considerations apply as discussed in 2.7.2 for the TCR valve.


As with the TCR (see 2.7.4), gating of the thyristors is initiated from the control system at ground potential through a fiber optic system. For electrically triggered thyristors, the auxiliary energy needed for the gate pulse is obtained from both the voltage across the thyristor when the TSC is OFF and from the current through it when the TSC is ON. It is not sufficient to obtain the energy from one source only because the TSC may be ON or OFF for an extended period of time. For direct light triggered thyristors, the energy is provided by the light pulse from ground potential and circuits to obtain auxiliary energy for the gate pulse at thyristor potential are not required.

2.8.4 Protection of the Valve

Overvoltage protection of the valve is achieved by a combination of arresters protecting the complete string of series connected thyristors and voltage breakover (VBO) circuits protecting

2-31

individual thyristors. Arresters are often connected in parallel to the combination of valve and series reactor; alternatively, arresters may be connected both directly in parallel to the valve and to the combination of valve and series reactor, or in parallel to the capacitor and the combination of valve and series reactor. In any case, the selection of the arrester rating must consider the particular voltage stress of the valve at turn-off, i.e., the superposition of the system voltage on the voltage trapped in the capacitor. It must be assured that the arrester is not overstressed when the TSC is switched in and out with a high repetition rate.

VBO protection is used to protect individual thyristors by gating them into conduction if their voltage exceeds the permissible level. For a TSC valve, this protection needs special consideration as it must not interfere with the operating strategy of the TSC: to be switched off in case of system overvoltage. The VBO level of the complete valve, i.e. the collective sum of all individual thyristor VBO levels, must be chosen sufficiently above the arrester protective level. Also, if the VBO protection of a single thyristor operates periodically, e.g., because of a missing gate pulse, the resultant slight delay in valve turn-on results in harmonic currents and it must be checked whether those have a detrimental effect on system operation.

Overcurrent protection of the valve aims at maintaining a thyristor temperature that permits safe blocking of the valve according to the operating strategy adopted. In most applications, the TSC is switched off during overvoltage conditions and then will not be exposed to overcurrents. However, high overcurrents may be a result of system faults or false firing of the valve. Though the risk of the latter is extremely low with properly designed control and gating systems, in some cases it is used as a design criterion for the current capability of the valve.

Overcurrent protection strategy depends on system requirements. The valve may be designed to withstand the worst fault currents specified (which need not be the highest currents expected) and subsequently block. The protection strategy would then be to initiate continuous gating of the valve if higher than the specified currents occur and trip the SVC with the a.c. breaker. The same strategy could be applied to false firing: design for one incident with subsequent blocking and initiate continuous gating with breaker tripping for repeated incidents.

2.9 Design Considerations of GTO Thyristor Valves for Voltage-Sourced Converters

As with transmission controllers using valves made from phase control thyristors, the power rating of voltage sourced converters for transmission controllers by far exceeds the capability of a single GTO thyristor device (see 2.6). Adequate ratings of valves/switches can be achieved by either one or a combination of three options: (i) series connection on the a.c. side of complete single phase converters made from a single GTO per switch; (ii) parallel connection on the d.c. side of complete converters made from a single GTO per switch; (iii) series connection of GTOs in each valve of the converter. In all options, converters may have various topologies (see chapter 3); depending on the total configuration, there may be a need for parallel connection on the a.c. side to make the complete controller – typically by some kind of transformer. An example for option (i) is the chain-link STATCOM [7]; option (ii) has also been used in several applications [8]. An example for option (iii) is the Inez UPFC [9] that also makes use of parallel connection of converters.

2-32

The main advantage of options (i) and (ii) would be that the individual converters can be controlled such that their combination functions as a multilevel converter, resulting in, e.g., a 48-pulse system with nearly sinusoidal a.c. voltage output. This performance in some cases is further enhanced by low frequency PWM control. Also, if a GCT is used as the basic switch rather than a GTO, snubber circuits may not be necessary and the rather high snubber losses can be avoided.

Series connection of GTOs has proven to be more difficult to accomplish than series connection of thyristors. The reason is the rather long storage time of the GTO (see figure 2-14), which is in the order of 25µs and is subject to manufacturing scatter, see 2.9.2.

During operation of the voltage sourced converter, the GTO valve (including antiparallel connected diodes) acts as a switch to apply positive and negative voltage to the load. During the switching (commutation) process, the energy stored in the stray inductance has to be absorbed by the snubber circuits. To minimize the resultant losses, care is taken in the physical configuration to keep the stray inductance as low as possible, e.g. by choosing a suitable layout and/ or by using special low-inductance buswork.

2.9.1 Turn-on of Series-Connected GTO Thyristors

In contrast to thyristor valves in a TCR or TSC, GTO valves in a voltage sourced converter are always turned on into the (constant) d.c. capacitor voltage. Depending on the circuit impedances, it may be necessary to include a current limiting reactor in the valve to limit di/dt at turn-on. This reactor then would have to be designed to withstand the discharge current of the d.c. capacitor during internal faults. As with phase control thyristors, individual GTOs exhibit differences in delay time, resulting in the need for snubber circuits. On the other hand, the rise time of GTOs is shorter: due to the heavy segmentation of the cathode (figure 2-6) with each segment being surrounded by the gate there is no spreading effect and with a sufficiently large gate pulse all the segments will turn on simultaneously, reducing the duty of the snubber circuits and the turn-on losses.

The current limiting reactor may be avoided if the turn-on gate pulse is made very steep, in the order of kA/µs. However, there is a limitation due to the inherent inductance of the gate circuit and the permissible gate voltage. A way out is the use of GCTs rather than conventional GTOs; they exhibit an inductance of the gate circuit of a few nano henries only. Use of GCTs with very steep turn-on gate pulses would potentially also eliminate the need for turn-on snubbers, as the delay time and thus its variance between devices in the series string is a function of gate pulse di/dt and becomes insignificant if that number is in the kA/µs range. However, snubbers are also needed to assure even voltage distribution between devices in the series string, see 2.9.2.

Although the GTO is a latching device and will keep conducting without a gate pulse until it is turned off, during dynamic conditions the load current may momentarily fall to a low value, then rise again. Under such conditions, a GTO without gate current may partially unlatch, that is, the current density may be insufficient to maintain the entire cathode area in a latched thyristor mode. This is, in itself, harmless but the consequential undefined conducting area results in an equally indefinable di/dt capability. Should the load current rise rapidly, current crowding could result, possibly provoking an apparent di/dt failure. This can be circumvented by supplying a

2-33

continuous gate current (also known as the “back-porch” current), typically around 5A at operating temperature, in order to keep all segments in the conducting state.

This continuous gate current is also important when the load current becomes negative, and load current flows through the free-wheeling diode. In this condition, the GTO returns to its off-state, and may then be unable to reabsorb the load current, when this latter becomes positive again. With inadequate gate current to ensure re-triggering, and no negative gate voltage to ensure blocking, the device could spuriously and destructively re-trigger at some anode voltage below its normal rating.

2.9.2 Turn-off of Series-Connected GTO Thyristors

GTOs are turned off by a negative gate pulse with an amplitude of 20% to 30% of the GTO anode current. As with all series connected semiconductor switches, all GTOs in the valve need to turn off simultaneously. If one device were to turn off while the remainder are still conducting, with no special provisions that one will be stressed by a recovery voltage that is a multiple of its blocking capability and it will certainly break down. As it would be uneconomical and even impractical to select devices with equal turn-off time (tf in figure 2-14), each GTO is shunted with a RC snubber circuit. The capacitor size (typically 6µF for a 4.5kV, 3kA GTO) is chosen to limit the voltage rise and assure that all series devices will turn off before any of them experiences dangerous stresses.

Another issue is the rather long storage time of typically 25µs before the turn-off gate pulse shows any effect. Like any semiconductor parameter, the storage time is subject to manufacturing tolerances that result in uneven voltage distribution between devices in the series string. The imbalance can be estimated with the simple formula ∆V = 1/Cs * ITGQ * ∆ts where Cs is the snubber capacitance and ITGQ the current to be turned off. Assuming a snubber capacitance of 6µF and a current of 3kA, then a spread of ± 10% in the storage time would result in voltage unbalance of 2500V, which is more than half the peak blocking voltage of a 4.5kV GTO – a totally uneconomic approach.

There are several options to improve the situation:

• Increase the size of the snubber capacitor – at the expense of increasing the snubber losses, which are a substantial part of the operating losses.

• Match GTOs in one valve for close-to-zero tolerance of storage times. This option is undesirable because it requires extra effort for qualifying and matching devices, reduces production yield, increases device cost, and complicates spare part warehousing.

• Adapt gate units. In principle, gate units could be tuned to the storage time of the associated GTO so that tolerances are compensated. This approach requires the warehousing of GTO/gate unit combinations for spares.

• An alternative method of addressing series operation is adaptive gate control, where the turn-off timing is individually and automatically adjusted for each individual device, so that the resultant re-applied voltage is the same for each device in the series stack. Gate units can be made "intelligent" so that they are able to learn the characteristic of the associated GTO, e.g.

2-34

during commissioning or following replacement. This technology works well but is complex and costly.

• Still another approach is to use GCTs. The hard-drive commutation mode shrinks device storage times to about 1 µs, such that a ±10% spread results in a ∆ts of only 200ns, yielding a ∆V of only 100 V in the above example.

When the GTO is in the off state, the gate must be biased with a negative voltage >2V (typically 15V) in order for the device to be able to block full rated voltage at any operating temperature.


The GTO gate drive has to fulfill the following functions:

• Turn the GTO on by means of a high current pulse

• Maintain conduction through provision of a continuous gate pulse during the on-state (also known as the “back-porch current”)

• Turn the GTO off with a high negative gate current pulse

• Reinforce the blocking capability of the off-state device, by negative gate voltage.

GTOs require more gate current for turn-on than conventional thyristors, because:

• The numerous segments are in fact so many parallel connected small thyristors that each require their own gate pulse for turn-on

• •GTOs have no amplifying gate, thus the whole current at the turn-on front must be delivered by the gate unit alone.

Depending on wafer diameter and desired anode current di/dt, turn-on gate pulses may vary from a few tens to several hundred amps. They should have a short rise time (high di/dt) to reduce turn-on switching times and variations between devices in the series string. The duration of this high current pulse should be as long as the sum of the delay and rise times (see figure 2-12), after which a current of typically five amps should be provided during the on-state to assure continuous conduction of all segments.

The amplitude of the turn-off gate pulse should be 20% to 30% of the current to be turned off. The rate-of-rise should be as high as practical considering the stray inductance of the gate circuit to reduce the storage time and its variation between devices in the series string. The decay should last until the end of the tail time of the GTO (see figure 2-14).

In addition, the minimum on-time and minimum off-time as specified in the GTO datasheets must be observed as well as the required discharge time of the snubber capacitor during the on state. The gate unit must be able to provide turn-on and turn-off pulses several times during one electrical cycle of power frequency as required by the control strategy of the converter, e.g. pulse-width-modulation (PWM).

2-35

The gate unit, also termed gate drive, is associated to each GTO in the series string at high potential to ground. Typically, it is controlled by converter control systems at ground potential through a fiber optic link while pulse transformers have also been used for high voltage insulation. Auxiliary energy may be provided from ground potential through transformers but it may also be extracted from the power circuit, e.g. the voltage across the GTO in the off-state.

2.9.4 Protection of the GTO Valve

A well designed GTO valve will not be stressed beyond the capability of its components for normal steady state and dynamic operating conditions. However, these limits may be exceeded for abnormal conditions, such as control malfunction, severe a.c. system faults, component failures, or disruptive discharges anywhere in the valve structure. Protection of the GTO valve is an integral part of the overall converter protection strategy that depends on the converter structure, the application, and the design philosophy.

Protection of the GTO valve aims at avoiding or at least limiting damage to the equipment and thus maintaining availability of the converter to the power system. The preferred protective action is to act on the converter controls because of inherently faster response, with disconnection of the converter from the power system as a last resort. This approach finds its limitation if the control system itself malfunctions. An example would be the "shoot-through": if a valve is turned on before its counterpart in the same phase leg of a two-level converter is fully turned off, a direct bypass to the d.c. capacitor would be established, resulting in a potentially destructive discharge current in the GTO valves and requiring rapid action of the a.c. breaker. Protection against this fault can only be by an appropriate design of the control system.

Assuming that the a.c. interface of the VSC is protected in a conventional way by surge arresters, the voltage stress of a GTO valve is defined by the voltage of the d.c. capacitor, which in turn is a function of the charging current that corresponds to the supply of real power from the a.c. system. Protective action upon the occurrence of excessive d.c. voltage can then be to reduce the real power drawn from the a.c. system by control action or to turn-off the valves. Alternatively, the d.c. bus voltage may be limited by an auxiliary clamping circuit, including an electronic switch in series with a high energy/low ohmic resistor that temporarily diverts the charging current from the capacitor.

Overcurrent protection of the GTO valve would go along with overcurrent protection of the converter: instantaneous overcurrent protection in the microsecond time frame and overload protection (tens of milliseconds). The former is required to assure that the maximum turn-of current specified for the GTOs is not exceeded and thus needs to be very fast. It may include a change in switching pattern to provide current limitation until the current returns to normal, or brief turn-off of the valves. Inclusion of a fast-operating bypass to the converter valves may also be considered. Overload current in a VSC persisting beyond the time limits provided for in the design, would be an indication that the control system is malfunctioning. Protective action therefore would be to trip the a.c. breaker or to use a mechanical by pass to divert the current from the valves.

Additional considerations apply to GTO valves of a VSC that is connected in series into a power line. In this case the current in the a.c. interface is impressed by the a.c. system and is not directly

2-36

controllable by the converter. Also, the voltage at the a.c. interface is not limited by the system. The converter then is protected against the rapidly rising current following an a.c. system fault by a bypass consisting of a two-way thyristor switch with a parallel mechanical breaker.

Protection of the GTO valves against ground faults in the converter can be provided by an appropriate configuration of the converter circuit if the a.c. interface includes a transformer. Grounding the circuit at a single carefully selected point through a resistor not only minimizes voltage stresses on the insulation to ground; by appropriate selection of the resistance value it can also be assured that the fault current that could result from a ground fault anywhere in the circuit will not be large enough to damage the valves.

Failure of components within the valve itself, e.g. snubber circuit or gate unit, results in failure of the associated GTO, which always fails short. Valves made from series connected GTOs include a number of redundant devices in the series string, including their associated circuits. Thus failure of components up to the redundancy does not overstress the remaining devices in the string and operation can be continued until the next maintenance opportunity. The GTO valve includes component and subsystem monitoring functions that provide the operator with continuous information about the status of the valve.

2.10 Design Considerations of Transistor Based Valves for Voltage-Sourced Converters

The only type of transistor that is practical for transmission controllers is the IGBT. It combines high current rating –though not as high as thyristors- with moderate voltage blocking capability and low gate power requirements. It has higher switching speed than the GTO and thus it is suitable for higher operating frequency up to around two kilohertz. Its disadvantage against the GTO is the increased conduction loss.

The use of IGBT based valves in voltage sourced converters for transmission controllers would be subject to some limitations. The maximum rated d.c. current presently is stated as 2000A, the surge current as 23kA. This would appear to prevent their use in converters that are connected in series to the transmission line in many cases. 2000A class IGBTs in a press pack housing are available with a peak blocking voltage of 2500V only, resulting in approximately twice the number of series connected devices compared to GTOs.

On the positive side is the possibility to avoid turn-on and turn-off snubber circuits and the associated losses, due to the high switching speed. However, voltage grading circuits are needed in the series connection and the gate unit needs to be designed with some built-in control power to optimize the switching process for the prevailing operating condition, including appropriate circuits to measure the IGBT voltage at turn-off and transmit the information through a fiber optic link to the controls at ground potential. Auxiliary power for the gate units would be extracted from the voltage divider circuit.

The high switching speed facilitates operation at frequencies up to two kilohertz. Voltage sourced converters with IGBT based valves therefore can be effectively used for pulse-width-modulation (PWM) control, reducing harmonic voltages at the a.c. connection. This eliminates

2-37

the need for complex multilevel converter topologies. However, overall losses are increased due to the large number of switching operations and efficient cooling with deionized water is needed.

IGBTs are voltage controlled; the gate units therefore are low power equipment. As such, they are more susceptible for electromagnetic interference. Being located in close proximity to the IGBT with its high switching speed, they need to be well shielded.

2.10.1 Turn-on of Series-Connected Transistors

As with all series connected switches, it is essential that all IGBTs turn on precisely at the same time. In contrast to thyristors, it is not sufficient to apply a short gate pulse only; the gate voltage must rise above a threshold level and then stay positive during the desired conduction interval. Each IGBT controls its current according to the applied gate voltage (see figure 2-10). To achieve even sharing of the applied total d.c. voltage, it is therefore essential that all devices are provided the same gate voltage waveform, both during turn-on and during conduction. Load current rise is determined by the time constant of the external circuit only; the IGBT does not require additional provisions for limitation.

2.10.2 Turn-off of Series-Connected Transistors

Turn off of IGBTs is controlled and can be influenced by the gate voltage. It is therefore important that the gate units provide all devices connected in series with the same gate voltage waveform. Turn-off snubbers then are not required, because spread in turn-off delay time will be below one microsecond.


To hold the conducting channel within the IGBT open, a positive gate voltage must be applied at all series devices during the conduction period. For turning the devices off, in principle it would be possible simply remove the gate voltage. However, to speed up and control the turn-off process, a negative voltage is applied at the gate during the off-state. By monitoring the voltage across each IGBT and controlling the gate voltage at each level by the gate unit in coordination with the controls at ground potential, the turn-on and turn-off processes are optimized and each IGBT in the series string is regulated to the correct voltage level. Details of the gate voltage vs. time characteristics and associated drive circuits are proprietary to the manufacturer.

2.10.4 Protection of the Transistor Valve

Being an element of a voltage sourced converter, the IGBT valve basically needs to be protected against the same types of potentially dangerous stresses as the GTO valve (see 2.9.4). Some differences in protective action may result from the capability of the IGBT to limit a fault current if it is within the Safe Operating Area (SOA), and for not more than approximately 10 microseconds. That implies that very fast sensors are used in the protection of the IGBT valve.

2-38

Selected References

[1] Benda, Gowar, Grant: Power Semiconductor Devices, Theory And Applications; 1999; John Wiley & Sons, Inc.

[2] CIGRE Working Group 14.17: Semiconductor Power Devices For Use In HVDC and FACTS Controllers; CIGRE Brochure 112, 1997

[3] Mohan, Undeland, Robbins: Power Electronics; 1995; John Wiley & Sons, Inc.

[4] Hingorani, Gyugyi: Understanding FACTS; 2000; IEEE Press.

[5] Vancers, Christofersen, Leirbukt, Bennett: A Survey Of The Reliability Of HVDC Systems Throughout The World During 2003-2004; CIGRE Report B4-202; Paris; 2006 Session.

[6] CIGRE Task Force 14.01.02 (Milan Cepek, Editor): Voltage and Current Stresses on Thyristor Valves for Static VAR Compensators; CIGRE Brochure 78, 1993.

[7] Knight, Young, Trainer: Relocatable GTO-Based Static Var Compensator for NGC Substations; CIGRE Report 14-106; Paris; 1998 Session.

[8] G. Reed, J. Paserba, T. Croasdaile, R. Westover, S. Jochi, N. Morishima, M. Takeda, T. Sugiyama, Y. Hamazaki, T. Snow, A. Abed: SDG&E Talega STATCOM Project – System Analysis, Design, and Configuration; Panel Session on FACTS Technologies: Experiences of the Past Decade and Developments for the 21st Century in Asia and the World; Proceedings of the IEEE PES T&D-Asia Conference and Exposition; Yokohama, Japan; October 2002.

[9] Renz, Keri, Mehraban, Kessinger, Schauder, Gyugyi, Kowalski, Edris: Word's First Unified Power Flow Controller on the AEP System; CIGRE Report 14-107; Paris; 1998 Session.

[10] Huang, Chen, Han, Du, Bhattacharya, Baran, Edris, Ingram, Atcitty: Emitter Turn-Off (ETO) Thyristor: An Emerging Power Semiconductor Switch With Lower Cost, Improved Reliability And Scalability Of FACTS Controllers; CIGRE Report B4-107; Paris; 2006 Session.

3-1

3 AC-DC CONVERTERS (OUTLINE)

Introduction

Chapter 3 will review the voltage-sourced converter with various implementations.

3.1 Operating Principles of Switching Converters

3.1.1 Concept of Switch Matrix

3.1.2 Input and Output Complementary Termination Requirement

3.1.3 Voltage- and Current-Sourced Converters

3.1.4 Line- and Self-Commutation of Switches

3.2 Role of AC-DC Converters in Transmission Systems

3.2.1 DC transmission

3.2.2 Asynchronous System Intertie

3.2.3 FACTS Controllers

3.2.4 Comparison of Converters for Transmission Applications

3.2.4.1 Voltage-Sourced vs. Current-Sourced

3.2.4.2 Self- vs. Line-Commutated

3.3 Voltage-Sourced AC-DC Converters

3.3.1 Basic Operating Principles

3.3.2 Methods of Output Voltage Synthesis

3.3.2.1 Pulse-Width Modulation (PWM)

A. Sinusoidal Modulation

B. Selective Harmonic elimination (Waveform “notching”)

3.3.2.2 Harmonic Neutralization (Multi-Pulse Structures)

A. 12-Pulse Converter

3-2

B. 24-Pulse Converter

C. 48-Pulse Converter

3.3.2.3 Multi-Level Approach with “Diode Clamp” Pole Structure

A. Three-Level

B. Five-Level

3.3.2.4 Multi-Level Approach with “Flying Capacitor” Pole Structure

A. Three-Level

B. Five-Level

3.3.2.5 Multi-Level Approach with “Chain Circuit” Structure

A. Three-Level

B. Five-Level

3.3.2.6 Hybrid Approaches and Structures (Harmonic Neutralization and/or Notching and/or Multi-Level and/or PWM)

3.3.3 Methods of Output Voltage Control

3.3.3.1 DC Voltage Control

3.3.3.2 Pulse-Width Variation

3.3.3.3 Pulse-Width Modulation

3.3.4 High Power Converter Structure Options and Trade-offs for High MVA Output

3.3.4.1 Single Large 2- or 3-Level Converter with Large Number of Series Connected Semiconductors

3.3.4.2 Several Single Device “Module” Converters in Parallel Connection

3.3.4.3 Several Single Device “Module” Converters in Series Connection

3.3.4.4 Hybrid Approaches: Few Module Converters with Modest Number of Series Connected Semiconductors

3.3.4.5 Practical Methods of Summing the Outputs of Module Converters

3.3.4.6 Comparison of Options

3.3.5 Converter Voltage and Current Rating Considerations and Issues in Transmission Applications

3.3.6 Voltage-Sourced Converter Protection Consideration and Issues in Transmission Applications

3.3.6.1 Output Current Limitation

3-3

3.3.6.2 Coupling Transformer Saturation

3.3.6.3 AC System Overvoltage

3.3.6.4 DC Capacitor Overvoltage

3.3.6.5 “Shoot-through”

3.3.7 Reliability, Availability and Redundancy Issues

3.4 Selected References

4-1

4 GUIDE FOR AND DEFINITIONS OF FACTS CONTROLLERS (OUTLINE)

Introduction

This chapter will provide generally used terminology and approved definitions that evolved in the relevant transmission controller literature, and that will be used in the Reference Book.

4.1 Basic Categories Determined by the Type of Semiconductor Valves Employed

4.1.1 Controllers Using Line-Commutated Conventional Thyristor Valves

4.1.1.1 Inherent Advantages

4.1.1.2 Inherent Disadvantages

4.1.2 Controllers Using Self-Commutated Semiconductor Valves

4.1.2.1 Inherent Advantages

4.1.2.2 Comparative Disadvantages

4.2 Categories by Functional Connection

4.2.1 Shunt-Connected Controllers

4.2.1.1 Thyristor-Controlled Type

A. Thyristor-Controlled Reactor

B. Thyristor-Switched Capacitor

C. Static Var Compensator (SVC)

D. Thyristor-Controlled Resistor

E. Thyristor-Controlled Voltage Regulator (Tap Changer)

4.2.1.2 Self-Commutated (Converter-Based) Types

A. Static Synchronous Compensator (STATCOM)

B. Static Synchronous Compensator with Energy Storage

4-2

4.2.2 Series-Connected Controllers


A. Thyristor-Controlled Series Capacitor (TCSC)

B. Thyristor-Controlled Series Reactor (TCSR)


A. Static Synchronous Series Compensator (SSSC)

B. Static Synchronous Series Compensator with Energy Storage

4.2.3 Combined Shunt- and Series-Connected Controllers


A. Thyristor-Controlled Voltage Regulator (TCVR)

B. Thyristor-Controlled Phase Angle Regulator (TCPAR)


A. Unified Power Flow Controller (UPFC)

4.2.4 Combined Shunt- and Shunt-Connected Controllers

4.2.4.1 Thyristor-Controlled Type: Line Commutated Back-to-BackConverter Asynchronous Tie

4.2.4.2 Self-Commutated Type: Self-Commutated Back-to-Back Converter Asynchronous or Synchronous Tie (BtB)

4.2.5 Combined Series- and Series-Connected Controllers

4.2.5.1 Thyristor-Controlled Type: None

4.2.5.2 Self-Commutated Type: Interline Power Flow Controller (IPFC)

4.2.6 Generic Characteristics of Thyristor-Controlled Type Controllers

A. Can Realize Controllable Shunt Admittance or Series Impedance

B. Reactive and Active Power Absorbed or Generated by Passive Components; Thyristor Valves Only Provide Control.

C. Controlled Active Power Exchange with the System is Inherently Associated with Reactive Power Absorption

4.2.7 Generic Characteristics of Self-Commutated Types

A. Can Realize Controllable Reactive Power Generator with Inherent Four- Quadrant Operation.

4-3

B. Variable Reactive Power Generation/Absorption is Accomplished by the Self- Commutated Controller without Passive AC Capacitors or Reactors.

C. Active Power Exchange with the System is Independent of the Reactive Power Generation or Absorption.

4.3 Categories by Main Application Function

4.3.1 Shunt Compensators: SVC and STATCOM and Related Combinations

4.3.2 Series Compensators: TCSC and SSSC and Related Combinations

4.3.3 Voltage and Angle Regulators: Thyristor Controlled Various Tap Changers

4.3.4 Multi-Function Power Flow Controllers: UPFC and IFFC

4.3.5 Back-to-Back System Interties


5-1

5 SHUNT COMPENSATORS (SVC AND STATCOM) (OUTLINE)

5.1 Objectives of Shunt Compensation

5.1.1 Voltage Regulation and Prevention of Voltage Collapse

5.1.2 Stability Improvement

5.1.3 Var Flow Management

5.2 Concepts for Controllable Var Generation

5.2.1 Thyristor-Controlled and Switched Capacitors and Reactors

5.2.2 AD-DC Converters

5.3 Thyristor Controlled/Switched Schemes for Static Var Compensators (SVCs)

5.3.1 Thyristor-Controlled Reactor (TCR)

5.3.1.1 Delay Angle Control

A. Basic Concept

B. Delay Angle vs. Fundamental and Harmonic Relationships

C. Methods of Harmonic Reduction

5.3.1.2 Basic Control Scheme for Output Control

5.3.1.3 Valve Rating and Protection

5.3.2 Thyristor-Switched Capacitor (TSC)

5.3.2.1 Issues of Capacitor Switching

5.3.2.2 Conditions for Repeated (Transient-Free) Switching



5.3.3 Fixed-Capacitor, Thyristor-Controlled Reactor (FC-TCR) Scheme

5-2

5.3.3.1 Basic Concept

5.3.3.2 Practical Circuit Arrangements with Harmonic Filter

5.3.3.3 System Voltage vs. Output Current Characteristic

5.3.3.4 Loss vs. Var Output Characteristic

5.3.4 Thyristor-Switched Capacitor, Thyristor-Controlled Reactor (TSC-TCR) Scheme

5.3.4.1 Basic Concept of and Reasons for the TSC-TCR Arrangement


5.3.4.3 Practical Circuit Arrangements

5.3.4.4 System Voltage vs. Output Current Characteristic

5.3.4.5 Loss vs. Var Output Characteristic

5.3.5 Hybrid Circuit Arrangements with Mechanically Switched Capacitors

5.3.5.1 Basic Concept

5.3.5.2 Features and Limitations

5.4 Voltage-Sourced AC-DC Converter Schemes for Static Synchronous Compensators (STATCOMs)

5.4.1 Operational Analogy to the Rotating Synchronous Generator

5.4.2 Functional Control Scheme

5.4.3 Methods of Output Control

5.4.3.1 Internal Control of DC Voltage

5.4.3.2 Pulse-Width Variation or Modulation

5.4.4 Converter and Filter Circuit Issues and Trade-offs

5.4.5 Converter Protection

5.4.6 System Voltage vs. Output Current (V-I) Characteristic

5.4.7 Loss vs. Var Output Characteristic

5.4.8 Output V-I Characteristic “Tailoring” for Application Requirements

5.4.9 Combined Converter and Thyristor Controlled Schemes

5-3

5.5 Compensator System Control Scheme for Network Shunt Compensation

5.5.1 Application Requirements for Reactive Shunt Compensation

5.5.2 Regulation Loop with Adjustable Slope

5.5.3 Transfer Function and Dynamic Performance

5.5.4 Special Control Functions

5.5.4.1 Var Reserve Control

5.5.4.2 Transient Stability Enhancement

5.5.4.3 Power Oscillation Damping

5.5.5 Practical Control System Structure

5.5.5.1 Interface with the Valves

5.5.5.2 Internal Operational and External System Functional Controls

5.5.5.3 Operation Supervisor, Diagnostic Control, and Status Monitor

5.5.5.4 SCADA and Local Display

5.6 Comparison Between SVC and STATCOM Schemes

5.6.1 V-I Characteristics

5.6.2 Compensation Performance Under Small System Disturbances

5.6.3 Compensation Performance Under Large System Disturbances

5.6.4 Operation Losses

5.6.5 Real Power Compensation and Exchange

5.6.6 Installation Size and requirements

5.7 Application Examples

5.7.1 SVC Installations and Experience

5.7.2 STATCOM Installations and Experience

5.8 STATCOM for Arc Furnace Compensation

5.8.1 Electrical Characteristics of Arc Furnace Loads and Corresponding Power System Disturbances

5.8.1.1 Fluctuating and Large MW and MVar Demand

5-4

5.8.1.2 Large and Changing Load Unbalance

5.8.1.3 Generation of Large Harmonic and Other Components with Non-Synchronous Frequencies

5.8.1.4 Voltage Fluctuation and Distortion Causing Lamp Flicker,Television Disturbance and Telephone Interference

5.8.2 Compensation Requirements

5.8.2.1 Mvar Rating

5.8.2.2 DC Energy Storage Capability

5.8.2.3 Speed of Response (Frequency Bandwidth)

5.8.2.4 Harmonic Attenuation/Filtering

5.8.3 STATCOM Design Features to Meet Requirements

5.8.3.1 Power Circuit Structure

5.8.3.2 Control Approach

5.8.4 Installation Examples and Performance Obtained


6-1

6 SERIES COMPENSATORS (TCSC AND SSSC) (OUTLINE)

6.1 Objectives of Series Compensation


6.1.2 Stability Improvement (Transient Stability and Oscillation Damping)

6.1.3 Power Flow Control

6.2 Concepts for Controllable Series Compensation

6.2.1 Thyristor-Controlled and Switched Capacitors

6.2.2 AD-DC Converters

6.3 Thyristor Controlled/Switched Schemes

6.3.1 Thyristor-Switched Series Capacitor (TSSC)

6.3.1.1 Principles of Operation

A. Switching Strategy and Restrictions

B. Compensating Voltage and Impedance vs. Line Current Characteristics

C. Losses vs. Line Current Characteristic

6.3.1.2 Basic Control Scheme


6.3.2 Thyristor-Controlled Series Capacitor (TCSC)

6.3.2.1 Principles of Operation

A. Blocking Mode

B. Bypass Mode

C. Control in the Capacitive and Inductive Operating Regions

D. Impedance Characteristics in the Subsynchronous Region

6-2

E. Compensating Voltage and Impedance vs. Line Current Characteristics

F. Loss vs. Line Current Characteristic

G. Harmonics

6.3.2.2 Basic Operating Control Schemes


6.3.3 Installation Considerations for the TSSC and TCSC

6.4 Voltage-Sourced AC-DC Converter Scheme

6.4.1 Basic Concept of the Static Synchronous Series Compensator (SSSC)

6.4.1.1 Compensating Voltage and Impedance vs. Line Current Characteristics

6.4.1.2 Loss vs. Line Current Characteristic

6.4.2 Functional Scheme for Compensation Voltage Control

6.4.2.1 Internal Control of DC Voltage

6.4.2.2 Pulse-Width Variation or Modulation

6.4.2.3 Subsynchronous Characteristics

6.4.3 Harmonics

6.4.4 System Control Scheme for Series Line Compensation Requirements

6.4.5 Converter Circuit Issues and Trade-offs

6.4.5.1 Transformer Coupling

6.4.5.2 Transformerless Implementation

6.4.6 Converter Protection

6.4.7 “Tailoring” the Symmetrical Capacitive and Inductive Operating Range to Application Requirements

6.4.8 Installation Considerations

6.5 Comparison Between the TCSC and SSSC

6.5.1 Compensation V-I and X-I Characteristics

6.5.2 Controllability


6.5.4 Real Power Compensation and Exchange

6-3

6.5.5 Installation Size and Requirements

6.6 Special Application and Control Considerations for Damping Electromechanical Oscillations

6.7 Application Examples

6.6.1 TCSC Installations and Experience

6.6.2 SSSC Installations and Experience


7-1

7 VOLTAGE REGULATORS AND PHASE SHIFTERS (OUTLINE)

7.1 Objectives of Voltage Regulators and Phase Shifters


7.1.2 Power Flow Control

7.1.3 Real and Reactive Loop Power Flow Mitigation

7.1.4 Stability Improvement (Transient Stability and Oscillation Damping)

7.2 Concepts for Voltage Regulation and Phase Shifting

7.2.1 Thyristor-Controlled/Switched Voltage Regulator (TCVR/TSVR) and

Phase Angle Regulator (TCVR/TSPAR)

7.2.2 AD-DC Converter-Based Schemes

7.3 Thyristor-Controlled Voltage Regulator

7.3.1 Basic Circuit Configurations

7.3.2 Delay Angle Control

A. Resistive Load

B. Inductive Load

C. Capacitive Load

D. Delay Angle vs. Fundamental and Harmonic Relationships

7.3.3 Basic Control Scheme

7.3.4 Limitations for High Power Applications

7.4 Thyristor-Switched Circuit Schemes for High Power Voltage Regulators and Phase Shifters

7.4.1 Tap-Changer Schemes with Equal Winding Sections

7-2

7.4.2 Tap-Changer Schemes with Ternary Proportioned Winding Sections

7.4.3 Basic Control Scheme

7.4.4 Valve Rating and Protection

7.5 Voltage-Sourced AC-DC Converter Scheme

7.5.1 Application of Back-to-Back Four Quadrant Converter Structure

(Note: See circuit, control, and application details at the UPFC Chapter)

7.5.1.1 In-Phase Voltage Injection (Voltage Regulator)

7.5.1.2 Quadrature Voltage Injection (Phase Shifter)

7.5.1.3 Voltage Injection at an Arbitrary Angle (UPFC)

7.6 Comparison Between the Thyristor Controlled/Switched and Converter-Based Voltage Regulator and Phase Shifter Schemes

7.6.1 Transformer and Power Circuit Rating

7.6.2 Reactive Power Demand

7.6.3 Protection Requirements


7.6.5 Installation Requirements


8-1

8 GENERALIZED AC TRANSMISSION CONTROLLERS: UNIFIED POWER FLOW CONTROLLER (UPFC) AND INTERLINE POWER FLOW CONTROLLER (IPFC) (DRAFT)

A generalized transmission controller could be considered as one that has the combined functional capabilities of dedicated line compensators and controllers. Two such controllers, the Unified Power Flow Controller (UPFC) and Interline Power Flow Controller (IPFC) are discussed in this chapter

8.1 The Unified Power Flow Controller (UPFC)

8.1.1 Concept and motivation

There are several different ways to introduce and explain the concept of the Unified Power Flow Controller (UPFC). These include mathematical approach, vector or phasor diagrams, various plots showing graphically relationships among the main transmission parameters, power (active and reactive), voltage, line impedance, and transmission angle. In a similar manner, the understanding of the UPFC, i.e., the visual concept one’s mind ultimately develops, is heavily dependent upon the individual’s established way of looking the operation and control of power systems based upon previous studies and experience. Considering that the aim of the present book is to primarily help the utility engineers and operators, the introduction of the UPFC concept, in contrast to the customary approach, is carried out from well established power transmission operating and control principles.

8.1.1.1 UPFC as the functional combination of conventional transmission controllers

Consider the simple case of transmitting power between Bus 1 and Bus 2, as shown in Fig. 8.1. Each of the two buses may be represented with a generator and source impedance. The two generators are in steady-state synchronism and at zero power transmission the magnitude of the two bus voltages are the same.

8-2

Figure 8-1 Two machine, single-line power system with corresponding phasor diagram and transmission equations

It is well understood from basic power transmission theory that the transmitted power P between Bus 1 and Bus 2, and the reactive power Q supplied from the busses, are dependent on the magnitude V of the bus voltage V (considered as a phasor or vector), the prevailing angle δ, between the voltages at the two buses, and the (reactive) line impedance X, as expressed by the simple equations:

δsin2

XVP ≅

(8.1)

)cos1(2

δ−−≅X

VQ (8.2)

If we assume for this example that the utilization of the conduction capacity of the line, and corresponding desired power transmission, are not achievable with existing line impedance and prevailing transmission angle (due to long distance and other interconnections in the overall system), then existing practice would suggest the well established solutions of applying series capacitive compensation to obtain the desired overall transmission impedance and, if the required angle still remains insufficient, a phase angle shifting transformer as well. Thus, it can be concluded that in a general case to ensure complete control of power flow in the line, both the line impedance and transmission angle may have to be controlled. However, further examining the possible transmission problems in the present example, let us also assume, that the source impedance behind Bus 1 is relatively large and the corresponding voltage drop constrains the desired power transmission (and also cause voltage related operating problems in the system). The existing practice would suggest the obvious solution of using a shunt compensator to provide the necessary voltage support for Bus 1.

X

V1I

V2

V

X

1V

I

2V

δ

Bus Bus

Line (reactive)

System impedance

System impedance

P

Q Q

)cos1(2

δ−−≅X

VQδsin2

XVP ≅

8-3

Thus, as illustrated in Fig. 8.2, the full control of a transmission line may require three independent controllers. It should be noted here that, although each of the three controllers is independent with regard to controlling its “own” parameter (i.e., impedance, angle, voltage), they may have an interaction on each other. For example, the action of changing the effective transmission impedance by series compensation, may also require an appropriate change in the overall transmission angle to get the desired power plow and vice versa.

Figure 8-2 Full transmission control: reactive shunt compensation for voltage, series compensation for impedance, and phase-shifting for angle control

The example so far considered assumed a steady-state condition for the power transmission. Under this condition the desired transmission can be achieved with some fixed capacitance/reactance for series and shunt compensation, as well as with some fixed angular adjustment. However, it is evident that under different transmission conditions, these values would require readjustment, and, in particular, under the conditions of dynamic disturbances, the stabilization of the system could necessitate the rapid controllability of these parameters. In the above example, on the basis of classical transmission control practice, we employed conventional reactive series and shunt compensation, as well as phase-shifting to achieve the required transmission control. Although the name of each action seems to imply the ways the required result is achieved, it is still worthwhile to examine the actual physical phenomena taking place with these to prepare for the understanding of the UPFC concept. Considering first the series capacitive compensation (Phase Shifter is assumed bypassed), the voltage between Bus 1 and Bus 2 is applied across the series connected compensating capacitor Cse and reactance of the line X. The line current produces a voltage across X that leads, and across the capacitor that lags, this current by 90 degrees. The sum of these two voltages must equal the prevailing voltage between the two buses, as illustrated by the simple phasor diagram in Fig. 8.3. It is to be observed that, in a physical sense, the series capacitor in effect increased the voltage

Xline

1V

I

2V

Bus 1 Bus 2 System impedance

System impedanceat Bus 2

P

Q Q

Cse

Vσ

+

+

--

ShiftePhase

2V’

CompensatShun

CompensatSerie

Csh Lsh

(σ )

P V’ = V 1 2 Xline-XCse

1sin (δ±σ)

XCse

8-4

across the existing physical line reactance, and thereby increased the line current, and the transmitted power, in the same way, as if the line reactance had been reduced. This observation leads to the conclusion that series capacitive compensation in the actual physical sense achieves power flow control by changing, i.e., increasing, the voltage across the existing line reactance, and not the reactance itself. This means that the same end objective of power flow control can be achieved by any practical means that can increase (or decrease) the magnitude of the voltage across the transmission line. An additional observation needs also to be noted, i.e., the voltage across the line reactance is in quadrature with the line current, and therefore the adjustment of this voltage, whether achieved by a series capacitor or other means, does involve only an exchange of reactive power, but no active power.

Figure 8-3 Series capacitive line compensation and corresponding phasor diagram

Considering now the operation of the phase-shifter (bypassed series compensation), the voltage across the line is the (vectorial) sum of three voltages. To obtain the desired power transmission, the voltage at Bus 1 is modified to that at Bus 1’, to which the line is actually connected (see Fig. 8.4). Thus, the voltage across the line is the voltage between Bus 1’ and Bus 2. The Phase Shifter shown in Fig. 8.4 is a conventional transformer type which can inject, via the insertion transformer, a voltage composed of two components, one in phase, the other in quadrature with the voltage at Bus 1. This is necessary to maintain the magnitude of the phase-shifted voltage at the required value independent of the amount of phase-shift applied. (It is to be noted that this arrangement would also facilitate the control of the bus voltage while maintaining the phase shift at a fixed value.) The physical operation of the Phase Shifter leads again to the same conclusion as reached for series line compensation, that is, the phase shifting needed to obtain the transmission angle necessary to achieve the desired power flow is done by injecting a specific voltage between one of the two buses and the line, to establish the needed voltage across the line. It is evident that, from the standpoint of this objective, the means by which the voltage injection is accomplished is unimportant.

X

V1I

V2

XCse X

1 V

I

2

δ

Cse

V

V

V

Bus 1 Bus 2

8-5

Figure 8-4 Phase-shifting for transmission angle control and corresponding phasor diagram

Besides the conclusion that power flow control using series reactive compensation or phase shifting is essentially achieved by the injection of an appropriate voltage between the line and the “sending-” and/or “receiving-end” buses of the power system, there are some important differences between the two approaches. One important difference is that series (capacitive) compensation by definition is reactive, the injected voltage is always in quadrature with the line current. By contrast, phase shifting injects a voltage to establish and maintain the desired transmission angle, and therefore this voltage has no predefined relationship with the line current. Consequently, the Phase Shifter does, in general, exchanges both active and reactive power with the system. Thus, whereas the series capacitive compensation is self supporting (the capacitor supplies the reactive power that the line reactance absorbs), the transformer type phase sifter has no capability to this and both the reactive and active power exchanged in the process of phase shifting must be supplied externally, usually by the system itself. (In practice, shunt capacitive compensation is often applied to supply the reactive power.) In comparing further the Phase Shifter to the series reactive compensator, the deduction can readily be made that the generalized Phase Shifter with combined in-phase and quadrature voltage injection capability could perform the specific function of series compensation by keeping the injected voltage in quadrature with the line current (of course, without supplying the necessary reactive power itself). Moreover, because of the two dimensional voltage injection capability of the generalized Phase Shifter, the more general deduction that the such Phase Shifter could actually provide both functions of series compensation and angle control by injecting the resultant voltage, i.e., the sum of that the two controllers would separately provide to satisfy the specific power transmission, as illustrated in Figs. 8.5a and 8.5b. (Fig. 8.5a shows an example for combined series compensation and phase shift, and 8.5b illustrates the case when the same transmission line condition is accomplished by the Phase Shifter alone.) Note, however, that in the (b) case the reactive power, previously supplied by the series capacitor, is supplied now by the system from Bus 1.

V σ

V

-V σ

δ

δ+σ

+σ −σ δ−σ

V

V x σ = 0 ( )

(1

1V’ )+σ

−σ (xV )

2 ( ) −σ

+Vσ

+σ ( xV )

X

1V

I

2V

+

+

--

1V’ 1V’

shiftePhase

Bus Bus Bus

8-6

Figure 8-5 a. Series capacitive compensation plus phase-shifting; b. identical transmission end-result with the phase shifter providing independent two-dimensional (in-phase and quadrature) voltage injection

In spite of the attractive functional flexibility of the generalized Phase Shifter for power flow control, it has very rarely been applied in practice. (Even when it was, the in-phase injection capability was used separately for bus voltage control and the quadrature one for angle regulation.) There are several reasons for this: the multi-transformer arrangement with taps is complex and difficult to realize for high voltage levels; the tap-changer gear is slow and requires regular maintenance; since it has no internal capability to supply reactive power, the total VA the arrangement exchanges through the series voltage injection with the line, must be provided at the primary side by the system itself, which, due to the possible high reactive power content, can make the overall transformer rating and reactive power flow in the primary-side supply line high, usually necessitating a separate reactive power compensator. For these reasons, it has been realized that a single, generalized power flow controller capable of controlling all the parameters of determining power transmission (voltage, impedance, angle) selectively or in combination, would require an implementation with inherent reactive power generation/absorption capability. A conceptual model for the generalized power flow controller with inherent reactive power generation capability can be derived once again with a conventional power system component, the rotating synchronous compensator (or, with the traditional name, synchronous condenser) or, in a more general form, the synchronous machine. Such a model is illustrated in Fig. 8.6.

Vσ

V

-V σ

δ

−σ δ−σ

V1 2

X

1V

I

2V

+

+

--

1V’

1V’

shiftPhas

X

CV xV

x,noV

1V’’

1V’’

Vσ

V

-Vσ

δ

δ−σ

V 1 2

X

1V

I

2V

+

+

--

1V’

shiftPhas

xV

x,noV 1V’

(a) (b)

8-7

Figure 8-6 Two-machine synchronous power flow controller model with internal reactive power generation for self-sufficient series and shunt reactive compensation, and corresponding phasor diagram

To understand the operation of this model, one has to recall that a synchronous machine can be operated both as a generator and as a motor. In the former case, the machine receives mechanical input power and it converts to electrical output power; in the latter case, its input is electrical power that it converts to mechanical output power. Independently whether it is operated as a generator or as a motor, it can generate or absorb reactive power (var) by excitation control (var generation by over excitation; var absorption by under excitation). If two synchronous machines are mechanically coupled to run at a common synchronous speed, as shown in Fig. 8.6, then controllable active electrical power can from one electrical terminal to the other in either direction. At the same time, regardless of the direction of active power flow, reactive power can be generated or absorbed at the two terminals independently of each other.

The operation of the two-machine model is illustrated by the phasor diagram in Fig. 8.6b. Machine 2 inserts a voltage Vpq between the Bus 1 and the line at Bus 1’. The voltage across the line, VX, the line current I, and the transmitted power is thus determined by the voltage between Bus 1’ and Bus 2. The magnitude of the inserted voltage Vpq is fully controllable from zero and a maximum, and its phase between 0 and 360 degrees, relative to the voltage at Bus 1. This inserted voltage, in general, result in an exchange of both active and reactive power with Machine 1 and the line. The reactive power exchanged is generated by Machine 2 itself. However, the active power is supplied by the system at Bus 1 via Machine 1 that, as a motor, drives Machine 2. In addition, Machine 1 can generate (or absorb) reactive power, if needed, to regulate the voltage at Bus 1.

From the previous discussion it should be evident that the angularly unconstrained voltage injection can not only emulate the operation any one of the traditional transmission compensators/controllers used for voltage regulation, series compensation and angle control (for an example, see Fig. 8.5), but provide an equivalent control action as if all three of those would be separately operated by synthesizing the resultant voltage of the three components they would separately provide.

V X

1V

I

2V1V’

Synchronous Power Flow Controller Model

Voltage control

Voltage control

V VV x1 1V’

x V

2

δ

oδ

ρ

o

p

Machine 2

Machine 1

(a) (b)

Vp

Bus 1 Bus 2Bus 1’

8-8

It should be observed that if the generalized power flow controller formed from two synchronous machines is mechanically decoupled, then its operation is restricted to that of two independent, synchronous, reactive compensators, one connected in shunt and the other in series with the line.

The synchronous power flow controller model, shown in Fig. 8.6, has not been realized practice due probably to cost, possible operating difficulties under dynamic conditions, maintenance requirements, and also to relatively slow response to handle disturbances. However, this model can be eminently realized with fast response and unprecedented control flexibility by static, high power converters using semiconductor switches. Such a realization is shown schematically in Fig. 8.7, and called Unified Power Flow Controller (UPFC). The operation of this arrangement is analogous to that of the rotating synchronous machines: Converter 2 is used to inject the compensating voltage, Vpq, in series with the line, which is in synchronism with the ac system, but its magnitude and phase angle with respect to Bus voltage V1 is freely variable between zero and a maximum, and between 0 and 360 degrees, respectively. Converter 2 is able to internally provide the reactive power exchanged as a result of the voltage injection, but the active power must be provided to it, or absorbed from it. This function is provided by Converter 1, which absorbs/supplies active power from the system at Bus 1, and delivers it to Converter 2 via the common dc link. Converter 1 can also provide reactive shunt compensation by appropriately controlling both the magnitude and phase angle of its voltage. (Recall the operation of the voltage-sourced ac to dc converters discussed in a previous Chapter.)

Figure 8-7 Realization of the Unified Power Flow Controller (UPFC) with two, back-to-back connected voltage-sourced converters providing identical characteristics to that of the synchronous power flow controller model with fast response time

8.1.1.2 UPFC directly providing line current forcing function

The ultimate function of a transmission controller is to optimize or control the power transmission over a transmission line. In this sense, one could argue that the conventional controllers, such as Phase Shifters, series capacitors, etc., really aim to vary the line current in order to control the transmitted power by means of changing a particular transmission parameter, such as the transmission angle or line impedance, to accomplish indirectly this objective. The

V X

1V

I

2V1V’

Unified Power Flow Controller using back-to-back converters

AC to DCConverte V VVx

1 1V’

x V

2

δ oδ

ρ

o

p

(a) (b)

Vp

ContrV d

Transmission parameters

AC to DCConverte

Bus Bus Bus

8-9

UPFC by its multifunctional capability can, with sufficient rating, directly force the line current, with the use of an appropriate closed-loop control, to assume an angle and magnitude with respect to a given bus voltage to deliver the desired active power and, simultaneously, minimize the reactive power flow in the line. In other words, the UPFC has the functional capability of directly, and independently, controlling the active and reactive power flow in the line, as illustrated schematically in Fig. 8.8.

Figure 8-8 Unified Power Flow Controller providing closed-loop control of the reference defined active and reactive power flow

Conclusions on the UPFC functional capability: 1. The UPFC functional capability is equivalent of the combined functional capabilities of the Phase Shifter, controllable series reactive compensator, and controllable reactive shunt compensator. The UPFC can provide these functions separately, or in combination. The practical UPFC employs solid-state switches, thus its functional and dynamic performance equals, or usually exceeds, those of the Thyristor Controlled Phase Shifter, Thyristor Controlled Series Capacitor (TCSC), and Static Var Compensator (SVC). 2. The UPFC can be viewed as a controller with the unique functional capability to force the line current so as to establish the desired active and reactive power flow in the line.

8.1.2 Basic characteristics of the UPFC

In a conventional uncompensated transmission line connecting two buses with equal voltages, the relationship between the active power P, reactive line power Q, and the transmission parameters bus voltage V, line impedance X, and transmission angle δ, can be described by two simple equations, as given in (8.1) and (8.2). These equations result in simple curves when P and

V X

1V

I

2V1V’

Unified Power Flow Controller

pq

AC to DC Converter 2

Control V d

AC to DC Converter 1

P

PrefQ ref

1V 1V’ I

8-10

Q plotted against the usually only controllable variable, the transmission angle δ, showing graphically the variation of the active and reactive line power flow with the transmission angle. Similarly, a single curve graphical relationship can also be established between the active and reactive line power.

When a transmission line is controlled by a UPFC, the mathematical and corresponding graphical relationships become somewhat more complex, because the UPFC can vary the conventionally fixed transmission parameters, the magnitude and angle of the transmission voltage as well as the line impedance, in addition to the effective transmission angle. Thus the simple curves of the uncompensated transmission become operating areas or regions, the boundaries of which are defined by the rating or control imposed restrictions of the UPFC. These characteristic curves will be presented here in a manner analogous to those of conventional power transmission for easy understanding, and also for the visual illustration of the hitherto unattainable transmission control capability of the UPFC.

The mathematical expressions for the active and reactive power for the UPFC controlled line can be expressed in an analogous form of those give by Equations 8.1 and 8.2, that is:

⎟⎠⎞

⎜⎝⎛ ++= ρδδ

2sinsin

2

XVV

XVP pq (8.3)

and

( ) ⎟⎠⎞

⎜⎝⎛ ++−−= ρδδ

2sincos1

2

XVV

XVQ pq (8.4)

where Vpq is the amplitude and ρ is the angle (with respect to the Bus 1 voltage, V1) of the Vpq voltage injected by the UPFC in series with the line (refer to Fig. 8.8). It can be seen that Equations 8.3 and 8.4 contain the customary terms of the uncompensated line given by Equations 8.1 and 8.2, plus an additional term that is a function of both the magnitude and angle of the injected voltage Vpq. Recalling that angle ρ is freely variable between 0 and 360 degrees (and is independent of the MVA rating of the equipment), a plot showing the expansion of the single P versus δ curve into an operating band, the upper and lower boundary of which is determined my the magnitude Vpq (i.e., the MVA rating of the UPFC), can be constructed for both P and Q, as shown in Fig. 8.9. Inspection of Fig. 8.9a and 8.9b shows that at any transmission angle δ the

8-11

Figure 8-9 Range of transmittable active power P and receiving-end reactive power demand Q, versus the transmission angle δ

active power P or the reactive power Q can be increased or decreased by amount VVpq/X, which in the illustration is equal to 0.5 p.u. Mw or Mvar, since Vpq/V=0.5. (Note that this ratio in practical applications would be rather large, requiring a UPFC with high Mva rating.) It is important to realize that the possible maximum increase in power is a fixed amount, which is independent of the operating point defined by the transmission angle. This means that as the transmission angle is decreased from 90 degrees (the theoretical operating point for maximum power transmission), the percentage capability of the UPFC to increase the transmitted power continuously increasing. Thus at small transmission angles, a UPFC with relatively low Mva rating can provide highly effective power control. This particular characteristic of the UPFC (similarly to the SSSC discussed elsewhere) is due to its inherent capability to inject and maintain a fixed series compensating voltage in face of a variable line current (in contrast to, for example, the series capacitor, which provides a fixed compensating impedance).

Although Equations 8.3 and 8.4 indicate that both the active and reactive line power can be controlled over the same range, and increased or decreased by the same amount, they do not imply that the maximum values are simultaneously attainable. Simple physical reasoning suggests this, since control of active line power flow requires quadrature voltage injection with respect to the line current, whereas that of the reactive line power is accomplished by in-phase voltage injection. The overall control capability of the UPFC with regard to simultaneous active and reactive line power flow control can best be illustrated by a P versus Q plot, which show the corresponding active and reactive power values (in per unit) over the transmission angle range of interest. The starting point for the derivation of such a plot is again the two simple Equations, 8.1

VV

0.0

0.5

π/2 δ π

(=0.5)

(=-0.5)

VV pq

X

X

pq

-Q

1.0

π/2 0.0

-0.5

0.5

δπ

1.5

P

P (δ )

) o

pq

pq

X

X

VV

VV

(= -0.5)

(=0.5)

-1.0

-0.5

-1.5

-2.0

Q (δ) o

(a) (b)

8-12

and 8.2, characterizing the uncompensated transmission line. Normalizing these equations by making V2/X=1, and by summing the squares of them, the following expression can be obtained:

(Qo+1)2+Po 2 = sin2δ + cos2δ = 1 (8.5)

where the subscript zero indicates that Po and Qo represent an uncompensated transmission. Equation (8.5) describes a circle with a radius of 1.0 around the center defined by coordinates Po=0 and Qo=-1 in a Qo,Po plane, as illustrated for positive values of Po in Fig. 8.10a. Each point of this circle gives the corresponding Po and Qo values of the uncompensated system at a specific transmission angle δ. For example, at δ=0, Po=0 and Qo=0; at δ=30°, Po=0.5 and Qo=-0.134; at δ=90°, Po=1.0 and Qo=-1.0; etc.

Figure 8-10 a. Reactive vs. active power of the uncompensated line in the Q,P plane at transmission angle range of 0≤δ≤180; b. Control range of the UPFC in the corresponding Q,P plane

Inspection of Figs. 8.9a and 8.9b suggest that with the maximum positive value of Vpq, the radius of this Po versus Qo semi-circle would increase from 1 to (1+Vpq/V), and the maximum negative value would decrease it by that amount. (In the present example used in Fig. 8.9, Vpq/V=0.5 and thus (1+Vpq/V) is equal to1+0.5=1.5.) Of course, these values are obtained at particular values of ρ, the angle of voltage Vpq, which is freely variable from 0 to 360 degrees. The operation of the UPFC could be visualized in the Q,P plane by correlating the voltage phasor diagram shown in Fig. 8.7b to the P versus Q plots shown in Fig. 8.10. The transmission angle δ defines a related Po and Qo value pair for the uncompensated line with given transmission parameters. These Po and Qo value pairs define a semi-circular curve in the Q,P plane as δ varied from zero to 0 180

-

-

-

-

Q -

=

180° δ =

δ

0.

0.

0.

δ =0 1.

δ

Po

90°

= δ 30°

= δ 60°

o

-

-

-

-

Q-

=

180° δ =

δ

0.

0.

0.

δ =01.

δ

P

90°

-

1.

δ =30°

ρ

(Qo+1)2 + Po2 = 1 (Qo+1)2 + Po

2 = 1

(a) (b)

8-13

degrees. When compensating voltage Vpq is injected between Bus 1 and the line, the uncompensated Po and Qo is pushed off the defined semi-circle curve in the Q,P plane into a location that is determined by the magnitude and angle of Vpq. It is evident that as Vpq.is rotated over the full range of zero to 360 degrees, the P and Q value pair will move along a full circle around the point in the Q,P plane defined by the original Po and Qo value pair of the uncompensated line. It is further evident that, as the magnitude of the compensating voltage varied between zero and its maximum value, a circular area around the original Po and Qo value pair defined point will be established, in which any P and Q can be established by setting the amplitude and angle of Vpq to corresponding values. Changing the transmission angle δ will, of course, move the original Po and Qo values along the semi-circular curve, and around this new operating point the UPFC will establish its own circular operating are, expanding it into an eventual operating band around the uncompensated semi-circle curve, as illustrated in Fig. 8.10b.

The operating circular area of the UPFC with a center defined by coordinates Po and Qo and a radius of VVpq/X in the Q,P plane can be expressed mathematically from Equations 8.3 and 8.4:

( ) ( ) 2

max22

⎭⎬⎫

⎩⎨⎧

=−+−X

VVQQPP pq

oo ρρ (8.6)

This circular control region defined by equation (8.6) is shown in Fig. 8.10b for V=1.0, Vpqmax=0.5, and X=1.0 (per unit or p.u. values) with their centers on the semi-circular arc characterizing the uncompensated system (Equation 8.5) at transmission angles δ=30o. It can be observed in the illustrated example, that the UPFC would be able to increase the transmitted active power P almost by a factor of two (or reduce it to zero) and, at the same time, maintain zero reactive power flow in the line. It is mentioned again that to accomplish this, the UPFC would need a significant Mva rating (equal to half of the power transmitted by the uncompensated line), but considerably smaller rating could still provide significant control for the active and reactive power flow in the line.

Conclusions on the fundamental UPFC characteristics:

The UPFC is primarily a power flow controller. Its operation can best be characterized in the Q,P plane with an operating area around the basic operating point defined by the transmitted active power P and reactive line power Q obtained with the uncompensated line at the desired transmission angle. Within this, theoretically circular operating area (which is primarily determined by the MVA rating of the equipment, but may also be influenced by operating constraints), the UPFC can independently control the active and reactive power flow in the line.

8.1.3 Comparison of the UPFC control characteristic to that of other power flow controllers: series reactive compensator and Phase Shifter

8.1.3.1 UPFC versus reactive series compensators

Controlled series compensators provide a series compensating voltage that, by definition, is in quadrature with the line current. Consequently, they can affect only the magnitude of the current

8-14

flowing through the transmission line. As discussed in other parts of the reference book (see Chapter 6), there are various types of series compensators using different solid-state circuit configurations to emulate variable capacitors or provide quadrature voltage injection directly. However, whatever technique is used, at any given setting of the effective capacitive/inductive impedance, an overall transmission impedance is defined at which the transmitted power is strictly determined by the transmission angle (assuming a constant amplitude for the end voltages). Therefore, the reactive power demand at the end-points of the line is determined by the transmitted active power in the same way as if the line was uncompensated but had lower impedance. Consequently, the relationship between the transmitted power P and the reactive-power demand of the line can be represented by a single Q-P circular locus, similar to that shown for the uncompensated system in Fig. 8.10a. This means that for a continuously controllable compensator an infinite number of Q-P circular loci can be established by using the basic transmission relationships:

δsin2

qXXVP−

≅ (8.7)

)cos1(2

δ−−

−≅qXX

VQ (8.8)

where the effective compensating reactance Xq is varying between 0 and Xqmax. (Xq=XC for series compensators emulating a variable capacitor, and Xq=|Vq/I|=Vq/I for compensators executing quadrature voltage injection.) Evidently, a given transmission angle defines a single point on each Q-P locus obtained with a specific value of Xq. Thus, the progressive increase of Xq from zero to Xqmax could be viewed as if the point defining the corresponding P and Q values at the given transmission angle on the first Q-P locus (uncompensated line) moves through an infinite number of Q-P loci representing progressively increasing series compensation, until it finally reaches the last Q-P locus that represents the system with maximum series compensation. The first Q-P locus, representing the uncompensated power transmission, is the lower boundary curve for the type of series compensators that emulate a variable series capacitor and identified by (Q-P)Xq=0. The last Q-P locus, representing the power transmission with maximum capacitive impedance compensation is identified by (Q-P)Xcmax. The lower and upper boundary curves for the type of series compensators that directly inject quadrature compensating voltage are different and therefore they are identified by (Q-P)+Vqmax and (Q-P)-Vqmax. The difference is partially due to their capability to inject the compensating voltage both with 90° lagging (capacitive) or 90° leading (inductive) relationship with respect to the line current. Thus, they can both increase and decrease the transmitted power. Also, they can usually maintain maximum compensating voltage with decreasing line current (theoretically, even with zero line current). For these reasons, the series compensators with quadrature voltage injection capability have a considerably wider control range at low transmission angles than those emulating variable series capacitors. Note that all Q-P circular curves are considered only for the normal operating range of the trans-mission angle (0≤δ≤90o).

The plots in Fig. 8.11 correspond to the plots shown for the UPFC in Fig. 8.10 at δ=30° transmission angle. Both types of series reactive compensators have the same 0.5 p.u. maximum voltage rating stipulated for the UPFC. The previously derived circular control region of the

8-15

UPFC is also shown in Fig. 8.11 for comparison. In each figure the upper and lower boundary curves identified above for the variable capacitor type and voltage injection type compensators.

Figure 8-11 Attainable Q and P values with controlled capacitive compensation (points on the heavy green/blue straight line inside the circle) and those with the UPFC (any point inside the circle) at δ=30°

Inspection of Fig. 8.11 indicates that the relationship between Q and P, for variable capacitor type series compensator a straight line (shown in blue in the figure) connecting the two related points on the lower and upper boundary curves obtained at Xq=0 and XCmax, which represent the power transmission at δ=30° with zero and, respectively, maximum series compensation. Series compensators with direct voltage injection capability can provide maximum compensating voltage, theoretically down to zero line current. They can also reverse the polarity of the compensating voltage to reduce the power. For these reasons, this type of compensators can provide compensation over a wider range, between the boundaries (shown in green in the figure) determined by +Vqmax and -Vqmax, than can the variable capacitor types. However, although they lengthen the attainable control interval, their basic control characteristic, fundamental to all series reactive compensators, remains unchanged: active power increase or decrease in the line inevitably associated with a proportional change of reactive line power, independent active and reactive power flow control is not possible.

Fig. 8.11 vividly illustrated the superior power flow control characteristic of the UPFC in that providing full control for the transmitted active power and, at the same time, also able to control reactive line power flow, and thereby facilitating the maximum utilization of transmission assets and minimizing losses.

8.1.3.2 UPFC versus Phase Shifters

As discussed in other parts of this reference book, the function of the Phase Shifter (also called Phase Angle Regulator or PAR), is to change the angle of a given bus voltage to make it

Controllable interval of Variable capacitor

δ

qX = 0

=30°

Controllable region of UPFC

qmaV

=0° δ

-

Q

=90° δ

P

-

-

qmaxV+X Cma

Controllable interval of Quadrature voltage source

8-16

compatible with prevailing transmission requirements. Ideal Phase Shifters provide voltage injection in series with the line so that the voltage applied at their input terminals appears with the same magnitude but with ±σ phase difference at their output terminals. For simplicity, mechanical and thyristor-controlled Phase Shifters provide controllable quadrature voltage injection (“Quadrature Boosters”) in series with the line and thus the magnitude of their phase-shifted output voltage increases with increasing angular change. These Phase Shifters employ a shunt-connected regulation (excitation) transformer with appropriate taps on the secondary windings, a mechanical or thyristor-based tap-changer, and a series insertion transformer. The excitation transformer has wye to delta (or delta to wye) windings and thus the phase to phase secondary voltages are in quadrature with respect to the corresponding primary phase to neutral voltages.

Apart from their different power flow control characteristic, and in spite of their similarity in employing both a shunt-connected and series connected transformer for coupling their inputs and outputs to the transmission system, there is a fundamental, and from the application standpoint important, difference between the above (mechanical and thyristor-controlled) Phase Shifters and the UPFC. In the case of the former, the total VA (i.e., both the active power and the vars), exchanged by the series insertion transformer appears at the primary of the regulation transformer as a load demand on the power system. However, in the latter case only the active power exchanged is supplied by the shunt transformer from the power system, because the reactive power exchanged is generated by the UPFC converter itself.

The purpose of the present investigation is to compare the power flow control capability of the UPFC to that of the Phase Shifter. For this comparison, the practical implementation of the Phase Shifter and its ability to generate reactive power is unimportant. For simplicity, Phase Shifters are considered as “ideal”, i.e., their “quadrature boosting” is ignored, and thus they are assumed to be able to vary the phase angle between the voltages at the two ends of the insertion transformer in the control range of -σmax ≤ σ ≤ σmax, without changing the magnitude of the phase shifted voltage from that of the original line voltage.

As previously established, the transmitted power and the reactive power demands at the sending-end and receiving-end can be described by relationships analogous to those characterizing the uncompensated system:

)−≅ σδsin(2

XVP

(8.9)

)cos(12

σδ −−−≅X

VQ (8.10)

It follows from these equations that Phase Shifters cannot increase the maximum transmittable power, P=V2/X, or change Q at a fixed P. Consequently, the Q-P relationship, with transmission angle δ'=δ−σ (0≤δ'≤90°) controlling the actual power transmission, is identical to that of the uncompensated system. The function of the Phase Shifter is simply to establish the effective transmission angle, δ', required for the transmission of the desired power P, by appropriately adjusting the phase-shift angle σ. In other words, the Phase Shifter can vary the transmitted power at a fixed δ, or maintain the actual transmission angle δ' constant in the face of a varying

8-17

δ, but it cannot increase the maximum transmittable power or control the reactive power flow independently of the active power.

The plots in Fig. 8.12 show the Q-P relationship for the ideal Phase Shifter in comparison to that of the UPFC at δ=30°, the transmission angle used in the previous examples of transmission control characteristic. It can be observed that the control range centers around the point defined by the given angle between the two ends of the line (δ=30°) on the Q-P locus characterizing the uncompensated power transmission (Vσ=0 for the Phase Shifter and Vpq=0 for the UPFC). As the phase-shift angle σ is varied in the range of -30° ≤ σ ≤ 30° (which corresponds to the maximum inserted voltage of Vσ=0.5 p.u.), this point moves on the uncompensated Q-P locus in the same way as if the fixed angle δ was varied in the range of (δ-30°) ≤ δ ≤ (δ+30°).

Figure 8-12 Attainable Q and P values with phase-shifting (points on the heavy blue arc inside the circle) and those with the UPFC (any point inside the circle) at δ=30

The Q versus P plots in Fig. 8.12 clearly show the superiority of the UPFC over the Phase Shifter for power flow control: The UPFC has a wider range for active power control and facilitates the independent control of the reactive power demand of the line over a broad range. For example, it is seen that the UPFC can facilitate up to almost 1.0 p.u. active power transmission with unity power factor at the receiving-end (Q=0), whereas with phase-shift the reactive power demand at the receiving-end would increase with increasing active power (Q would reach 0.5. p.u. at the end of the Phase Shifter’s control range, at about P=0.9 p.u.).

Conclusions on the power flow control characteristic of the UPFC in comparison to the Series Reactive compensator and the Phase Shifter.

1. The control capability of the UPFC can be characterized in the Q,P plane by a (rating dependent) circular region, in which all points define an active and reactive power transmission value (i.e., to every P value there is an infinite number of Q within the boundaries of the region). The UPFC can operate the transmission line at any of the P and the corresponding Q values

Controllable interval of Phase Shifter

δ =30°

Controllable region of UPFC

σmaV =0° δ

-

Q

=90° δ

P 0.5 1.0 1.5

0.5

-0.5

-1.0

0.0

σmaV

8-18

within this region, in other words, the UPFC can independently control the active and reactive power flow within its control region.

2. Both the Series Reactive Compensator and the Phase Shifter define a (different) Q-P locus (which functionally defines a single Q value to each value of P), within the control region of the UPFC, over which the active power can be controlled. However, since the control in the Q,P plane is restricted to a given locus, the reactive power is functionally coupled to the active power, i.e., with increasing active power, the reactive line power will also increase.

3. From the application viewpoint, the UPFC is a general power flow controller with unique capability to control the transmitted active power independently of the reactive line power. It also has a multi-functional capability to provide specific, and changeable, control roles for the transmission system, if contingencies would so require. By contrast, the Series Reactive Compensator primary role is to control power transmission in lines where the natural line impedance is not compatible with transmission line power capacity. Similarly, the Phase Shifter is a specific controller for applications in which the system provided transmission angle is either not compatible, or not stable, for executing the desired power transfer.

8.1.4 UPFC control system

The superior operating characteristics of the UPFC, described in the previous sections, are due to its unique ability to inject an ac compensating voltage with variable magnitude and angle in series with the line to independently control the active and reactive components of the line current, and thereby the transmitted active power and the reactive power demand of the line. Apart from establishing and maintaining the desired power flow, the inherent characteristic features of the UPFC offer a great potential to effectively counteract dynamic system disturbances if its operation is controlled by a suitable electronic control systems. With this, the UPFC can cause the series-injected voltage to vary rapidly and in magnitude and angle as desired to increase transient stability and provide highly effective oscillation damping.

The control of the UPFC is based upon the so-called “vector-control”, which was originally developed for controlling converter supplied ac motor drives, but today it is generally used for transmission compensators and controllers employing the converter-based approach. The term vector, instead of phasor, is used in this type of systems to represent a set of three instantaneous phase variables, voltages or currents, that sum to zero. The symbols v and i are often used for voltage and current vectors. Since these vectors represent instantaneous phase variables, they are not stationary, but move around a fixed point in the plane as the values of the phase variables change, describing various trajectories. It is important to note that these trajectories become circles when the phase variables represent a balanced, steady-state condition, which means that under these conditions the vectors become phasors that are well known in traditional power engineering. Thus, utility (and other) engineers who may not be familiar with “vector- control”, may consider vectors as a phasors (mentally changing the low-case letters into capital letters, like v to V, i to I, p to P, q to Q, etc., without affecting the understanding of the basic control principles and structures. (This simplifying replacement is used in the illustration of the basic control system structure in Fig. 8.13.)

8-19

Figure 8-13 Overall UPFC control structure with separated Internal converter control and System operation control functions

For the purpose of power control it is useful to view these vectors in an orthogonal coordinate system with p and q axes such that the p axis is always coincident with the instantaneous voltage vector v and the q axis is in quadrature with it. In this coordinate system the p-axis current component, ip, accounts for the instantaneous active (real) power and the q-axis current component, iq, for the reactive power. Under balanced steady-state conditions, the p-axis and q-axis components of the voltage and current vector are constant quantities. This characteristic of the described vector representation makes it highly suitable for the control of the UPFC by facilitating the de-coupled control of the active and reactive current components.

The UPFC control system may be divided functionally into internal (or converter) control and functional (or system) operation control, as illustrated in Fig. 8.13. The internal controls operate the two converters so as to produce the commanded series injected voltage for power flow control and, simultaneously, draw the desired shunt reactive current for bus voltage regulation. The internal controls provide gating signals to the converter valves so that the converter output voltages will properly follow the internal reference variables, ipRef, iqRef and vpqRef.

The basic structure for internal (or converter) control is shown in Fig. 8.14. The functions of the two converters are different, and thus they are controlled according to their functional roles. The role of the series converter is to execute the voltage injection as commanded. Accordingly, the series converter, being supplied from the shunt-converter regulated dc source, can respond

d

ParameteSetting

Internal converter control

Shunt transforme

A+ V

Transmission

Converter 1

A

Series transforme

Converter 2 (series)

IV p

V1 V’ 1

I s

I s

VpqRef

V’ 1

V 1

I qRef

IVdc

System variables: P,Q,V1,V’1, etc.

System operation control

Gate

Breake P

Q

8-20

directly and independently to the demand for series voltage vector injection. Changes in series voltage vector, vpq, can therefore be affected virtually instantaneously. The role of the shunt converter is to maintain the dc voltage level to support the series voltage injection (by providing the necessary active power) and reactive shunt compensation demand. Thus, in contrast to the series converter, the shunt converter operates under a closed-loop current control structure whereby the active power component, needed for the regulation of the common dc link voltage, and the reactive power component, used for the ac bus voltage regulation, are independently controlled. The shunt converter therefore fulfills two functions with two independent control loops: it provides shunt reactive power in response to reactive shunt current reference, and it maintains the necessary voltage level on the dc link, and thereby provides the active power supply or sink needed for the support of the series voltage injection. In other words, the control loop for the active power ensures the required active power balance between the output of the two converters (i.e., Pshunt+Pseries=0). As emphasized previously, the converters do not (and could not) exchange reactive power through the dc link.

Figure 8-14 UPFC Internal control structure

The Internal converter control is thus composed of two separate control blocks, one for the series, and the other for the shunt converter, as shown in Fig. 8.14. They are operated in synchronism with the ac system, but their functional roles are executed independently. The synchronization is provided by a common phase-locked loop which ensures that all variables run in a common, synchronous time frame.

s

1~ v

~ i

qRei dcRefV

control converter

Shunt

converter Shunt

si ~

1v ~

pqRefv~

converter control

Series

1v ~ v’

1

~ i ~

converter Vdc

Series

~i p

~V’

1

v~

P

Q

-+

loop locked Phase-

1v~

θ θ

Internal converter control

8-21

The external (or functional) operation control defines the operating mode of the UPFC and is responsible for generating the internal references, vpqRef for the series converter, and iqRef for the shunt converter to meet the prevailing demands of the transmission system. As explained in the previous sections, the UPFC has, in addition to providing an unconstrained two dimensional voltage injection (in a defined region) to independently control active and reactive power flow, capability to emulate the operation of line impedance compensation and transmission angle regulation. The functional operating modes, such as defined vpq voltage injection (determined by an independent power system optimization control), active and reactive power regulation to given PRef and QRef references, line impedance compensation to a ZRef or XRef reference, or transmission angle regulation to a σRef reference, can be can be set manually (via a computer keyboard) by the operator or dictated by an automatic system optimization control to meet specific operating and contingency requirements. An overall control structure, showing the internal, the system operation, and system optimization controls with the internal and external references is shown in Fig. 8.15.

s

1~ v

~ i

shpRefi

control converter

Shunt

converter Shunt

si~

1v ~

pqRev~

converter control

Series

1v ~ v’ 1

~ i ~

converter Vdc

Series

~i p

~ V’ 1

v~

P

Q

loop locked Phase-

1v~

θ θ

shqRefi

V dc

θ

1v ~ v’ 1

~ i ~ System operation control

System optimization control

1RefV shqRefI pqRefv~ P

RefQ Ref Z Ref

σ Ref

Mode selection

Operator inputs

Power systemvariables

UPFC control system

Transmission line

8-22

Figure 8-15 UPFC control operated by the inputs of a System optimization control managing the area transmission

The unique functional capability of the UPFC, based on unrestricted series voltage injection together with independently controllable reactive shunt compensation is facilitated by the circuit structure of two back-to-back converters. This flexible structure not only facilitates several operating and control modes in normal operation, but it also allows the total de-coupling of the two converters (i.e., separating the dc terminals of the two converters) to provide independent reactive shunt compensation (STATCOM) and reactive series compensation (SSSC) without any active power exchange if network conditions or contingency situations so require.

The possible functional operating modes of the UPFC with the available control options for the shunt and series converters are detailed below.

8.1.4.1 Functional Control of the Shunt Converter

The shunt converter is operated so as to draw a controlled current, ish, from the line. One component of this current, ishp, is automatically determined by the requirement to balance the active power of the series converter by maintaining the common dc link voltage. The other current component, ishq, is reactive and can be set to any desired reference level (inductive or capacitive) within the capability of the converter. The reactive compensation control modes of the shunt converter are, of course, very similar to those commonly employed for the STATCOM and conventional static var compensator.

Reactive Power (VAR) Control Mode. In reactive power control mode the reference input is a capacitive or inductive var request. The shunt converter control translates the var reference into a corresponding shunt current request and adjusts the gating of the converter to establish the desired compensating current. The control in a closed-loop arrangement uses current feedback signals obtained from the output current of the shunt converter to enforce the current reference. A feedback signal representing the dc bus voltage, Vdc, is also used to ensure the necessary dc link voltage.

Automatic Voltage Control Mode. In voltage control mode (which is normally used in practical applications), the shunt converter reactive current is automatically regulated to maintain the transmission line voltage to a reference value at the point of connection, with a defined droop characteristic. The droop factor defines the per unit voltage error per unit of converter reactive current within the current range of the converter. The automatic voltage control uses voltage feedback signals, usually representing the magnitude, V1, of the positive sequence component of bus voltage v1.

8.1.4.2 Functional Control of the Series Converter

The series converter controls the magnitude and angle of the voltage vector vpq injected in series with the line. This voltage injection is, directly or indirectly, always intended to influence the flow of power on the line. However, vpq is dependent on the operating mode selected for the UPFC to control power flow. The possible operating modes include:

8-23

Direct Voltage Injection Mode. The series converter simply generates the voltage vector, vpq, with the magnitude and phase angle requested by the reference input. This operating mode may be advantageous when a separate system optimization control coordinates the operation of the UPFC and other FACTS Controllers employed in the transmission system. Special functional cases of direct voltage injection include those having dedicated control objectives, for example, when the injected voltage vector, vpq, is kept in phase with the system voltage for voltage magnitude control, or in quadrature with it for controlled “quadrature boosting”, or in quadrature with the line current vector, i, to provide controllable reactive series compensation.

Bus Voltage Regulation and Control Mode. The injected voltage vector, vpq, is kept in phase with the “input” bus voltage vector v1 and its magnitude is controlled to maintain the magnitude “output” bus voltage vector v’1 at the given reference value.

Line Impedance Compensation Mode. The magnitude of the injected voltage vector, vpq, is controlled in proportion to the magnitude of the line current, i, so that the series insertion emulates an impedance when viewed from the line. The desired impedance is specified by reference input and in general it may be a complex impedance with resistive and reactive components of either polarity. A special case of impedance compensation occurs when the injected voltage is kept in quadrature with respect to the line current to emulate purely reactive (capacitive or inductive) compensation. This operating mode may be selected to match existing series capacitive line compensation in the system.

Phase Angle Regulation Mode. The injected voltage vector vpq is controlled with respect to the “input” bus voltage vector v1 so that the “output” bus voltage vector v’1 is phase shifted, without any magnitude change, relative to v1 by an angle σ in response to the reference input. One special case of phase shifting occurs when vpq is kept in quadrature with v1 to emulate the “quadrature booster”.

Automatic Power Flow Control Mode. The magnitude and angle of the injected voltage vector, vpq, is controlled so as to force such a line current vector, i, to assume a magnitude and angle that results in the desired active and reactive power flow in the line. In automatic power flow control mode, the series injected voltage is determined automatically and continuously by a closed-loop control system to ensure that the desired P and Q are maintained despite system changes. The transmission line containing the UPFC in this operating mode appears to the rest of the power system as a high impedance power source or sink. This operating mode, which is not achievable with conventional line compensating equipment, provides great potential for power flow scheduling and management, as well as asset utilization. It can also be applied effectively to handle dynamic system disturbances (e.g., to damp power oscillations).

8.1.4.3 Stand Alone Shunt and Series Compensation

The UPFC circuit structure offers the possibility of operating the shunt and series converters independently of each other by disconnecting their common dc terminals and splitting the capacitor bank. In this case, the shunt converter operates as a stand-alone STATCOM, and the series converter as a stand-alone SSSC. This feature may be included in the UPFC structure in order to handle contingencies (e.g., one converter failure) and be more adaptable to future system changes (e.g., the use of both converters for shunt only or series only compensation). In the

8-24

stand-alone mode, of course, neither converter is capable of absorbing or generating active power so that their operation is restricted to the reactive power domain, and their control modes to those of the reactive shunt and series compensators.

8.1.5 Basic control structure for the series and shunt converters

Although the UPFC has many possible operating modes, the basic control structures for the series and shunt converter within the overall internal control remain substantially the same. The block diagram-based explanations here focus on the automatic power flow control mode providing independent control for active and reactive power flow in the line. This control mode utilizes most of the unique capabilities of the UPFC and it is expected to be the basic mode of series compensation in the majority of practical applications. Similarly, the UPFC shunt compensation capability is normally used for automatic bus voltage regulation. These control modes could easily altered for different series and shunt compensation objectives by deriving appropriate references for other system variables to be controlled.

The block diagram for the control scheme of the series converter is shown in Fig. 8.16. (Note that for clarity only the most significant features are shown in this and related figures ignoring some of signal processing and limiting functions that may be used in practical circuits.) It is assumed that the magnitude and phase angle of the output voltage generated by both the series and shunt converter are independently controllable at a fixed dc link voltage.

Figure 8-16 Functional block diagram of the series converter control

As shown in Fig. 8.16, the automatic power flow control for the series converter is achieved by means of a vector control scheme that regulates the transmission line current using a synchronous reference frame (established with an appropriate phase-locked loop producing reference angle θ) in which the control quantities appear as dc signals in the steady state. (Recall

p i

i q

current computer

Real and reactive

-+

i pRef

Q Ref

P Ref v’

1~

v

Series converter

2θ +ρ V

pq

v 2q

~i V

d

θ

ρ

i qRef

v’ 1

~v

1~

v 1

~

v’ 1

~ ~ i amplifier

Error

- +

computer Limit

current computer

Real and reactive

voltagelimiter

Seriesinjected

loop

Phase-locked

anglecomputer

Magnitude and

gate pattern logic

Series converter

+ +

8-25

that the synchronous reference frame with coordinates q and p rotates synchronously with the vector/phasor variables at the power system frequency. Consequently a steady-state vector – i.e. phasor - with some angle with respect to the p axis will have a constant, but generally different, projection to both the p and the q axes, defining the in-phase and quadrature or “active” and “reactive” components of the vector.) The appropriate references for active and reactive line current components, ipRef and iqRef, are determined for the desired active and reactive line power, PRef and QRef. These are compared with the measured line currents, ip and iq, and, within a closed control-loop, used to derive the magnitude and angle of the series converter voltage, Vpq and ρ, respectively. Note that a voltage limiter in the forward path is employed to enforce practical limits on the series voltage injected. These limits may result from system restrictions (e.g., voltage and current limits) or equipment and component ratings.

The block diagram for the control scheme of the shunt converter is shown in Fig. 8.17. This is also a vector control scheme in which the controlled quantity is the current ish drawn from the line by the shunt converter. The active and reactive components of this current, however, have a different significance. The reference for the reactive current component of the shunt converter, ishqRef, is generated by an outer voltage control loop that is responsible for regulating the ac bus voltage. The reference for the active current component, ishpRef, on the other hand, is generated by a second voltage control loop that regulates the dc bus voltage. This second control loop thus achieves the automatic balancing of the active power need of the series converter in executing the series voltage injection for the desired power flow control. The dc voltage reference, VdcRef, may be kept substantially constant.

Figure 8-17 Functional block diagram of the shunt converter control

v

Shunt converter

shθ +α V

sh

v sh

θ

α

v 1

~

loop

Phase-locked

anglecomputer

Magnitude and

gate patternlogic

Shunt converter

+ +

V d

v 1~

computer

Voltage magnitude current

computer

Real and reactive

-+

i shp

si ~

i shq

-+

computer Limit -

+ 1RV

Amplifier Error

limiter

Reactive current

amplifierError

shpRei

shqRefi

amplifier Error

- + V dcRe

8-26

It is pointed out here that shunt reactive compensation may also be controlled by varying the dc voltage and thereby changing the magnitude of the shunt converter’s output voltage. In this case, the outer voltage loop would regulate the ac bus voltage and would also control the dc capacitor voltage by changing the angle α of the converter voltage with respect to the ac bus voltage until the dc capacitor voltage reaches the value necessary to achieve the reactive compensation demanded. With this operation of the shunt converter, the closed-loop controlling the output of the series converter would be responsible for maintaining the magnitude of the injected voltage, vpq, in face of the variable dc voltage. This method may be used to minimize harmonics and losses due to direct voltage magnitude control (e.g., executed by pulse-width modulation). However, disadvantage of this “indirect” control technique is that the dc voltage is allowed to vary (typically up to ±12%) according to the prevailing shunt compensation demand, which would inevitably reduce the attainable magnitude of the injected series voltage when the shunt converter is operated with high reactive power absorption. However, this, in many applications, may be an acceptable trade-off.

It can be observed that in the control scheme of Fig 8.16, a limit is imposed on the reactive current component of the shunt converter. This shows that operating priority is given for power flow control and, for this reason, the shunt reactive compensating current may be limited in order to use the maximal current capability of the shunt converter to satisfy the active power demand of the series converter. Of course, priority could be given to bus voltage regulation, if system operation would prefer that. Also, current limit could be imposed without priority, affecting both series and shunt compensation equally. As indicated earlier, the control block diagrams shown in Figs. 8.16 and 8.17 represent only a selected operating mode to illustrate the basic control structure. Other control algorithms needed for additional operating modes of the UPFC could similarly be implemented. It should also be noted that the block diagrams here omit control functions related to converter protection, as well as sequencing routines during operating mode changes and start-up and shut-down procedures. These are similar to those used in other converter based compensators and controllers, such as the STATCON, SSSC and others.

8.1.6 Practical Control Considerations

Apart from the basic active time control functions discussed above in the Internal and External Controls of the UPFC, the total control system also has many other elements to manage the proper and safe operation of the equipment with high reliability and availability, as well as to accommodate proper interface with local and remote operators. The main elements of a modern UPFC control system, shown schematically in Fig. 8.18, include:

8-27

Figure 8-18 Main elements of modern, practical UPFC control system

1. Interface between a high power, high voltage semiconductor valves of the UPFC switching converters and a highly sophisticated real time control system, incorporating the Internal and External Control functions. This interface, transmitting gating commands from the control to the valves and status information from the valves to the control, is usually implemented by optical links

2. Signal measuring and processing circuits for system and equipment variables. The real time control and protection relays (and operator displays) need as inputs certain system variables, such as appropriate bus voltages and line currents, as well as corresponding internal voltages and currents of the equipment, from which magnitude, phase, frequency and other relevant information characterizing the prevailing system conditions to provide the desired transmission control and monitoring of equipment operation and performance.

3. Supervisory control and status monitor which interfaces with the all parts of the UPFC, including all essential components of the converters and their support equipment (e.g., cooling system, power supplies, breakers, switches, interlocks, etc.) It collects status information from every part of the system, usually via serial communication links, organizes and interprets the status data to determine the operational integrity of the UPFC and to provide diagnostics for possible malfunctions and failures. It also carries out the start-up and shut-down sequencing and other operating routines of the compensator and provides appropriate communication links for the local and remote operators.

4. User interface with CRT graphical displays is usually provided by a stand-alone computer with an appropriate CRT monitor, key board, and pointing device for data entry. This computer usually has a serial link to the status processor and runs a graphical display and control software. Through the interface a large amount of information is available for the operation, diagnostic, and maintenance purposes in graphical and numerical form. The information includes: Status information from the valves, identifying failed power semiconductors and other components and associated circuits;

Internal control

Supervisory control &

Measured system variables

system status information

SCADA

Inter-

face

Auxiliary equipment and

UPFC

converters

AC system

Optical links Inter-

face

Control

status monitor panel and

CRT monitor

Real time control

External control

Transformer

coupling

8-28

selected operating modes of the compensator and associated control and operational parameter settings; control operating and redundancy; and, status of support equipment such as cooling system, auxiliary power supplies, breakers, switches, etc., and building climate status (temperature, humidity, etc.).

Conclusions on the UPFC control system.

1. The basic structure of the UPFC control system is composed of the Internal converter control and System operation control. The Internal converter control operates the power semiconductor switches on a micro scale to reproduce at the output the input reference signals provided. The System operation control develops the reference input variables for the Internal converter control from the desired, operator determined system references (Voltage, power, etc.) and the measured system variables. The control system may also include automatic operating and protection limit setting, diagnostic algorithms, monitoring, communication and other hardware defined by the customer for user to system interface.

2. The two constituent shunt and series converters have a different operating function within the UPFC structure. The series converter function is to inject the required compensating voltage for power flow control. The shunt converter function is to enable the series converter to do that by providing the necessary active power to it from the line. Consequently, the Internal converter control has two major sub-controls: one is for the series converter to execute the desired voltage injection rapidly; the other is for the shunt converter to maintain the necessary power balance within the UPFC by maintaining the common dc voltage. The shunt converter control also has an additional control loop to regulate the ac bus voltage by controlling the reactive power exchange.

3. The UPFC control system may also include an System optimization control, the function of which is to derive appropriate “operator” references from measured system variables representing a part of the system that the UPFC operation can influence, in order to take into account the operation of other compensating and control equipment, as well as system conditions, to provide an overall optimal operation for that part of the system.

4. The UPFC control system is, in general, based on the so-called vector-control approach, which is particularly advantageous to separate and rapidly control the “active” (in-phase) and “reactive” (quadrature) components of system variables, represented as rotating vectors.

8.2 Interline Power Flow Controller (IPFC)

8.2.1 The concept and motivation

It has been shown that the Unified Power Flow Controller has the unique capability of independently controlling both the active and reactive power flow in the line. The UPFC concept provides a powerful tool for the cost effective utilization of individual transmission lines by facilitating the independent control of both the active and reactive power flow, and thus the maximization of active power transfer at minimum losses, in the line. The Interline Power Flow Controller (IPFC) is essentially an extension of the UPFC concept to the economic control and optimization of multi-line power transmission systems.

8-29

The IPFC addresses the problem of compensating a number of transmission lines at a given substation. Conventionally, series capacitive compensation (fixed or controlled) is employed to increase the transmittable active power over a given line and also to balance the loading of a normally encountered multi-line transmission system. However, series reactive compensators are unable to control the reactive power flow in, and thus the proper load balancing of, the lines. This problem becomes particularly evident in those cases where the ratio of reactive to resistive line impedance (X/R) is relatively low. Series reactive compensation reduces only the effective reactive impedance X and, thus, significantly decreases the effective X/R ratio and thereby increases the reactive power flow and losses in the line. The IPFC scheme, in addition to independently controllable reactive series compensation of each line, provides means to transfer active power between the compensated lines. This capability makes it possible to: equalize both active and reactive power flow between the lines; reduce both the active and reactive loop-flow of power; compensate against resistive line voltage drops and the corresponding reactive power demand; increase the effectiveness of the overall compensating system for dynamic disturbances. In other words, the IPFC can potentially provide a highly effective scheme for power transmission management at a multi-line substation.

8.2.2 Basic Operating Principles and Characteristics of the IPFC

In its general form the Interline Power Flow Controller employs a number of dc to ac converters each providing series compensation, similarly to that used in the UPFC, for one of the lines of a multi-line system. However, within the general concept of the IPFC, the dc terminals of these converters are not supplied from a shunt-connected converter like in the UPFC scheme (although this possibility in some applications can be advantageously utilize), but are linked together at their dc terminals, as illustrated in Fig. 8.19. With this arrangement, each converter, in addition to providing independently controllable series reactive compensation, can be controlled to supply active power to, or draw it from, the common dc link from, or to, its own transmission line. Thus, an overall surplus power can be made available from the under utilized lines which then can be used by other lines to reduce reactive power flow by active power compensation. In this way, some of the converters, compensating lines with a heavy burden of reactive power flow, or high priority lines, can be equipped with full two-dimensional, reactive and active power control capability, similar to that offered by the UPFC. Evidently, this arrangement mandates the rigorous maintenance of the overall power balance at the common dc terminal to make the total active power exchange between the IPFC and all the lines equal to zero by proper coordination of the converters control.

8-30

Figure 8-19 Interline Power Flow Controller (IPFC) scheme

Consider an elementary IPFC scheme consisting of two back-to-back dc to ac converters, each compensating a transmission line by series voltage injection. This arrangement is shown functionally in Fig. 8.20a, where the outputs of two back-to-back dc to ac converters are connected in series with transmission Lines 1 and 2 to inject voltage V1pq and V2pq, respectively.

The common dc link of the converters provides a path for active power flow between the two lines. Transmission Line 1, represented by reactance X1, has a sending-end bus with voltage phasor V1s and a receiving-end bus with voltage phasor V1r . The sending-end voltage phasor of Line 2, represented by reactance X2, is V2s and the receiving-end voltage phasor is V2r. For clarity of the illustrations, all the sending-end and receiving-end voltages are assumed to be constant with fixed amplitudes, V1s = V1r = V2r = V2r = 1.0 p.u. The impedance of the two lines and their transmission angle, as well as the rating of the two compensating voltage sources, are also assumed to be identical, i.e., X1 = X2 = 0.5 p.u., δ1 = δ2 (= 30o), and V1pqmax = V2pqmax = 0.25 p.u. Thus, the maximum active or reactive power increase is VVpq/X=0.5 p.u. for both lines Although Lines 1 and 2 could be (and in practice are likely to be) different with different transmission line voltage, impedance and angle, in order to make the relationships governing the operation of the IPFC clear, the above stipulated operational identity of the two lines is maintained throughout this section. Note that the operational identity of the two lines does not violate their assumed independence; they could even belong to different systems, having arbitrary phase relationship with each other since the active power transfer between them through the dc link of the converters is asynchronous.

Transmission lines

+ - +

HV 2

HV n

HV 1

- + -

DC bus

Converter 1 Converter 2 Converter n

8-31

Figure 8-20 a. Two-line IPFC scheme; b. phasor diagram for “priority” Line 2 and c. corresponding operating region in the P,Q plane at δ=30°

In order to establish the transmission relationships between the two independent lines, Line 2 is arbitrarily selected to be the one with transmission priority for which free controllability of both active and reactive line power flow is stipulated. The reason for this stipulation is to derive the constraints which the free controllability of Line 2 imposes upon the power flow control of Line 1.

A phasor diagram representing transmission in Line 2, which defines the relationship between V2s, V2r, V2X (the voltage phasor across X2) and the inserted voltage phasor V2pq, with controllable magnitude (0 ≤ V2pq ≤ V2pqmax) and angle (0 ≤ ρ2 ≤ 3600), is shown in Fig. 8.20b. It should be noted that this phasor diagram is similar to that characterizing a UPFC compensated line (refer to Fig. 8.7).

As illustrated for the UPFC in Fig. 8.11 the effect of the rotating compensating voltage V2pq on the power transmission can best be illustrated in the Q2r,P2r plane, where each value of V2pq

11V V

1pq V’ X

1

Line 1

I 1 1rV

V 2pq

22V V’ X

2I 2 2r

V

Line 2

IPFC +

12

P

(b)

V V

V 2xo

2s 2s V’

2x V

2r

δ

ρ

V 2pq

2

2

2

δ ’ 2

Q

0

-0.5

δ

P

=30°

δ 2

1

=0

1.5

2rρ

2

2r1.00.5

V 2pq

-1.0

(c)

(a)

8-32

(magnitude plus angle) defines a pair of corresponding P2r and Q2r values that represents the total MVA the transmission line carries. As established for the UPFC, the full (360°) rotation of phasor V2pqmax produces a circular locus around the operating point of the uncompensated line defined by transmission angle δ2, which determines the attainable maximum P2r and Q2r. At the selected transmission angle, δ2 =30o, the uncompensated Line 2 (V2pq = 0) with the stipulated X=0.5 p.u. reactance value transmits P2r,30° = 1.0 p.u. active power to, and absorbs Q2r,30° = 0.268 p.u. reactive power from the receiving-end bus. These P2r,30° and Q2r,30° coordinate values define the center of the circular control region. With the maximum amplitude of the compensating voltage, V2pqmax=0.25 p.u., the radius of the circular region in the Q2r,P2r plane is [P2pq

2 + Q2pq

2]1/2 ( = 0.5 p.u. VA), as illustrated in Fig. 8.20c. This circular locus provides the boundary for the two-dimensional control range within which any corresponding Q2r and P2r values are achievable by the proper setting of the magnitude, V2pq, and angle, ρ2, of the compensating voltage phasor V2pq.

The compensation of the priority Line 2 described above is identical to that characterizing the operation of the UPFC. However, in the case of the UPFC, the active power exchanged through the series voltage insertion is supplied via the shunt-connected converter from the sending-end bus. This power can be positive or negative, as the prevailing compensation requirements dictate with usually no other limit, but the rating of the shunt converter. In the case of the simple two-line IPFC considered here, the active power is obtained from the other line via the series-connected compensating converter of that line. In order to establish the possible compensation range for Line 1, under the constraints imposed by the unrestricted priority compensation of Line 2, it is helpful to decompose the total compensating MVA provided for Line 2 into reactive power Q2pq [Mvar] and active power P2pq [Mw]. To this end, the injected voltage phasor V2pq is decomposed into two components, one, V2q, in quadrature with the line current and the other, V2p, in phase with it. The products of these with the line current define the reactive and active power, Q2pq and P2pq. The component Q2pq, generated internally by Converter 2, evidently provides self-sufficient series reactive compensation for Line 2, just like a controllable series capacitor. The component P2pq, on the other hand, provides active power compensation for Line 2, but this power must be supplied for Converter 2 by Converter 1 from Line 1.

Consider first the special case in which the compensation of Line 2 requires no active power. In this case P2pq is zero, Converter 2 operates as a controllable series compensator injecting a compensating voltage that is in quadrature with the line current, i.e., V2pq= V2q. This means that this voltage directly adds to or subtract from the voltage across the reactive line reactance X2, which is, by definition, also in quadrature with the line current. The loci of the compensating voltage is on a straight line, termed here as reactive compensation line, passing through the end points of the sending- and receiving-end voltage phasors, as illustrated in Fig. 8.21b. Suppose now that a compensating voltage component, V2p that is in phase with the line current is added with a fixed magnitude to the variable quadrature component V2q. This will have an effect of shifting the reactive compensation line up or down, depending the polarity of V2p with respect to the line current phasor I, creating a new voltage compensation line. This line can be visualized as loci of the end points of the compensating phasor, V2pq=V2p+V2q, when in-phase component V2p is kept constant and quadrature component V2q is continuously varied over its total range of -V2qmax≤V2q≤+ V2qmax. In a different way, it may also be said that if the magnitude of voltage phasor V2pq is controlled over the attainable range of angle ρ so that its end point stays on a straight line trajectory (“voltage compensation line”), parallel to the line connecting the end-points of the sending-end and receiving-end phasors (“reactive voltage compensation line”), as illustrated in

8-33

Fig. 8.21b, then the product of V2pq with the prevailing line current phasor I2 will result in constant active power demand (that the converter must supply or absorb) independent of angle ρ. This means that within the circular operating region of V2pq, the reactive compensation is freely variable along a “voltage compensation line” at any value of the active power compensation within this operating range. As indicated, the active power demand is, by definition, zero when the trajectory of V2pq coincides with the “reactive voltage compensation line” (V2p=0), which divides the circular operating region into two equal halves. An increasing amount of active power is to be supplied to Line 2 as the voltage compensation line is shifted higher and higher above the “reactive voltage compensation line” in the upper half of the control region. Conversely, increasing active power is to be absorbed from Line 2 as the compensation line is shifted lower and lower below the “reactive voltage compensation line” in the lower half of the compensation region.

Figure 8-21 P-Q control lines in the in the operating region of the P,Q plane corresponding to different amount active power compensation of the “priority” Line 2 controlled by in-phase voltage injection

The “voltage compensation lines” prescribing the trajectory of the phasor V2pq for constant active power demand in the phasor diagram characterizing Line 2, define a linear relationship between the receiving-end reactive and active power, Q2r and P2r, respectively, within the circular locus representing the boundary of the control region in the Q2r,P2r plane, as shown in Fig. 8.21c. Thus a “reactive compensation control line”, which crosses the center of the boundary circle, defines the Q2r versus P2r relationship for purely reactive variable compensation. An infinite number of parallel Q2r versus P2r control lines of decreasing length, corresponding to the “voltage compensation lines” of the phasor diagram, can be established above and below the reference control line corresponding to the “reactive compensation line”, as a function of the active power exchanged. A number of such lines, corresponding to the “voltage compensation lines” of the related vector diagram, are drawn in the same manner (dashed with different spacing), in the Q2r

V 2p

22V V’ X 2I 2 2rV

Line 2

IPFC

+

12P

(a)

=0 V p

-

+

an

= 0 P 1

(b)

V V

V 2x

2s 2

δ

ρ V 2pqm

2

2

2

Q

0

-δ

P

=30°

δ21

=0)

1.2 ρ

22r 1.0.

(V 2p-

L 2 supplies active

P2r Q2 vs δ2

L 2 receives active

L 2 receives

L 2 supplies

Reactive voltage

Voltage compensatio

+

-

(c)

Reactive compensation

P - Q control

8-34

versus P2r control region of the Q2r,P2r plane shown in the figure. Comparison of Figs. 8.21b to 8.21c indicates that the reactive power flow decreases (pink lines) or increases (blue lines) in proportion to the active power supplied to, or absorbed from the line by series compensation.

With respect to the total IPFC scheme shown in Fig. 8.20, the following conclusions can be drawn from the above discussion. The operating point of Converter 2 compensating the “priority” Line 2 can be considered to be a point of a particular “voltage compensation line.” Consequently, a corresponding point on the related Q2r versus P2r control line in the Q1r,P1r plane defines the resultant active and reactive power flow in the transmission line. In general, at a selected operating point, Converter 2 has to exchange both active and reactive power with Line 2. However, the converter can internally generate only the reactive power and thus it must be supplied with the active power (plus or minus) it exchanges. The active power demand remains constant, while the internally generated reactive power changes, as the operating point of the converter is shifted along a selected “voltage compensation line.” With the shifting operating point (i.e., with the resultant variable reactive compensation), the receiving-end active and reactive power moves along the relevant Q2r versus P2r control line in the Q1r,P1r plane, changing primarily the transmitted active power P2r. Moving the operating point from one “voltage compensating line” to another changes the active power demand of the converter and shifts the resulting receiving-end active and reactive power to a parallel Q2r versus P2r control line in the Q2r,P2r plane, and thus primarily changing the reactive power Q2r in the line.

It follows therefore, that in order to satisfy the active power demand of Converter 2 operated along a selected “voltage compensation line,” Converter 1 must be operated along a complementary “voltage compensation line” so as to precisely supply this demanded active power from Line 1 via the common dc link for Converter 2. (The term “complementary” means here that if Converter 2 is operated on a voltage compensation line above the reference “reactive compensation line”, then Converter 1 must be operated on a corresponding voltage compensation line below “reactive compensation line”.) In other words, the relationship P1pq = - P2pq must be continuously satisfied. (In practice, this condition is usually satisfied by controlling Converter 1 so as to maintain the voltage of the common dc link in face of the variable active power demand of Converter 2.)

The operation of the two converter IPFC scheme is illustrated with the help of the complementary regions of “voltage compensation” and Qr versus Pr control lines, characterizing the operations of the “priority” Line 2 and the “support” Line 1, in Fig. 8.22. For clarity, the previously stipulated transmission line and operating parameters are assumed (two independent lines with identical impedances and operating parameters, transmission voltage and angle). In this case, the selection of “priority” and “support” is arbitrary; in practice, the priority line usually would have heavier reactive load.

8-35

Figure 8-22 Illustration of the basic complementary relationship governing the operation of a two-line IPFC. Line 2 is operated as the selected “priority” line and Line 1 is as the “support” line

The right hand side of Fig. 8.22b shows the circular operating area of the “priority” Line 2 in the Q2r,P2r plane. The “priority” designation, however, here means that the desired power transmission necessitates the reduction of the reactive line power flow, while maintaining the full controllability of the active power up to the maximum power of 1.5 p.u. This requirement puts the operation of the IPFC controlled Line 2 in the upper region of the circular operating area, above the reference control line representing purely reactive compensation (P2pq=0). (As shown in the figure, this semi-circle region may further be limited on the top by the available maximum transferable active power or by the allowed maximum magnitude of corresponding voltage.)

The left hand side of Fig. 8.22b shows the operating area for the IPFC injected voltage phasor V2pq. As can be seen, this semi-circle area is above the “reactive compensation line”, meaning that Converter 2 supplies active compensating power P12 for Line 2 by series voltage injection. This P12 power must, of course, be provided for Converter 2 by Converter 1 via the common dc link from “support” Line 1. To absorb the P12 power from Line 1, the active component of voltage phasor V1pq must be in phase opposition to the line current phasor I1, which requirement constrain the operation region of phasor V1pq in the semi-circle area below the “reactive compensation line”, as shown at the left hand side of Fig. 8.22 a. (Note that the operating regions of both V1pq and V2pq show a possible limit for the maximum value of allowed for the injected active voltage component due to the increase of transmission voltage.)

The operating region of Line 1 in the Q1r,P1r plane is shown at the right hand side of Fig. 8.22 a. As could be expected, the restriction of the V1pq to the operating region below the “reactive

2 δ -

0

-

1P IPF

pqP

1 δ -

-

=00

P2

ρ

12,limitP-

12,limit

=30°δ2

Q=2

2r

δ1.

δ

=60° δ 2

2

2. P 2r

1ρ

=30δ

12,limit P-

1

12,limit P+Q1r

=0δ1

1. P P =

=60° 1

δ 1

12

δ

2. 1r

1s 1s V V 1pq V’ X 1

Line

I 1 1r V

V 2pq

2s 2s V V’ X 2 I 2 2r V

Line

1s V 1r V V 1pq

1 ρ

2 ρ V 2pq

2s V 2r V

pqP =0

+

V

Control range with selected priority

Control range with selected priority

P =12

(a)

(b)

8-36

compensation line” constrains the corresponding semicircle operating region of the attainable P1r and Q1r values below the reference control line representing purely reactive compensation. This means that, although the active power in Line 1 is still well controllable, but the reactive power flow in Line 1 increases in proportion to the active power it supplies to Line 2. It is emphasized again that in practice the IPFC type compensation would be typically applied if the two lines reactive loading would be different or if the capacity of one line would be appreciably greater then that of the other, or some other network or transmission reasons which reactive power flow control or balancing would increase asset utilization or aid network optimization.

Conclusions on the operating principles and characteristics of the IPFC.

1. The IPFC is composed of two back-to-back connected ac to dc converters. The output of each converter provides compensation for a different line by voltage injection in series with the line.

2. Each converter of the IPFC can provide an independent series reactive compensation to control the active power flow in its line. The back-to-back arrangement also makes it possible to transfer active power through the common dc link of the converters from one line to the other to provide series active compensation to control the reactive power flow in the line.

3. The active power transfer capability of the IPFC facilitates independent, UPFC type active and reactive power flow control in one, the selected priority line. However, although the IPFC substantially maintains its reactive compensation capability for the other line, but the reactive power flow will change in it in the opposite direction than in the priority line since it supplies this lines active power compensation demand.

4. The IPFC can effectively be used to control active and reactive power flow to manage transmission capacity, minimize loop flow of power and optimize transmission asset.

8.2.3 Generalized Interline Power Flow Controller for multi-line systems

Although the detailed explanation of the IPFC has been restricted to a simple two-line, two converter arrangement, the general principle is applicable to an n-line transmission system in which each line is compensated by a dc to ac converter which is part of an n-converter IPFC compensator with common dc terminals, as illustrated schematically in Fig.8.19. The operation of this is similar to the two-converter IPFC in that a number of “priority” lines could be designated as long as their active power demand could be satisfied by the remaining “support” converter, since it is a fundamental operating requirement that the total of active power exchanged between the converters and the transmission lines must be zero. However, there can be compensation requirements for particular multi-line transmission systems which would not be compatible with this basic operating constraint of the IPFC. This fundamental requirement could be particularly restrictive under an emergency contingency when those lines which were to support the “priority” lines would also become overloaded or even disconnected due to fault. Such potential problems can be solved, and a more general applicability of the multi-line compensator established, by combining the UPFC and IPFC concepts to realize a generalized Interline Power Flow Controller arrangement, in which a shunt connected converter is added to the number of converters providing series compensation, as illustrated in Fig 8.23. With this

8-37

scheme the net power difference at the dc terminal is supplied or absorbed by the shunt converter, and ultimately exchanged with the ac system at a suitable the shunt-bus where active power exchange is not subject to rigorous limitation. Apart from removing the restriction for series active power compensation for the lines, this generalized IPFC scheme greatly simplifies the control by using the shunt-connected converter for the exclusive role of maintaining the common dc link voltage and thereby automatically maintain the required power balance. The arrangement can also be economically attractive because the shunt converter has to be rated only for the maximum active power difference anticipated for the whole system. Furthermore, it can also facilitate relatively inexpensive shunt reactive compensation, if this is needed at the particular substation bus, because the shunt converter would need to be rated only for the vectorial sum, i.e., the square-root of the squares, of active power it exchanges with the other converters and the reactive power it provides for shunt compensation.

Figure 8-23 Generalized multi-line Interline Power Flow Controller scheme in which a shunt-connected converter is used to support the common dc link voltage and thereby allow unrestricted reactive power flow control in each line

It is also possible that the compensation and power flow control requirements at a given substation change due to system contingency, maintenance, and system operational changes. Under these conditions and scenarios, complete functional changes in compensation and control requirements may be possible. For example, in a weakened system or under heavy line loading, voltage support of a given bus may be more important than the power flow control of a given line. In this case a series compensator could be converter into an additional shunt compensator within or outside of generalized IPFC system. Overall, a multi-converter compensating system could provide a hitherto unachievable operational flexibility together with almost unrestricted functional convertibility.

Transmission lines

+ - +

HV 2

HV n

HV 1

- + -

DC bus

Converter 1 Converter 2 Converter n

+ -

Shunt Converter

8-38

8.2.4 Basic control system

The control structure of the IPFC is similar to that of the UPFC shown in Figs. 8.15, 8.16, and 8.17 with appropriate changes in the controlled variables and the necessary constraints imposed by the possible limitations of active power transfer. A possible IPFC internal control scheme is shown in the form of a block diagram in Fig. 8.24. For this structure the assumption of the previous section, stipulating System 2 with IPFC Converter 2 as the “priority” system requiring the independent control of both active and reactive power is for continuity and clarity retained. This stipulation makes the control of the two converters functionally somewhat different. However, the reader should note that in practice the two converter controls would be identical with control inputs selecting either converter for the “prime” or “support” operating role.

Figure 8-24 Block diagram for the basic control of a two-line IPFC

As shown in Fig. 8.24, the operation of Converter 1 is synchronized to line current I1 and Converter 2 to line current I2 by two independent Phase Locked Loops. This enables each converter to provide independent series reactive compensation and to keep operating under contingency conditions when the other line or converter is out of service.

The input to the “priority” control is either the desired active and reactive line power, P2 and Q2 (indicated in the figure by P2Ref and Q2Ref) or it could be the desired quadrature and in-phase compensating voltage V2q and V2p, shown in the figure as internal references, V2qRef and V2pRef, derived from P2 and Q2. Voltage component V2qRef, being in quadrature with the prevailing line current I2, represents series reactive compensation to control the transmitted active power, and component V2pRef, being in phase with that, represents series active compensation to control the reactive power flow in the line.

~i1

~ v 1p

V1

Pp limi

2i~

1i~

2ψ2

θ+ ψ

2 2θ

v’ 2~

reference computer

Real & reactive voltage

Q 2Ref P 2Re

reference computer

Real & reactive voltage

v 2pRef

v 2qRef

2i~ v 2p

~

v 2q v 2p

-+

-+ amplifier

Error

amplifierError

V dc

-+V dcRe

voltagelimiter

Seriesinjected

computerLimit

2i~v’2

~2i~

loop

Phase-locked

anglecomputer

Magnitudeand

V2p

Series converter 2

gate patternlogic

Series converter

++

power limit computer

Reactive voltage &

-+

V 1qRe

amplifierError

1ψ1

θ+ ψ

1 1θ

loop

Phase-locked

anglecomputer

Magnitudeand

V1p gate patternlogic

Series converter

++

Series converter 1

8-39

The internally derived references, V2qRef and V2pRef, are compared to the actual voltage components, V2q and V2p, derived from the measured line current and injected voltage vectors, i2 and v2pq. The thus obtained error signals after appropriate amplification and possible limitation provide the input for the computation of the magnitude, V2pq, and angle, ψ2, of the injected voltage vector v2pq.

The limiter preceding the computation of V2pq and ψ2 is an important function of the IPFC control to ensure system operation within predefined constraints. One set of constraints may be provided by the voltage and current limitations of Line 2. (This set of constraints would also apply to the series converter of the UPFC.) The second set of constraints is unique to the IPFC and related to the possible limitations of Line 1 to supply the active power demand resulting from the “priority” compensation of Line 2. As indicated in Fig. 8.24 these limitations may result from insufficient current in Line 1 to supply active power for the maintenance of the dc bus voltage (which is burdened by the active power demand of Line 2) or from limitations imposed for active power transfer by the allowed reactive power flow constraints on Line 1.

The control of Converter 1 is different from that of Converter 2 due to its functional role to support the operation of the “priority” Converter 2 by supplying the necessary active power from Line 1. This requirement means that, since the in-phase component of the injected compensating voltage is imposed on Line 1 by the active power demand of Line 2, the control of Converter 1 can vary only the transmitted active power in its own line by controlling the quadrature component, V1q, of the injected voltage vector v1pq. Thus, the reference input to the control of Converter 1 is the desired quadrature compensating voltage, V1qRef. (The reader should note that an equivalent input could be the desired compensating reactance. It would also be possible to provide a closed regulation loop for the desired active power P1 that would yield V1qRef as an internal reference.) Reference voltage V1qRef is compared to voltage component V1q derived from the measured injected voltage v1pq. From this, and from the amplified error representing the difference between the desired and actual dc bus voltage, VdcRef and Vdc, the magnitude, V1pq, and angle, ψ1, of the injected voltage vector v1pq are derived to generate the output voltage of Converter 1, as illustrated in Fig. 8.24, to provide the required series reactive compensation and maintain the DC link voltage.

The constituents and functional arrangement of the overall practical control system for the IFFC would be similar to that illustrated for the UPFC in Fig. 8.18.

In the case of a multi-converter IPFC compensating several lines, without a shunt-support converter, the control system could become quite complex. The fundamental problem is to maintain the balance for the active power the IPFC would exchange with a multi-line system. There is yet no precedent for the practical realization of such a system, although its functional advantages are well established. Qualitatively it appears that all lines compensated would have to have an defined priority order (or criteria thereof) together with rigorous limits for critical variables (e.g., maximum line current, reactive power, line voltage increase/decrease, etc.) and a computer employed in the control system would determine the attainable decrease and allowed increase in the reactive power flow in the lines according to this priority order and within the limits, and set the compensating voltages for each line accordingly.

8-40

However, if a shunt converter is employed than the control system would be simple and rugged, and the series and shunt converter controls shown in Figs. 8.16 and 8.17 could be used practically without change. The reason for this is that in this case each series converter could be controlled freely as an independent compensator without worry whether the active power exchanged is positive or negative, since the resultant net active power manifesting itself at the common dc link would be balanced by the shunt converter whose control would ensure the maintenance of the dc voltage. Of course, this arrangement would require that the rating of the shunt converter is determined under the expected worst case condition under which the operation of the IPFC is still expected.

8.3 Structural and rating considerations

8.3.1 Introduction

This section is intended for providing some guidance in defining the electrical capabilities of the equipment at its terminals where it is connected to the electrical system, and it does not deal with the ratings of the individual components used in the equipment. Clearly, the ratings of the internal components used in the converter are the responsibility of the manufacturer and it involves detailed design knowledge of the electrical components and their applications, design philosophies and margins the designer applies, protection strategies and other, often proprietary operating considerations, which would generally vary from one manufacturer to the other. The typical user would usually not have sufficient knowledge and expertise in these highly specialized issues to judge them and, also, the manufacturer should have a relatively free hand in these to be able to provide the required guaranty on the equipment. On the other hand, the terminal capabilities would have to be clearly defined by the user with the understanding of the special nature of an electronic equipment in order for the manufacturer to understand the application requirements and be able to properly rate the internal components for the application duties and, at the same time, minimize the overall cost of the equipment.

It is usually the normal, and often the necessary practice for the user to carry out necessary studies, such as load flow, transient and/or voltage stability studies, power oscillation damping, and others, which would provide the required data in terms of voltage and current magnitudes and angles (when applicable). The aim is to select the appropriate values which would define the electrical capability of the equipment in r.m.s. and peak values both for steady-state and short terms transient conditions. It is also an aim to give guidance for the manufacturer regarding the desired functional capabilities and corresponding operating regions to enable the manufacturer to consider different structural options for cost minimization.

In order to appreciate the rating issues and its effect on the cost of the equipment, it is necessary to appreciate the significant difference between conventional and power electronics-based transmission equipment in that the conventional electrical components are quite tolerant against short term over-voltage and over-current excursions, whereas the electronic ones are generally not. Thus, transient rating requirements set without due consideration of their impact on the equipment rating can, on the one hand, unnecessarily increase the cost or, on the other hand, impact the reliability of the power-electronics-based transmission controller. For this reason, in

8-41

the next section the nature of, and the basic considerations for, power electronics based equipment will be briefly summarized.

The technological bases of the transmission controllers considered here, the UPFC and IPFC, is the Voltage-Sourced Converter (VSC). As explained elsewhere in this book, the VSC is a power electronics based switching converter using high power semiconductors to generate from a dc voltage source the necessary, usually three-phase alternating (ac) voltage for compensation. The required rating of the converter in transmission applications is at least several tens of MVA, and often falls in the range of 100 to 300 MVA or more. The maximum rating of suitable power semiconductors is presently in the range of 1-3 kA and 4-6 kV; therefore, a considerable number of individual semiconductor devices are required to provide the desired output rating. Depending on the design, the desired converter rating may be achieved with high voltage, solid-state valves using an appropriate number of individual semiconductor devices in series connection. An alternative approach is to use a number of basic, single device converters in parallel or series connection to get the necessary rating by summing their voltage or current at the output. A frequently used approach, particularly for large ratings, is the combination of the above two, using series connected semiconductor valves to get a basic converter of appreciable rating, and applying a number of such converters in parallel to get the final output.

Various types of power semiconductors (Gate Turn-Off [GTO], Gate-Commutated Thyristors [GCTs] or Insulated Gate Bipolar Transistors [IGBTs], etc.) may be used in the power converter. These devices have different advantages and disadvantages related to switching capabilities and operating losses, which are still debated by technical experts and representatives of various manufacturers. From the standpoint of equipment rating and optimized circuit configurations, the issues related to the merits of the individual semiconductor device are not important; those are to be resolved by the equipment manufacturer in meeting the overall requirements of the user.

All power semiconductors used today are made of thin wafers of large silicon crystals (up to 150 mm, or 6”, in diameter) with rigorously controlled impurities and must be operated within a defined temperature range not exceeding 125 oC or 150 oC). From this physical reality it follows that their capability in handling transient ratings (excess voltage and/or current) could be rather limited. Indeed, the inherent limitations of semiconductors is absolute for voltage rating, strictly defined for maximum magnitude of instantaneous current in terms of milliseconds, and short (in terms of seconds) by conventional standards for sustained overcurrent.

In the actual converter there are usually a sophisticated control and protection circuit arrangements to ensure that the semiconductor voltage and current limits are not exceeded even under extreme system conditions. However, the activation of these arrangements usually means a limitation or interruption in the normal function of the equipment. Therefore, it is extremely important that the user provides credible information regarding voltage and current magnitudes and necessary time durations under worst case steady-state and contingency transient system conditions, during which the substantially uninhibited functional operation of the equipment is expected. It is then the manufacturer’s responsibility to establish the ratings and internal safety margins for the semiconductors and related components used in the converter, taking into account statistical and temperature dependent parameter variations, internal voltage transients resulting from the turn-off actions of the semiconductors, and all other factors that could unduly increase device voltage stresses, unmanageable rate of current increases, and dangerously high internal (junction) temperatures. It is also evident that an overly generous specification requiring

8-42

operation without limitation under high overvoltage and overcurrent conditions for long time intervals characterizing conventional equipment could lead to considerable increase in the cost, since these requirements would generally necessitate substantial increase in the steady-state rating of the power electronics based equipment. Thus, judicious trade-off between cost and functional operation during system contingencies and fault conditions is required to arrive at an optimal system rating meeting essential system requirements at minimum capital cost.

8.3.2 Circuit structural considerations

The UPFC is composed of two, back-to-back connected, voltage sourced converters. Thus the basic, and only fundamental restriction of the constituent converters is that their dc terminals must be freely available and of the same voltage. This restriction allows the application of practically all types converters, except the some of the multi-level types, such as the so-balled “chain-link” type, which have multiple and isolated dc capacitors making it impossible or practically too complex to connect and operate.

As discussed in the Converter Chapter The main converter types with different advantages and disadvantages are the harmonic neutralized, multi-level and pulse-width modulated converters, and their combinations. Their operations are not detailed here, but their main operating characteristics, with advantages and disadvantages, are summarized for convenience.

The harmonic neutralized converter is composed of n basic so-called 6-pulse elementary converters which are operated with some phase displacement and their individual output are combined to yield a high, 6xn pulse number that approximate a sine-wave sufficiently well to avoid filtering. The harmonic neutralized converter has the following advantages:

• Valve switching at the fundamental 60 Hz – low operating losses

• Multi-pulse output voltage – little or no output filtering

• High MVA output by parallel operation – relatively low dc voltage (insulation)

The main disadvantage of the harmonic neutralized converter is a relatively complex and expensive magnetic structure.

The multi-level converter can produce multi-pulse type waveforms by using an increasing number of dc capacitors in different configuration and, in addition to retaining the same main advantages (low switching frequency and no or little output filtering) as the harmonic neutralized converter, it has no auxiliary magnetic components and requires only a single “standard” coupling transformer. However, its dc terminal becomes complex often with additional semiconductors, and the total rating of its dc capacitor, and the cost thereof, is many times that of the harmonic neutralized converter.

The pulse-width-modulated (PWM) converter can have a simple 6-pulse structure with a single coupling transformer. However, its valves are switching at several or even many times of the fundamental frequency, which requires semiconductor switches capable of high frequency operation. Their output voltage contains relatively large unwanted components generated as the side-bands of the basic high frequency switching frequency. Therefore, the operating losses of

8-43

the PWM converter are significantly higher that those of the other two types, and it does require considerable high frequency filtering at the output. It is mentioned that two or more PWM converters can be combined at the expense of some circuit complexity to achieve carrier cancellation and thereby considerable reduction in filtering requirement.

It can be concluded that there is no basic converter circuit approach which would provide optimum solution for output terminal simplicity (magnetics), dc terminal simplicity (dc capacitors), and converter operation and performance (switching frequency, losses, and filtering). With present semiconductor technology the best approach may be the combination of all three basic approaches, each only with a minimum level complexity (e.g., 12-pulse, 3-level converter with relatively low switching, like 180 Hz or 300 Hz).

Generally, it appears the best policy for utilities not to explicitly specify the type of converter to be used in the UPFC or IPFC application, unless there is a definite reason for doing so (e.g., unsatisfactory experience with a particular approach or the necessary constituent of that approach, such as a special magnetic component, or filter, etc.). It appears the best policy to specify the desired performance and operating features and leave to the manufacturer to propose their best structural alternative for that. The specification numbers in the critical areas (e.g., operating losses) will ultimately influence the manufacturers in selecting the converter approach to meet customer requirements and be price competitive.

8.3.3 Rating considerations for the UPFC

As have been shown, the UPFC consists of two converters in back-to-back connection, one connected in series and the other in with the transmission line. In order to establish the necessary rating for the UPFC, it is convenient to separate initially the application requirements for the series and shunt compensation. Indeed, the reactive compensation provided by the series and shunt converters are operationally independent; there is no direct power exchange or necessary functional rating coordination between the two converters.

The initial rating requirements for the UPFC can easily be derived from load flow studies and, depending on the application, this base rating can be subsequently modified by superimposed additional requirements obtained from appropriate dynamic (transient stability, power oscillation damping, etc.) studies.

8.3.3.1 Series converter rating to meet line compensation requirements

The primary objectives of the load flow studies are:

(1) Establish sustainable limits for power transmission (transmitted active power and allowed reactive line power) under normal as well as contingency system conditions that the UPFC has to maintain or, in other words, to define the boundary limits for the operating domain(s) in the Q,P plane;

(2) Derive the corresponding compensating voltage phasors (magnitudes and angles) need to be injected in series with line;

8-44

(3) Ensure that the electrical parameters of the compensated line stay within established transmission operating constraints (e.g., transmission voltage limits);

(4) Establish the required maximum values of the compensating voltage, active power, and line current the UPFC needs to produce to meet all compensation requirements; in other words, define the rating the series converter has to have at the transmission line voltage insertion terminals. (Note that the maximum values of the three parameters, i.e., the magnitude of compensating voltage, of the line current, and injected active power, which determine the rating of the series converter of the UPFC may not occur simultaneously, each may be related to a different operating condition.)

An operating region for the UPFC, representing all possible values of active and reactive line power, would generally be derived around an initial operating point in the Q,P plane obtained with a given system set-up (or contingency) at a given transmission angle with no series compensation (the UPFC is non-operational). Several such regions, covering all the operating conditions of interest, would need to be established and evaluated for maximum compensation and rating requirements. The worst case requirements, in terms of the injected voltage, active power, and line current would then be selected individually from these regions (i.e., the maximum value of each variable could come from a different region representing non-concurrent operating conditions), and checked against prevailing transmission parameter constraints to finally define the system-side, series injection capability of the UPFC.

The above process is illustrated in Fig. 8.25, where at (b) a circular operating region under a particular system condition is shown (in dashed pink line) around an initial operating point, marked Po, Qo, representing the power transmission at the prevailing transmission angle δo without any series compensation (the UPFC is non-operational). The operating point “i”, defined by coordinates Qi, Pi at the boundary of this region represents an extreme transmission condition which requires the UPFC to inject the maximum series compensating voltage Vpqmax at angle .

8-45

The phasor diagram shown at (a) in Fig. 8.25 illustrates the relationship between the system bus voltage V, the injected voltage Vpq and its corresponding circular operating area (in dashed blue line), and the line current I, under the system operation defined by point “i” in the Q,P plane. (Note that these phasors, Vpq=Vpqe

j and line current I=Iejϕ, corresponding to operating point “i” define the active and reactive power, Pi=IVpqcos( + ) and Qi=IVpqsin( + ), that the series converter of the UPFC has to supply for the line to sustain the operation under the system conditions “i”.)

Note that in Fig. 8.25a limits are established for the maximum and minimum allowed transmission voltage. These limits constrain the boundary for the injected voltage phasor at the upper and lower part of the original circular region as shown in the figure. The restriction on the allowed voltage injection results in a corresponding reduction in the related original circular operating region in the Q,P plane, as illustrated in Fig. 8.25b.

It follows from the above simple illustration that the rating requirements for the series converter can be simply derived by establishing quantitative relationships between the worst case P and Q values (defining the transmission requirements at the boundary of the operating region in the Q,P plane), and the corresponding compensating voltage and active power that the UPFC has to supply to maintain the resulting line current. The steps of the process can be summarized process with the help of the illustration in Fig. 8.25, as follows:

1. Derive the circular locus of the largest compensating voltage phasor Vpqmax around the end point of bus voltage phasor V from the boundary of the operating area in the Q,P plane representing the worst case transmission requirements. This automatically defines the maximum magnitude of the compensating voltage.

(b) (a)

φ

ρ

V

Vpq

I

Vp

Vq

Pi = VpI

Qi = VqI

“i”

Alloweddomain for voltage injection

P

Q

δo Qo,Po

Qi,Pi “i”

δ

ρ Limit for max line voltage

Limit for min line voltage

Figure 8-25 Illustration for determining the operating region for injected voltage from power flow control requirements and practical system and

8-46

2. Check if the injection of Vpqmax at any value of angle ρ will violate the allowed maximum or minimum value of the line voltage. If so, change the operating area of Vpqmax with the given voltage limits.

3. Modify accordingly the corresponding operating area of power transmission in the Q,P plane.

4. Check if the worst case operating point yielding the maximum value for Vpqmax is still within the modified area in the Q,P plane. If not, determine the new value for Vpqmax (and make a decision if the attainable transmitted active power and reactive power load with the defined voltage limit are still acceptable or additional measures are needed.)

5. Once the locus corresponding to the worst case Vpqmax is determined, the locus of the resultant line current phasor I can be derived and, with these, the active and reactive power, PUPFC and QUPFC, the series converter exchanges with the ac system at the boundary points of the P, Q operating domain, can be calculated.

6. The maximum magnitudes of the three variables Vpq, I, and PUPFC, obtained from the sets of values defining the boundary of the operating area, will define the rating of the UPFC to meet the series compensation requirements for this operating region at the terminals of voltage injection. (Note that the computational process for generation of the necessary data, illustrated above in a simplified manner, would, of course, be carried out automatically by appropriate computational routines of the power flow program and, possibly, only the selection and evaluation of the data would require special human attention.)

In the case of having a number of operating regions around a different initial Po, Qo, values in the Q,P plane due to changes in the transmission angle δ or other system parameters, it would become necessary to determine the maximum magnitudes of the three variables of interest, Vpq, I, and PUPFC, for each of the regions. The largest individual value found for voltage magnitude Vpq, line current magnitude I, and active power PUPFC among these, regardless whether they belong to the same or different operating regions, would define the ultimate rating requirements for the series converter of the UPFC.

If the application would involve power oscillation damping, or some other temporary compensation functions, then it would be necessary to assess by additional studies, or appropriate risk analysis, whether the incorporation of the additional function would require an increase in the rating of the series converter of the UPFC.

8.3.3.2 Issues of transient ratings

Considering possible transient rating requirements for the series converter of the UPFC, it should be remembered that practically any required increase over the specified maximum steady-state compensating voltage will likely become the design value for the actual steady-state rating of the converter, since, as mentioned before, power semiconductors do not have repetitive overvoltage capability. However, it may be possible to obtain some short term overvoltage operation at the expense of reducing the voltage margin of the converter (that the equipment is likely to have by design) for these infrequent and short time durations. (This approach implies certain operating

8-47

risks and thus its feasibility should be explored with the equipment manufacturer.) It should be made sure, too, that the coupling transformer of the series converter has sufficient voltage rating to accommodate the increased compensating voltage without saturation. In considering the need for overvoltage capability, it should also be kept in mind that the series converter of the UPFC, in contrast to conventional series capacitive compensation, is able to maintain (or even to decrease) the compensating voltage with increasing line current and, therefore, it can continue the operation if it has sufficient overcurrent rating to accommodate this system condition.

The series converter, depending on the design, is likely to have some (perhaps 20%) overcurrent capability for a short time, about one second, until the junction temperature of the power semiconductors reaches the maximum operating limit, typically 125 oC. The maximum overcurrent with the time duration during which operation is required has to be specified for the manufacturer and for economic reasons, need to be kept at minimum. Operation with higher than about 20% overcurrent and/or time durations exceeding about one second, will necessitate the increase of the actual steady-state rating of the equipment. This actually means that the converter valves or structure will be modified so that under normal operating conditions the current through the individual power semiconductors will be reduced (together with the corresponding junction temperature) relative to their rated value, in order to provide sufficient margin accommodate the necessary overcurrent conditions. Thus, overcurrent requirement can affect the equipment cost considerably. However, in contrast to transient overvoltage rating, short term operation with line currents significantly higher than the steady-state maximum will not require an increase in the rating of the coupling transformer.

If the line current, under fault or some contingency conditions, exceeds the overcurrent (magnitude and/or time) limit, then the UPFC protection (to be discussed later) would be activated to provide an appropriate by-pass for the line current.

Conclusions on the rating of the series converter.

1. The maximum rating of the series converter in the UPFC is defined by the maximum line current, the maximum injected voltage, and the maximum active power provided for the line for compensation under the worst case operating conditions when sustained operation is required.

2. The maximum line current, injected voltage, and active power may not occur simultaneously under a single operating condition. Still, the converter must be rated individually for each of these variables.

3. The series converter of the UPFC is an equipment based on power semiconductors. As such, in contrast to conventional power equipment, it has no inherent overvoltage and only very short duration overcurrent capability. The manufacturer usually includes some safety margins for voltage and current ratings for increased reliability. If system operation does require equipment functioning under overvoltage and/or overcurrent conditions, this should be precisely specified for the manufacturer with the understanding of probable cost consequences. Overvoltage requirements usually considered as steady-state, almost independent of practical time durations, thus will increase equipment rating and cost. Moderate overcurrent capability may be provided for short time, up to about one second, after which it would also become a steady-state requirement with cost increase.

8-48

8.3.3.3 Shunt converter rating to meet UPFC operation requirements

In considering the rating requirements for the shunt converter of the UPFC, the following application options and asset investment issues should be considered:

(1) The sole purpose of the UPFC is to provide the prescribed two-dimensional series compensation for active and reactive power flow control in the line. Support of transmission voltage at the bus of connection is not necessary (now or ever).

(2) In addition to the required two-dimensional power flow control, the transmission voltage needs now support by reactive compensation at the bus of connection.

(3) In spite of not needing now transmission voltage support in addition to the required two-dimensional power flow control, there is a concern that unpredictable system, load or operational changes will necessitate the support of the transmission voltage to at the UPFC bus in the future.

The rating considerations for shunt converter of the UPFC under the above three operational scenarios would be distinctly different.

Under option no. 1, the rating of the shunt converter is determined simply by the active power, PUPFC, that the shunt converter needs to supply at its dc terminal from the ac bus for the series converter to execute the prescribed two-dimensional compensation. (The shunt converter thus will have no reactive compensation capability.)

Under option no. 2, the reactive power rating of the shunt converter is to be determined by normal load flow studies under prescribed system conditions to maintain the bus voltage within defined limits. Once the maximum reactive rating of the shunt converter is established, the total shunt converter MVA can be determined by the simple formula:

MVAshunt = [Q2

shunt + P2

UPFC]½

Where Qshunt is the maximum reactive power the shunt converter has to supply to, or absorb from the ac bus to keep the bus voltage within defined limits, and PUPFC is the maximum active (real) power that it has to provide at the dc bus of the series converter to sustain the two-dimensional series line compensation in the total P-Q operating domain specified. (Note that in practice PUPFC would also include the losses of the two converters - about 1.5-2.0 % of their total MVA rating at maximum output - in order to maintain the required dc link voltage.)

Under option no. 3, there is no definite requirement for reactive shunt compensation. However, the formula established for option no. 2 provides a useful indication, from the standpoint of asset investment, of the high economic incentive for incorporating reserve capacity into the shunt converter for reactive compensation. This is because, as seen in the simple expression above, the active and reactive power sum as two quadrature or phasors and, thus, for a moderate overall VA rating increase, much more than proportional dynamic var compensation capacity can be realized. Consider, for example, that the required active power rating of the shunt converter to support the series compensation is 100 Mw. Then, in order to also provide a maximum of 100 Mvar controllable capacitive (and inductive) reactive compensation capability, the rating of the shunt converter would have to increase only by 41% (•2) to 141 MVA. Considering further that

8-49

(a) the continuous variable reactive output from 100 Mvar inductive to 100 Mvar capacitive represents 200 Mvar dynamic compensation range; (b) this dynamic range can simply be converted into a zero to 200 Mvar range with the addition of a fixed 100 Mvar capacitor bank, should the requirement for voltage support would increase in the future; and, (c) the control, auxiliary equipment, and the operational infrastructure hardware, as well as much of the installation labor, are already included in the cost of the basic UPFC (having no shunt compensation capability), it is quite evident that the inclusion of reactive shunt compensation capability with the UPFC is, in general, an economically savvy decision for transmission asset investment.

8.3.3.4 Issues of transient rating

The main issue to consider is the maximum system overvoltage at which the shunt converter must be kept operating. Under overvoltage condition the converter would need to be absorbing reactive power by drawing inductive current, the magnitude of which must be limited to the maximum rated converter current. In order to do that, the converter control must increase the dc link voltage in order to enable the converter to produce a large enough output voltage so that the corresponding output current determined by the difference between the system and converter voltage will not exceed the maximum allowed value, i.e., (Vline-Vconverter)/Xtransformer ≤ Iconvertermax). It is evident that the system overvoltage condition must be carefully evaluated from the standpoint of operation benefit versus cost to specify the level at which the converter operation is required since, above a nominal design margin, it does directly affect the converter rating and cost. It is worth mentioning that at relatively high overvoltage (which typically would be caused by load rejection), not only the coupling transformer of the shunt converter, but all other system transformers in the affected part of the system would saturate, resulting in a very large overall var absorption, albeit with very distorted current. Thus, under this condition the converter’s var absorption may be irrelevant, and its operation could anyway be interrupted by harmonics caused high peak current exceeding the turn-off capability of the valves.

The overcurrent considerations within the normal operating voltage range are similar to those characterizing the series converter. With increasing overcurrent requirement (magnitude and time duration), the cost of the equipment will increase in proportion. It should be noted that if system operation would require higher than maximum rated current from the shunt converter, the control could be programmed to given either a pre-defined or condition based priority to supply first the active current need of the series converter for the desired power flow, or use the capacity primarily for reactive compensation to regulate the bus voltage.

Conclusions on the shunt converter rating

1. The shunt-converter rating is primarily determined by the active power of the series compensation requirement.

2. The shunt-converter provides an economic option for bus voltage regulation by reactive shunt compensation due to the fact that the total MVA is not the direct sum of the active P and reactive power Q, but the square-root of their squares, i.e., MVA = [P2 + Q2]½.

3. Transient rating consideration for overvoltage and current, although are somewhat different impact on application consequences, from the standpoint of cost are similar to those of the series converter.

8-50

8.3.3.5 UPFC rating optimization by combined compensation

As previously explained and illustrated (e.g., in Fig. 8.20), the unconstrained voltage injection of the UPFC is in a circular region around the end of the transmission bus voltage phasor. This means that the UPFC voltage injection in general results in an advance or a retard of the prevailing transmission angle. Thus, the UPFC could be considered to have two equal half circle segments as operating regions, one characterized by the advancement, the other by the retardation of the effective transmission angle. Although this inherent capability of the UPFC to operate symmetrically in both regions may well be utilized in many practical applications (e.g., power oscillation damping), there are other applications in which voltage injection resulting in unequal operating regions for advancement and retardation, or in unidirectional change, either advancement or retardation, of the transmission angle is quite satisfactory. For example, if the attainable transmission angle is too small for the desired power transmission, the UPFC would always have to provide first a leading an essentially fixed phase angle advancement to establish the correct steady-state operating point around which the control of active and reactive power would be executed under the prevailing system conditions. In this type of applications the series converter of the UPFC would not be utilized effectively because with a given MVA rating it could control, for example, the transmission angle over the range of -σmax ≤ σ ≤ +σmax, but it would actually be used to control the transmission angle over only either the positive range of 0 ≤ σmax or the negative range of -σmax ≤ 0. In other words, the rating capacity of the converter is utilized only 50 per cent. Another consideration in this type of application is that a significant portion of the UPFC rating may be used up for steady-state angle control, which could be provided by more economical means.

The operating requirements for unequal operating range and steady-state angular shift can be satisfied if the UPFC is combined with a phase shifting transformer providing fixed or selectable angle of advancement or retardation. A possible circuit arrangement to implement this hybrid arrangement is shown schematically in Fig. 8.26a. As seen, the overall circuit arrangement includes the usual UPFC configuration with two converters, one is in shunt- and the other is in series-connection with the transmission line. This hybrid arrangement, in its simplest form, however, also includes an additional winding on the secondary (converter) side of each phase of the shunt-connected coupling transformer. (In other arrangements, the phase shifting could also be accomplished by a separate transformer, and in both arrangements a tap-changer could be applied to adjust the steady-state phase shift.) The transformer connection is such (e.g., delta-connected primary) that the voltage obtained at the phases a, b, and c of the secondary windings are in quadrature with the phase a, b, and c phase-to-neutral primary (line) voltages. Considering one phase, for example phase a, it is seen that this winding is connected in series with the secondary winding of the series-connected coupling transformer and the phase a output of the series converter. As a result, the voltage injected in series with the line to control the flow of power through it is the vectorial sum of the fixed voltage provided by the shunt-connected coupling transformer and the controllable output voltage of the series converter.

8-51

Figure 8-26 UPFC combined with (fixed) phase-shifting to reduce its rating to that necessary only for handling frequent angle variations and dynamic compensation needs

The operation of the hybrid power flow controller is summarized with reference to the phasor diagram shown Fig. 8.26b. As explained above, the total injected voltage, used to control the power flow in the line is made up of two components: Voltage component Vσ, which is the fixed quadrature voltage provided by the shunt connected transformer to advance (or retard) the existing transmission angle by a fixed angle σ, and voltage component Vpq, which is the controllable component provided by the series converter of the UPFC. The magnitude of Vpq is variable in the range of 0 ≤ Vpq ≤ Vpqmax (<Vσ) and its angle ρ in the range of 0 ≤ ρ ≤ 2π with respect to the fixed π/2 angle of Vσ. The magnitude and angle of the effective transmission voltage, V’, at the bus of the transmission line is obtained by vectorially adding the total injected voltage Vσ + Vpq to the original bus voltage V. In the hybrid arrangement the circular operating region of the UPFC is centered around the end of the voltage phasor Vσ, providing in effect a fixed angular shift for voltage phasor V and, with the stipulation of Vpqmax •Vσ, the total control region allows only unidirectional change (retardation or advancement depending on the sign of angle σ) for the transmission angle. (Note that the figure illustrates the case of angle advancement.)

The economic benefits of the hybrid arrangement for suitable applications, such as the present example of steady-state angle advancement, can be further explained with reference to Fig. 8.26c that shows the resulting operating region of the power transmission in the Q,P plane.

Without any compensation, the “natural” steady-state transmission angle of the line is o. The corresponding power is too small for the adequate utilization of the line. The obtain the desired steady-state power transmission, the existing transmission angle is increased by an additional angle σ to yield the desired overall angle of o+σ. This steady-state angular phase shift is simply accomplished by a fixed-angle phase shifting transformer. In this way, the required rating of the UPFC is significantly reduced and used only for optimizing and managing the real–time power transmission demands (minimization of reactive power flow, matching load demands, counteracting dynamic disturbances, etc.) in the desired operating Q,P domain (shown by the

(c) Q-P

Controllable region with a large UPFC

Q

σδ +o

δoP

(b) Phasor

p=V’ +V Vσ+V

Vpρ

+V Vσ

Vσ π/2

V

ShunConverte Converte

Serie

∆

V

σ V pq V

I V V σ + pq V’

(a) UPFC with Phase

8-52

solid-line circle in Fig. 8.26c) around the fixed operating point corresponding to the effective transmission angle, o+σ, obtained by the fixed phase shift.

The stipulated transmission requirements could, of course, be met with a single UPFC of much higher rating than that needed with the phase-shifter (over 2.5 times higher in the illustrative example), as shown by the large circular operating domain inside the dashed line boundary in the Q,P plane of Fig. 8.26c. However, as can be observed, the advancement of the transmission angle to the required operating region would also establish a much larger operating domain (including capability for power flow reversal and reactive power flow control over a wide range from inductive to capacitive) than the application requires, at a cost that could not be justified.

8.3.4 Rating considerations for the IPFC

The basic Interline Power Flow Controller is comprised of two converters in back-to-back connection, the ac outputs of which are connected in series with a different transmission line. Each is able to provide independent series reactive compensation for its own transmission line to substantially control the active power flow in it. However, the back-to-back connection of the converters at the common dc bus establishes a path for active power flow between the two lines enabling them also to execute active power compensation in a pre-determined way to control reactive line power flow.

The operating role of the two converters is the same for reactive power compensation, but different for active power compensation. That is, whereas each converter can execute the reactive compensation independently of the other, one converter needs the active power support of the other for active power compensation. This means that one converter has the role of providing the desired reactive and active power compensation of the priority line (the line in need of power flow optimization), while the other is in support role, providing the necessary (positive or negative) active power from its own line to support the compensation of the priority line.

Since the capacity of the two lines may be different and, depending on the prevailing transmission requirements and line loading, the roles of the two converters may need to be reversed, the rating and operating considerations for the IPFC can become complex requiring repetitions of load flow studies to define the desired and sustainable operating conditions. However, the principles involved are quite simple and are summarized below for guidance.

1. Determine the series reactive compensation requirements for each line independently, as if only a controllable series reactive compensator was to be applied, to define the necessary rating of each converter to execute the reactive compensation under the stipulated operating conditions. From these studies, the maximum current and voltage rating of each converter, which may not occur simultaneously, for the required reactive compensation can be established. Note that under worst case contingency conditions, such as partial equipment failure, line failure, or other reasons which prevent active power transfer from one line to the other, this is the compensation that will be available for the line.

2. Define the criteria for priority compensation. Determine the rating of the priority converter as if it was the series converter of a UPFC, using the process described in the previous section with

8-53

the maximum (quadrature) voltage and line current obtained in 1. (Note that the maximum quadrature voltage will define the radius of the unconstrained circular operating area for the injected voltage.) The required end result of this process is the maximum voltage, line current, and active power rating that the priority converter of the IPFC must have for the maximum reactive and active power compensation the priority line may require.

3. Determine if the line in support role is able to provide the maximum active power for the priority line. The consideration here is that active power usually has to be supplied to the priority line to decrease the reactive power flow in it. Consequently, the corresponding active power absorption from the support line will increase the reactive power flow in it. Therefore, the general requirement for sustainable operation of the IPFC is that either the support line is a higher capacity line than the priority line, or that it is less loaded than the priority line.

4. If the support line has the capacity to provide the required maximum active power for the priority line, then the rating of the support converter is determined from the appropriate combination of the maximum reactive compensation required for its own line and the maximum active power need of the priority converter compensating the other line.

5. If the support line does not have the capacity to provide the required maximum active power for the priority line, then the process has to reversed to (a) finding first the attainable active power of the support line; (b) determining with that the ratings of both the support and priority converters; (c) assessing whether, with the constraint of reduced, or not always available active power transfer capacity, the IPFC still provides enough benefit in terms of power flow optimization, reactive loop flow minimization, damping capability, etc. to justify its practical implementation; considering the possibility of adding a third converter in shunt-connection to support both series converters in active power compensation, and also providing the option for bus voltage regulation. In assessing the possible options and benefits, it should be kept in mind that the extension of a two-line, converter-based series compensating system into an IPFC is practically “free”, since it requires no additional power components, only suitable control algorithm, and the addition of a shunt-connected converter for supporting reactive power flow control in the lines can also provide very economical means for shunt reactive compensation.

8.4 Protection considerations

The generalized power flow controllers using combinations of shunt and series connected converters with common dc link require different types of protection for their series and shunt connected constituents. Thus the protection issues of the UPFC and IPFC, and their generalized multi-line variants, can generally be broken down to the individual protection of the series-connected and the shunt-connected constituent converter.

8.4.1 Protection of the series converter

All types of compensators and power flow controller connected in series with the line require special rating and/or protection considerations to handle surge currents caused by line faults or other system disturbances. One consideration is, of course, whether the series connected equipment has sufficient current rating to handle the worst case line surge currents. The other

8-54

consideration is whether it would have sufficient voltage rating, or suitable voltage limitation characteristic, to support or limit the voltage that would develop between the two connection-points in response to the flow of line surge current. The voltage-sourced converter generally has some limited inherent surge current capability, but for rather short time duration (from a few cycles up to about 60 cycles). On the other hand, the output voltage of the converter (i.e., the injected series compensating voltage) can be controlled independent of the line current. Thus, the rating and protection issues of series-connected converter are very much dependent on the particular application, primarily on the prevalent line impedance (“weak” or “strong” line) and the desired behavior of the Controller during surge current conditions. In general, there are three possibilities:

(1) The line is “weak” and the possible surge current is within (the economically practical) rating of the converter;

(2) The converter during surge current conditions is put in a “bypass” operation to provide a path for the line current at zero voltage injection; and

(3) The converter limits the injected voltage to a maximum value for line current higher than a given maximum.

The simplest case is the rare (almost non-existent) very “weak” line case, in which the converter has sufficient current rating to handle the anticipated line fault current while maintaining the (reactive) voltage injection up to the maximum voltage rating of the converter. In this case, there is no other protection than appropriate control algorithm together with the practically compulsory breaker bypass to handle converter or control failures, or thermal problems due to a sustained fault.

The protection approaches based on bypass or voltage limitation ca be implemented on the basis of two protection philosophies. One is to provide a fast solid-state bypass and the other is to have some kind of voltage limiter parallel with the voltage injection terminal, which takes over the current conduction, but maintains the injected voltage below a given maximum, when the line current exceeds the normal operating range. Some implementation possibilities of these approaches are illustrated in Figs. 8.27 and 8.28. Due to the large range of possible surge currents at different line voltages and system structures, as well as to the variety of differing application requirements and constraints, it is not possible to devise or select a single protection solution that would uniquely meet in all cases the operation needs and cost objectives. For this reason, in this section the basic practical possibilities for protection are summarized only for the purpose of giving guidance for selection and evaluation in different application the designer may encounter.

The first arrangement in Fig. 8.27 illustrates the case of handling modest line surge currents by operating the converter switches so as to provide an internal bypass via two series-connected valves with zero voltage injection. The idea behind this solution is that the power semiconductors, particularly those in the thyristor family can conduct significantly higher current, and can be operated at higher junction temperature, when they are kept in full conduction than when it normally operated with repeated turn-on and off. The main feature of the arrangement is that it is “free” (no extra components are needed), and it may be a sufficient practical solution, particularly for relatively low voltage and weak transmission lines. (One

8-55

actual installed UPFC at AEP’s Inez substation employs this method.) In the future, when the anticipated silicon-carbide power semiconductor devices become available, this solution may be more generally applicable due to their anticipated much higher surge current and operating temperature capability.

Figure 8-27 Current bypass approaches to the protection of series-connected converter during line faults and major disturbances

The second arrangement allows the use of all four valves of the commonly used bridge- converter for bypass protection. This is achieved through the rapid disconnection of the dc capacitor by a fast auxiliary valve constructed of suitable power semiconductors (e.g., IGBTs), and by the simultaneous gating-on of all converter valves. This arrangement in effect doubles the bypass capability of the previous one at a very modest cost and thus it would likely be satisfactory in appreciably increased number of practical cases.

In connection with these two approaches it should be noted that in many actual converter configurations several individual converters may in effect operate in parallel. With this type of structures, of course, the role of the single valve shown in both illustrations would be replaced by a number of valves sharing the current in parallel connection.

The third arrangement in Fig. 8.27 illustrates the case of handling high surge currents by an external thyristor bypass switch of sufficiently high short term current rating. This approach is quite universally applicable due to the very high attainable surge current rating of conventional thyristors (i.e., those having no turn turn-off capability). It should be mentioned that this, as well as the previous two solutions would be backed-up by a fast mechanical breaker to handle sustained fault currents under unsuccessful fault clearing.

+

1. Bypass protection by single

converter valve

Line

Converter valves

+

2. Bypass protection

by two parallel

Line

Converter valves

+

Line

Thyristor b

3. Bypass protection by external

Converter valves

8-56

The first arrangement in Fig. 8.28 is similar to that used for the protection of series capacitors. A (set of) ZnO arrester(s) of appropriate rating provide a bypass for the line current at the break-over voltage of the arrester(s). Since the ZnO break-over voltage needs to be about twice as high as the peak (injected) voltage during normal operation, the converter voltage rating must be correspondingly high. This is because during surge currents, the converter gating would be blocked, and the dc capacitor would get charged to the maximum voltage determined by the limit of ZnO arrester via (the series string of) the anti-parallel diodes in the valves that the (string of) controlled semiconductors (e.g., GTOs) in off-state would have to be able to block.

Figure 8-28 Voltage limiting approaches to the protection of series-connected converter during line faults and major disturbances

The second arrangement shown in Fig. 8.28 solves the voltage rating problem of the conventional ZnO limitation by applying a thyristor-switched ZnO voltage limiter. The ZnO break-over voltage is now equal to the maximum voltage rating of the converter required for normal operation (with sufficient safety margin). In normal operation, the thyristor switch is open (blocked) and a low current voltage divider reduces the prevailing operating voltage across the ZnO arrester to a safe design value (about half of the converter injected voltage). Under surge current conditions, the thyristor switch is gated on, the converter gating is blocked, and thus the ZnO is forced to conduct the line current and limit the converter terminal voltage to the nominal maximum. Note that this can be an economical solution in many applications, since the voltage rating of both the thyristor switch and the ZnO arrester is half of that required in the previous arrangement.

ZnOLine

+

1. Protection by ZnO arrester with

converter off (high converter

voltage rating)

Converter valves

ZnO TSW

+

Line

2. Protection by switched ZnO arrester with

converter off

Converter valves

+

Line

3. Protection by dc voltage limitation with converter off

(high current converter diodes)

Converter valves

8-57

The third and final arrangement is different in that the voltage limitation during surge current conditions is provided by a (controlled) voltage limiter across the dc capacitor. In this case, the controlled semiconductors in converter valves are again blocked and the line current through the valve diodes is allowed charge the dc capacitor. However, as soon as a preset dc voltage level is reached, the auxiliary thyristor switch across the dc capacitor turned-on to discharge the capacitor via appropriately rated linear or non-linear resistors. At the end of the of the fault conditions, the thyristor switch is commutated off as the converter is brought back to normal operation. This protection arrangement would require the converter diodes to have a sufficient surge current rating which may not be compatible with their switching requirements under the normal converter operation, particularly if pulse-width-modulation (PWM) is used for output voltage construction and control.

It is pointed out that the above voltage limiting approaches, just as well as those with the previously discussed current bypass arrangements, are complemented with a suitable mechanical bypass breaker to handle equipment failures and other unforeseen contingences. In addition, it is usually a normal practice to apply appropriate disconnect switches to be able isolate and remove failed equipment from the line after breaker bypass is applied.

It should also be kept in mind that, as mentioned before, the control system of the series converter does have sophisticated protection functions which monitor internal circuit parameters, component failures, redundancy conditions, and the overall performance of the system and give alarm indication or, in extreme cases initiate shut downs, if parameter limits are exceeded or reliable operation no longer can be ensured due to internal component failures or circuit malfunctions.

8.4.2 Protection of the shunt converter

The protection of the shunt-connected converter doest, in general, require no external means, except for the normal circuit breaker connecting the coupling transformer to the system bus. However, the internal, control-based protection of the shunt converter has the functional task of ensuring that the maximum instantaneous current does not exceed the turn-off capability of the valves in the converter. Similarly, the converter protection must ensure that operating voltage is kept below the defined maximum. Since the internal protection usually results in temporary suspension of operation or total shut-down, it is important to determine the necessary operating voltage and current limits within the process of defining the specified ratings of the equipment as, discussed in the previous Rating considerations sections.

Overvoltage exceeding operation limit results in suspended gating and subsequent shut down. High instantaneous current magnitudes in the converter valves (usually due to harmonics caused by saturated transformers or large system voltage unbalances) result in temporary short duration stoppage (from a fraction of a cycle to one or two cycles, or until the condition sufficiently improves) in the operation of the converter by gate signal suppression, followed by automatic synchronous restart. The instantaneous current magnitude caused operation stoppage can be greatly reduced by sophisticated control techniques which can keep the converter output current substantially sinusoidal in spite of external transformer saturation and system voltage unbalance.

8-58

8.5 Application examples

The generalized power flow controller concept, using back-to-back voltage-sourced converters is relatively new. There only a few installations, some of these have convertibility for multi-functional operation. Two examples are described below, both are the first of their kind. The first example is AEP’s United Power Flow Controller (UPFC) at the Inez substation, and the other is NYPA’s Convertible Static Compensator that can be operated as Interline Power Flow Controller (IPFC).

8.5.1 UPFC at AEP’s Inez station

The first Unified Power Flow Controller (UPFC) in the world, with a total rating of ±320 MVA, was commissioned in the mid 1998 at the Inez Station of American Electric Power (AEP) in Kentucky, USA, for voltage support and power flow control. The project was jointly sponsored by the Electric Power Research Institute and AEP, and designed and manufactured by the Westinghouse Electric Corporation.

8.5.1.1 Background and planning information at the time of installation

System description. American Electric Power is an investor owned electric utility company which operates across seven Midwestern states, from Michigan in the north and west, to Tennessee in the south, and Virginia in the east. These facilities, including the generating plants, 765 kV extra high voltage (EHV), 345 kV high voltage, and 138 kV transmission lines, distribution system, and associated stations, are all interconnected and operated as a major power network within the eastern US power system. AEP provides electric service to approximately 1.7 million customers in an area of about 48,000 square miles (124,313 square km).

AEP produces over 124 billion kilowatt-hours a year, representing a peak system demand of over 26,000 MW. A major portion of the System's generation is located on the Ohio river and its tributaries. The areas south and east of these rivers depends on long transmission lines to meet the electric demand. The load area of concern regarding this project is one of these areas, the Inez Area identified in Fig. 8.29. The project involves the Inez Area and the neighboring Tri-State Area. These two areas are quite different in their physical and electrical characteristics. The reinforcement plan adopted to mitigate the problems strongly ties them as one area.

8-59

Figure 8-29 Inez and Tri-State areas of the AEP power system

The Inez Area located in eastern Kentucky is rural in nature and has a population of about 670,000 which is spread over an area of about 6,300 square miles (16,316 square km). Generating plants and EHV/l 38-kV substations are located only at the periphery of the area. The Inez Area's winter power demand of approximately 2000 MW is served by several long 138-kV transmission lines. System voltages and area reactive power requirements are supported by a ±125 Mvar Static Var Compensator (SVC) and a large assortment of switched shunt capacitor banks located at several 138-kV and lower voltage sub-transmission stations.

The Tri-State Area located adjacent to and north of the Inez Area has a population of about 480,000 people is spread over an area of approximately 3,500 square miles (9,064 square km). The Tri-State Area load is summer peaking with a requirement of about 1400 MW. It is served by a strong network of EHV lines, EHV/138-kV substations, and generating plants within and near its boundaries. The area transmission system supports power flows into the Inez Area via 138-kV inter-area line connections.

8-60

System problems. The Inez Area depends on long 138-kV transmission lines to support its customers' demand. Even under normal system conditions, many 138-kV transmission lines carry power flows reaching 300 MVA. These power flow levels are well above the surge impedance loading for 138-kV lines and as such are indications of the system stress. Moreover, these high loadings, as compared to the thermal capabilities of the lines, left little margin for system contingencies. In addition, voltage levels in the Inez Area were generally as low as 95% of nominal, which is considered to be the lowest acceptable voltage level for reliable service. Heavy 138-kV line loadings and comparatively long line distances resulted in excessive voltage gradients across the transmission network, further indicating stress on the system.

The Tri-State Area, on the other hand, is served by a strong network of EHV lines and EHV/138-kV station. The load is concentrated in and around a comparatively smaller area. The 138-kV line exits are heavily loaded over relatively short distances. Voltage levels are maintained at all substations within a narrow bandwidth around 98% of nominal.

Several single contingency outages in the Inez Area could have resulted in depressed voltage and/or thermal overload conditions. As an example, a simulation of a 765-kV line outage, during winter peak load periods, indicated increases in system active and reactive power losses by about 85 Mw and 700 Mvars, respectively. The need to supply these increased system losses under a contingency condition, would have further reduced the area voltages to levels which would certainly have resulted in wide area customer complaints. In addition, underlying transmission system would be loaded to well above their winter emergency capabilities. A second contingency under these circumstances was intolerable.

Single contingency outages in the Tri-State Area would not have caused any thermal overloads or depressed bus voltage conditions. However, double contingency outages involving one or more EHV/138-kV transformers or 138-kV generating units would have resulted in thermal overloads of the EHV/138-kV transformer and/or 138-kV line exits.

System Reinforcement Plan. The analysis of the system performance indicated that additional voltage support and power supply facilities were required to resolve normal, single contingency, and double contingency problems in the Inez Area. In addition, the EHV/138-kV transformer and line exit capacities of the Tri-State Area needed to be increased.

Following extensive analysis, the conclusion was reached that constructing a high capacity 138-kV line having thermal capabilities approaching a 345-kV line would provide an economical means of adding thermal capacity to the area. However, such a high capacity line would not carry its share of line loading according to its capacity. Power flow on such a line would still be governed by its impedance and other ac transmission system parameters. Concurrent with thermal considerations, peak and off-peak voltage performances in the Inez Area dictated the need also for a dynamic voltage support facility in the area.

Series capacitors together with Static Var Compensators and converter-based FACTS Controllers were then evaluated to enhance the power flow and provide sufficient voltage support. The evaluation showed that a combined converter-based FACTS Controller, such as the Unified Power Flow Controller, which can provide both the voltage and line flow control capabilities was a logical choice to be an integral part of the overall reinforcement plan for the Inez and Tri-State areas, which include the following:

8-61

• A high capacity 950 MVA, 138-kV line between the Big Sandy and the Inez Stations.

• A ±320 MVA UPFC at the Inez Station to fully utilize the high capacity of the new 138-kV line, provide dynamic voltage support, and control several mechanically switched capacitors in the area.

• A 345/138-kV transformer at the Big Sandy Station in the Tri-State Area to provide for the power flow requirements of the new high capacity 138-kV line.

• Series reactors to constrain loadings on existing thermally limited facilities.

The reinforcement plan was executed in two phases. The first phase of the plan included the installation of a 345/138-kV transformer bank at the Baker/Big Sandy Station and line power flow limiting series reactors at two stations. The last portion of this first phase involved the installation of the ±160 MVA shunt converter part of the UPFC at the Inez Station. The 345/138-kV transformers were designed to have sufficient capacity to provide the Tri-State Area needs and fulfill the future loading requirements of the high capacity Big Sandy-Inez 138-kV line. The series reactors installed to limit the loadings on critical, but low capacity lines. During the first phase the shunt part of the UPFC functioned as a STATCOM and supported the reactive power and the dynamic voltage needs of the Inez Area. In addition, it provided signals to control the switching operations of several 138-kV shunt capacitor banks in the area.

The second phase of the plan included the construction of a high capacity (950 MVA) 138-kV line between the Big Sandy and Inez substations, installation of the series part of the Inez UPFC and two 138-kV mechanically switched shunt capacitor banks at the Inez Station. The series converter was specified to be identical to the shunt converter in order to increase the operating flexibility of UPFC.

The third and the last phase of the project included additional 138-kV line construction, switching configuration changes at several stations, and additional reactive power correction to supplement the existing system reactive power margins.

8.5.1.1 UPFC Operation Strategy

Voltage Control. The UPFC is required to regulate Inez substation 138-kV bus voltages and control six 138-kV shunt capacitor banks (a total of over 330 Mvar) located at the Inez and three other nearby stations. This way proper capacitor switching strategy is established to reduce daily and seasonal voltage fluctuations to within acceptable limits. During system disturbances, mechanically switched shunt capacitor banks and associated controls are generally slow to react. Over sixty shunt capacitor banks are connected on the 138-kV and lower voltage transmission system in the Inez Area. Under actual system contingency conditions, all of these banks may not switch-on (hunting concerns) or some may over-correct the voltage and lock-out. To resolve this situation, the UPFC is required to maintain a pre-determined reactive power margin to maximize the shunt converter’s dynamic reactive power reserve for system contingency conditions. This ensures that the controllable reactive power range of the shunt converter (from -160 to +160 Mvar, i.e., a maximum control range of 320 Mvar) can be made available at all times to compensate for dynamic system disturbances. The shunt capacitor banks will be switched on and off to maintain the reserve UPFC and SVC margins during steady state load fluctuations.

8-62

Power Flow Control. With all transmission system facilities in service, the Big Sandy-Inez 138-kV line loading to be maintained at a level which would minimize system losses. For the peak and near peak load conditions, load flow studies indicate that a 300 Mw line loading would meet this objective. Actual flow on the line could be higher depending on other prevailing system conditions. The UPFC, therefore, may be required to slightly reduce the line loadings. Line reactive power flow and its direction will be monitored to help maintain the dynamic reactive power margin of the shunt inverter.

The series power flow control becomes important during contingency conditions. Major double contingency outages heavily load critical 138 kV facilities (especially the Big Sandy-Beaver Creek line). Controls of the UPFC have to monitor loadings on the three critical lower capacity lines. The control objective is to increase the Big Sandy-Inez 138 kV line loading to decrease loadings on the three lines. This control is to be activated as soon as any one of the three line loadings exceeds 90 percent of their respective emergency thermal ratings. The UPFC has to increase the Big Sandy-Inez line loading until the critical line loadings are reduced below the defined levels or the UPFC reaches its rating limit. Under severe contingency conditions, the UPFC-controlled Big Sandy-Inez line will be capable of transferring 950 MVA (4000 A).

8.5.1.2 Description of the UPFC

The Unified Power Flow Controller for the Inez Station was designed to meet the above defined system requirements. In particular, to provide fast reactive shunt compensation with a total control range of 320 Mvar (-160 Mvar to +160 Mvar) and control power flow in the 138 kV high capacity transmission line, forcing the transmitted power, under contingency conditions, up to 950 MVA.

In order to increase the system reliability and provide flexibility for future system changes, the UPFC installation was required to allow self-sufficient operation of the shunt converter as an independent STATCOM and the series converter as an independent Static Synchronous Series Compensator (SSSC). It is also possible to couple both converters together to provide either shunt only or series only compensation over a doubled control range.

Power Circuit Structure. The UPFC equipment comprises two identical Gate Turn-Off (GTO) thyristor-based converters, each rated for ±160 MVA. Each converter includes multiple high-power, 3-level, 48-pulse GTO valve structures feeding an intermediate (low voltage) transformer rated for approximately 80 MVA. The converter and magnetic configuration provides harmonic neutralization and thus the output is a three-phase voltage set of nearly-sinusoidal quality that is coupled to the transmission line via the intermediate transformer by a conventional (three-winding to three-winding) main coupling transformer rated for 160 MVA. The converter-side voltage of the main transformer is 37 kV phase-to-phase (for both shunt and series transformers.) The shunt-connected transformer has a 138 kV delta-connected primary, and the series transformer has three separate primary windings each rated at 16 percent of the phase voltage. The converter structure is shown schematically with relevant information in Fig. 8.30. The installed converter is shown in Fig. 8.31.

8-63

Figure 8-30 Basic structure of the UPFC converters and the corresponding technical data

Figure 8-31 The UPFC converter hall

To maximize the versatility of the installation, two identical main shunt transformers and a single main series transformer have been provided as shown by the power circuit schematic at the left hand side of Fig. 8.32. (At the right hand side of the figure, the major components of the actual installation are identified.) With this arrangement, a number of power circuit configurations are possible. Converter 1 can operate as a STATCOM with either one of the two main shunt transformers, while Converter 2 operates as an independent SSSC. Alternatively, Converter 2

Each

Nominal rating: ±160 MVA Transmission voltage: 138 kV Converter output voltage: 37 kV (phase (to coupling transformer) to phase) Pole output voltage: 9.3 kV No. of converter pulses: 48 Type of converter pole: 3-level No. of poles: 12 No. of 3-phase converters: 4 Type of semiconductor: GTO Rating of GTO: 4.5 kV, 4kA No. of GTOs per valve: 8 (7+1) DC voltage: 10.5+10.5 kV DC capacitor rating: 110 kJ Note: All ratings are nominal

0 2 4 6 8 10 12 14 16

1

-1

0

P.U.

msec.

Converter voltage

Main transformer

Intermediatetransformer

48-pulse, 3-level converterVoltage limiter

+

+

+

+

138kV Line

8-64

can be connected to the spare main shunt transformer and can operate as an additional STATCOM. With the latter configuration a shunt reactive capability of ±320 MVA becomes available. The power circuit arrangement indicates the priority established for shunt compensation at this location.

Figure 8-32 Simplified UPFC power circuit in the AEP system at Inez and the corresponding practical installation

Control System. Both converters comprising the UPFC are controlled from a single central control system housed in three cabinets in the control room. Two of the cabinets house the relay interface and signal conditioning, while a single cabinet contains the control electronics. The conceptual structure of the control system is shown in Fig. 8.33. The actual control algorithms that govern the instantaneous operation of the two converters are performed in the real-time control electronics which employs multiple digital signal processors. The real-time control communicates with the two-valve pole electronics mounted on each pole of the converter via the valve interface that is linked to the poles by fiber optic cables. The status processor is connected to every part of the system, including the cooling system and all of the poles, by serial communications. During runtime it continually monitors the operation of all subsystems, collecting and analyzing status information. It is responsible for all startup and shutdown sequences and for organizing and annunciation of all alarm conditions. The status processor is serially connected to a graphical display terminal which provides the local operator interface. A hierarchical arrangement of graphical display screens gives the operator access to all system settings and parameters, and provides extensive diagnostic information right down to the individual GTO modules of each valve.

Big Sandy 138 kV Line

Series Xfmer

Spare Shunt Xfmer

Main Shunt Xfmer

Shunt & Series Intermediate

Xfmers

Heat Ex- changer

UPFC Building (Converters & Controls)

Converte

Big Sand

Shunt xfmer

Intermediate transformers

shunt

DC

Spare shunt xfmer

Series xfmer

Converte

8-65

Figure 8-33 Basic functional control structure of the Inez UPFC

8.5.1.3 Performance of the UPFC

In the course of commissioning the UPFC, tests were performed to verify its predicted capability. For this purpose, the UPFC was directed to produce large controlled swings of active and reactive power on the Big Sandy line, and sizeable swings of voltage at the Inez station, while measurements were recorded. It is emphasized that these swings do not represent "normal" duty for the UPFC in this application. Arbitrary variations of this kind can be disruptive to the power system because they force active power flows to be redistributed in the network and affect voltage regulation at other stations. AEP system operators defined acceptable boundary limits for the tests and, in addition, slow ramping functions were applied to the control references for the UPFC automatic power flow controller.

Three representative cases have been selected for the purpose of demonstrating the capabilities of the UPFC. The first two cases show the UPFC independently controlling line P, line Q, and Inez bus voltage respectively. In both cases, the UPFC is operating with the shunt converter in automatic voltage control mode and the series converter in automatic line power flow control mode. In the third case the series converter of the UPFC is operated as an independent series reactive compensator, SSSC. Each set of results is annotated using the sign convention as defined in Fig. 8.34. Note in particular that P and Q for the line are measured at the line-side terminals of the series insertion transformer. This is the actual power at the end of the line, and is defined as positive towards Big Sandy. The active and reactive power for each of the two converters is also shown. Note that the converters independently generate reactive power but that their active power is substantially equal and opposite. In each case a few stationary points

INTERFACE

ANALOG SIGNAL

CONDITIONING

STATUS PROCESSORCollects status data from all UPFC

Provides local and remote operatorsub-systems via serial communications

VALVE INTERFACE 2

AEP MASTER

CONTROLLER

LOCAL OPERATORINTERFACE(GENESIS)

and protective functions forthe four valves in each pole

Provides all local house-keeping

interface and sequencing for UPFC start and stop

Controls all GTO valve switchingfor shunt and series inverters

Three DSPs perform all control algorithms

Series converter

POLE ELECTRONICSProvides all local house-keeping

and protective functions forthe four valves in each pole

VALVE INTERFACE 1Shunt converter

POLE ELECTRONICS

Fiber optic cables

RELAY LOGIC

REAL TIME CONTROL

for all UPFC operating modes

COOLING SYSTEM

Fiber opticcables

8-66

(between major transitions) have been chosen and a phasor diagram drawn to represent the operating condition at that point. This is considered to be helpful, because it is difficult to interpret each situation from the time plots alone. The interpretation is made even more difficult by the fact that the power network "adjusts" following each major transition, especially with regard to the voltage phase angles at Inez and Big Sandy. The following five sections will take the form of a specific commentary for each of the cases.

Figure 8-34 Definitions of polarity conventions used for the commissioning tests of the AEP UPFC

Case 1: UPFC Changing Active Power (P)

Refer to Fig. 8.35. This case starts with the UPFC idling near zero injected voltage and the active power flow on the line near the "natural" level of 150 MW from Big Sandy. The shunt converter is regulating the Inez bus to 1.0 p.u. by generating about 60 Mvar capacitive, and about 36 Mvar are being delivered into the line. The objective for this case is to maintain the Inez bus voltage and the line reactive power flow Q unaltered, while making big step changes in line P.

Converter 2 (series)

Transmission line

to Big Sandy

V21

P

Inez 138kV bus

conv1 conv1 Q P

Converter 1 (shunt)

conv2Q

conv2 P

Q

V 1

V 2

8-67

Figure 8-35 Inez UPFC controlling active power P while maintaining constant reactive power Q in the line

The UPFC is first commanded to raise the line power to 240 Mw. It does this by injecting a voltage of about 0.16 p.u. roughly in quadrature (lagging) with the Inez bus voltage. To satisfy the required conditions, the shunt converter drops its capacitive output to about 20 Mvar, and the series converter delivers about 40 Mvar capacitive to the line. Active power exchange between the converters is about 8 Mw.

The second transition commanded is a 170 Mw drop in the line power to 70 Mw. This is accomplished with little change in the magnitude of the injected voltage, but about 180 degrees phase shift, so that the injected voltage is now still roughly in quadrature with the Inez bus voltage, but leading. The shunt converter produces about 85 Mvars capacitive, the series converter reverses its output to 10 Mvar inductive, and about 8 Mw flows between the converters through the dc bus. The final transition returns the system to the initial operating point.

Throughout this case the Inez bus voltage is tightly regulated at 1.0 p.u. and the reactive line power Q stays constant. Note, however, that the voltage, V2, applied to the transmission line, is lowered by a few percent relative to V1 to achieve the first swing and raised by a few percent for the second. It should also be noted that the large changes in active power arriving at Inez must, of course, be balanced by an equal and opposite total change in the active power on the other lines leaving the station. For this to happen a change in the phase angle of the Inez bus voltage, V1, is unavoidable.

Phasor diagrams corresponding to

line P and Q settings at t1=10, t2=33, t3=60

Time=60

V1 V2

(0,0

I line

Time=33

V V

(0,0

I line

Time=10

V1 V2

(0,0

I line

V1 = 138 kV bus voltage (regulated) V2 = Transmission line voltage to Big Sandy (controlled)

Active and reactive power of

P

Q

-300

-200

-100

0

100

200

0 10 20 30 40 5 60 70 80 90 100

Line active power (P) (MW) Line reactive power (Q) (Mvar)

-100

-50

0

50

100

0 10 20 30 40 50 60 70 80 90 100

Shunt converter reactive power (Qconv1) (Mvar) Shunt converter active power (Pconv1) (MW)

-20

0

20

40

60

80

0 10 20 30 4 50 60 70 80 90 100

Time

Series converter active power (Pconv2) (MW) Series converter reactive power (Qconv2) (Mvar)

Q

P

P

Q

Q

(a) Transmission line

(b) Shunt converter

(c) Series converter

8-68

At the time when these tests were performed the natural power flow on the line was too high for the UPFC to demonstrate its unique ability to reverse power flow. On other occasions, however, when the line has been lightly loaded, this has been demonstrated successfully and the UPFC has driven active power back towards Big Sandy.

Case 2: UPFC Changing Reactive Power (Q)

Refer to Fig. 8.36. The initial conditions for this test are similar to Case 1 but the objective is to regulate the Inez voltage at 1.0 p.u. and keep the line active power, P, constant while causing large steps in the line reactive power, Q.

Figure 8-36 Inez UPFC controlling reactive power Q while maintaining constant active power P in the line

For the first swing, the UPFC reference for Q is changed from +30 Mvar to -30 Mvar. After the change 30 Mvar is being received from the line, compared with the 30 Mvar delivered to the line initially. The UPFC forces the change by injecting about 0.05 p.u. voltage roughly in anti-phase with V1. The line voltage, V2, is consequently reduced in magnitude by about five percent. For the second step, the Q reference is taken to +100 Mvar (i.e., 100 Mvar to the line). This time the injected voltage is in phase with V1 so that V2 is increased by about five percent. The final step reduces the Q reference to zero. The line is now fed at unity power factor with V2 reduced by about 2.5 percent relative to V1.

It is interesting to note that the changes in Q at the line terminals are balanced almost entirely by equal and opposite changes in the reactive output of the shunt converter, which acts to maintain the Inez voltage. This brings to light a fascinating capability of the UPFC. In essence, it can


line Q and P settings at t1=10, t2=33, t3=60 seconds

Time=60

VV2

(0,0

I line

Time=33

VV

(0,0

I line

Time=10

V V2

(0,0

I line

V1 = 138 kV bus voltage (regulated) V2 = Transmission line voltage to Big Sandy (controlled)

Reactive and active power of


(b) Shunt converter

(c) Series converter

P

Q

-300

-200

-100

0

100

200

0 10 20 30 40 50 60 70 80 90 100

Line active power (P) (MW) Line reactive power (Q) (Mvar)

-20

-10

0

10

20

30

0 10 20 30 40 50 60 70 80 90 100

Series converter active power (Pconv2) (MW)

Series converter reactive power (Qconv2) (Mvar)

Q

-150

-100

-50

0

50

100

Shunt converter reactive power (Qconv1) (Mvar)

Shunt converter active power (Pconv1) (MW)

Q

P

0 10 20 30 40 50 60 70 80 90 100

P

Q

Time (s)

8-69

“manufacture” inductive or capacitive Mvars using the shunt converter and “export” this reactive power into a particular transmission line (i.e., the one with the series insertion transformer), without changing the local bus voltage and without changing the reactive power on any of the other lines leaving the substation. It can therefore regulate the station bus voltages at both the sending and receiving ends of a transmission line, while still freely controlling the active power flow, P, on the line.

Case 3: The series converter operating in SSSC mode

Refer to Fig. 8.37. For this case, the shunt converter is disconnected from the dc terminals of the series converter and is completely out of service. Consequently the Inez bus voltage is not regulated. The series converter injects voltage into the line essentially in quadrature with the prevailing line current. The injection angle deviates from true quadrature to draw active power from the line for converter losses and to charge and discharge the dc bus capacitor banks as operation requires. By means of the quadrature voltage injection, the SSSC is able to raise or lower the line current, but cannot independently alter P and Q. The SSSC operation is an important subset of full UPFC operation, since it can be used when the shunt converter is not available. Also, a single series-connected converter installation may be the most cost effective solution for applications, particularly for lines of moderate length, where a simpler form of power flow control is sufficient.

Figure 8-37 Inez UPFC Operated as an Static Synchronous Series Compensator (SSSC)


line power P settings at t1=10, t2=40, t3=60 seconds

Time=60

V1 V2

(0,0

I line

Time=40

V1 V2

(0,0

I line

Time=10

V1V2

(0,0

I line

V1 = 138 kV bus voltage (unregulated) V2 = Transmission line voltage to Big Sandy (controlled)

Active and reactive power of


(c) Inez bus and line

(b) Series converter

P

-300

-200

-100

0

100

200

0 10 20 30 40 50 60 70 80 90 100

Line active power (P) (MW) Line reactive power (Q) (Mvar) Q

Q

-20

0

20

40

60

80

0 10 20 30 40 50 60 70 80 90 100

Time (s)

Series converter active power (Pconv2) (MW) Series converter reactive power (Qconv2) (Mvar)

P

Q

0 0.95

1

1.05 INEZ bus voltage (V1) (p.u.) (unregulated)

Line voltage (V2) (p.u.)

V2

V1

10 20 30 40 50 60 70 80 90 100

Voltage of

8-70

The objective of this case is simply to show the SSSC raising and lowering the power on the Big Sandy line. The SSSC is operated to control the magnitude and the polarity of the injected voltage (i.e., the line power is not automatically controlled). Voltage injections are selected to give a sequence of approximately 100 Mw, 180 Mw, 250 Mw, and finally 200 Mw on the line. Note the corresponding changes in Q. From the phasor diagrams it can be clearly seen how Iline maintains a constant phase relationship to V1 and is always in quadrature with V21 . The natural flow on the line is between 100 Mw and 180 Mw. Consequently the polarity of V21 is reversed from lagging Iline to leading in this transition. The polarity reversal is accomplished by taking the converter dc bus voltage to zero, then raising it again with 180 degree phase shift in the converter output voltage.

8.5.1.4 Importance of results and possible future trends

This first UPFC installation at Inez has fully proven that the conceptual predictions regarding the unique operating and functional characteristics of the UPFC:

• Unique capability to provide independent and concurrent control for the active and reactive line power flow, as well the regulation of the bus voltage.

• Functional adaptability for a variety selectable operating modes and control functions.

• Flexible circuit structure that be reconfigured for independent shunt (STATCOM) and series (SSSC) compensation, as well as for only shunt or only series compensation at double rating.

• Operation in utility environment.

Considering the present technological trends and advancement, the UPFC would benefit most from technological developments in two areas, which could have great impact on its capital and operating cost, and reliability. One is in the device area which expected to yield more rugged, faster switching power semiconductors (such as the silicon carbide devices) with higher voltage and current ratings and, in particular, lower operating losses. The other is in power converter circuit area which, as one objective, would come up with a simple converter structure producing low output distortion, and, as a second and possibly more important objective, would create a suitable circuit topology in which the series converter is connected directly, without an insertion transformer, in series with the high voltage line.

8.5.2 IPFC at NYPA’s Marcy substation

The first Interline Power Flow Controller (IPFC) in the world, with a total rating of ±200 MVA, was commissioned as one of the several operating configuration of the two-converter Convertible Static Compensator (CSC) in 2002 at the Marcy Substation of the New York Power Authority (NYPA). The project was jointly sponsored by the Electric Power Research Institute and NYPA, and designed and manufactured by the Westinghouse (subsequently Siemens) FACTS and Custom Power Division.

8-71

8.5.2.1 Background information, system and equipment requirements

System description. New York Power Authority as a public benefit energy corporation was founded in 1931. It is the largest non-federal electric utility in the United States and a sale power supplier throughout New York State (26% of the State’s electric energy needs) and neighboring states with more than 42.3 billion kwh energy sales. It has over 7,000 Mw generating capacity, more than 1,400 circuit miles with 115, 230, 345 and 765 kV transmission lines. The NYPA power system is shown in Fig. 8.38.

Figure 8-38 NYPA Power System

System problems. Prior to the CSC project in the late 1990s, the maximum transfer limit of the Central-East interface was voltage constrained to 2,850 Mw (and even to lower under system outage contingency) and operated 100 Mw below this limit 75 per cent of the time. During 1996and 97, the system had 18 major emergencies and 192 alarm states. NYPA system The Inez Area depends on long 138-kV transmission lines to support its Studies indicated that certain contingencies could ignite oscillatory system conditions. The projected South-East New York load growth of greater than 3 per cent indicated the need to increase the transmission capability of the Central-North transmission corridor.

8-72

Objectives of the CSC project. The main objectives of the CSC was to increase the power flow transfer capability of the Central-North transmission corridor, to improve voltage control capability at Marcy, to enhance system damping during oscillatory contingencies, and provide operational flexibility during system outages. To this end, the CSC was specified as a single, multi-function all solid-state, modular compensating system with expandable hardware structure, capable of providing any one, or any combination of the following compensation and control functions:

• Voltage support and regulation by reactive shunt compensation,

• Power flow control by reactive series compensation,

• Independent active and reactive power flow control,

• Active and reactive line load balancing,

• Stability improvement.

8.5.2.2 Description of the CSC/IPFC

The Convertible Static Compensator for the Marcy Substation was designed to meet the above defined system requirements. It is structurally similar to the UPFC at Inez with extended functional and control capability, in particularly, with the important addition of the IPFC operating mode that allows power flow control in two different lines.

Power Circuit Structure. The 345 kV 200MVA CSC comprises two identical Gate Turn-Off (GTO) thyristor-based converters, each rated for ±100 MVA. Each converter includes multiple high-power, 3-level, 48-pulse GTO valve structures feeding an intermediate (low voltage) transformer rated for approximately 50 MVA. The converter and magnetic configuration provides harmonic neutralization and thus the output is a three-phase voltage set of nearly-sinusoidal quality that is coupled to the transmission line via the intermediate transformer by main coupling transformer. The main coupling transformer for shunt-connection is rated for 200 MVA and has two sets of secondary windings to accommodate both converters in parallel for reactive shunt compensation. Two coupling transformers, each rated for 100 MVA, are used to connect their corresponding converter to the selected line, to provide the configuration required for UPFC or IPFC operation. Thus, the converters can be connected to function as a STATCOM, regulating the voltage at the Marcy bus. The converters can also be connected through the insertion transformers in series with Marcy-New Scotland and Marcy-Coopers Corners 345 kV lines. The series transformers allow the converters to operate as a UPFC from the Marcy bus with series power flow control in either of the two lines. Also, they can be operated as an IPFC with active power transfer capability between the two lines. And, of course, the converters can be disconnected at their dc terminal and operated independently as reactive shunt and/or series compensators (i.e., STATCOM and/or SSSC). The converter structure is shown schematically with relevant information in Fig. 8.39. The installed converter is shown in Fig. 8.40. The one line diagram in Fig. 8.41 shows the connection arrangement of the CSC in the NYPA system at the Marcy substation and Fig. 8.42 shows the actual installation.

8-73

Figure 8-39 Basic structure of the CSC converters and the corresponding technical data

Figure 8-40 The CSC converter hall

Each Converter

Nominal rating: ±100 MVA Transmission voltage: 345 kV Converter output voltage: 22 kV (phase (to coupling transformer) to phase)Pole output voltage: 5.5kV No. of converter pulses: 48 Type of converter pole: 3-level No. of poles: 12 No. of 3-phase converters: 4 Type of semiconductor: GTO Rating of GTO: 4.5 kV, 4kA No. of GTOs per valve: 6 (5+1) DC voltage: 6+6 kV DC capacitor rating: 72 kJ Note: All ratings are nominal

0 2 4 6 8 1 1 1 1

1

-

0

P.U

msec

Converter voltage

Main transformer

345kV Line

Intermediatetransformer

48-pulse, 3-level converterVoltagelimiter

+

+

+

+

8-74

Figure 8-41 Simplified UPFC power circuit in the NYPA system at the Marcy Substation

Converter 1

New Scotland

Shunt xfmer

Intermediate transformers

DC Capacitors

Series xfmer

Converter 2

Thyristor

Series xfmer

Thyristor

Coopers Corners

Marcy South bus

Marcy North bus

8-75

Figure 8-42 NYPA CSC installation at the Marcy Substation

Control System. Both converters comprising the CSC are controlled from a single central control system housed in the control room. The control system is similar to that described for the Inez UPFC, but facilitates more system configurations, operating modes and control functions. The CSC basic Internal control architecture is shown in Fig. 8.43 and the operational and functional options are illustrated in Fig. 8.44. The Internal control of the CSC is interfaced with NYPA’s specially designed Master Controller that provides the appropriate references, derived from measured system variables and operation requirements, for the CSC according to the overall system optimization control concept illustrated in Fig. 8.45.

8-76

Figure 8-43 Basic functional control structure of the NYPA CSC

Figure 8-44 Functional and operational capabilities of the NYPA CSC

LocOperatInterfa

RemotContrInterfa

CentralControls

- Real time control algorithms - Subsystem data acquisition - System configuration - Sequencing - Protection

Converter 1

Pole (12)

(1) DC Gat

Drive

GatinPowSuppl

CoolinSyste

Analog Inputs(System Voltages

and Currents)

Digital I/O(Relay Logic)

Converter 2

Pole (12)

(1) DC Gat

Drive

GatinPowSuppl

CoolinSyste

Convertible Static Compensator

Reference signals for selected operation modes

Operation and Control Mode Selector

- +

Var controlVoltage regulation

STATCOM

P flow

Line impedanceQuadrature voltage

SSSC

345 kV bus

Converter 1(100 MVA)

+ -

Line impedanceQuadrature voltage

Independent P flowwith selected Q flow

IPFC (each line)

Line impedanceP, Q flow

Voltage injectionPhase angle

UPFC

Converter 2(100 MVA)

Line 2

Line 1

Reference signals for selected operation modes

8-77

Figure 8-45 The CSC system control structure

8.5.2.3 Importance of the NYPA installation

This Convertible Static Compensator with the full implementation of the IPFC concept extended the functional capability of converter-based modular transmission Controllers to the series compensation of multi-line systems with active power transfer option for the control of reactive power flow. With this, the practicality of back-to-back converters with both shunt-series and series-series ac coupling and operation has been verified. The practicality of converting a two-converter based compensator into five different operating modes to provide transmission compensation control functions according to prevailing system requirements has been demonstrated. The NYPA installation also proofed the operation of series-coupled converters in the case of high voltage (345 kV) lines.

References [1] Gyugyi, L., “A Unified Power Flow Control Concept for Flexible AC Transmission Systems,” IEE Fifth International Conference on AC and DC Power Transmission, London, Publication No. 345, pp. 19-26. Reprinted in IEE PROCEEDINGS-C, Vol. 139, No. 4, July 1992.

[2] Schauder, C.D. and Mehta, H., “Vector Analysis and Control of Advanced Static Var Compensators,” IEE PROCEEDINGS-C, Vol. 140, No. 4, July 1993.

[3] Gyugyi, L., “Dynamic Compensation of AC Transmission Lines by Solid-State Synchronous Voltage Sources,” IEEE Trans. on Power Delivery, Vol. 9, No. 2, April 1994.

[4] Lerch, E., et al., “Simulation and Performance Analysis of Unified Power Flow Controller”, CIGRE Paper No. 14-205, 1994.

CSC with Internal

Control

Damping Controller

CSC Master

Controller

CSC status and operation information

Capacitor and reactor bank on/off

commands

References and operation setting System variables and

operation data (SCADA & other

Power System

Control mode & gain control

Σ

8-78

[5] Gyugyi, L. et al., “The Unified Power Flow Controller: A New Approach to Power Transmission Control,” IEEE Trans. on Power Delivery, Vol. 10, No. 2, April 1995.

[6] Maliszewski, R.M. and Rahman, M., “Planning and Operating Considerations for the World’s First UPFC Installation”, CIGRE Report No. 37-96 (US) 13 (E), Paris, 1996.

[7] Mehraban, A.S., et al., “Application of the World’s First UPFC on the AEP System”, EPRI Conference - The Future of Power Delivery, Washington DC, April 9-11, 1996.

[8] Mihalic, R., et al., “Improvement on Transient Stability Using Unified Power Flow Controller,” IEEE Trans. on Power Systems, Vol. 11, No. 1, January 1997.

[9] Rahman, M., et al., “UPFC Application on the AEP System: Planning Considerations,” IEEE Trans. on Power Systems, Vol. 12, No. 4, November 1997.

[10] Papie, I., et al., “Basic Control of United Power Flow Controller”, IEEE Trans. on Power Systems, Vol. 12, No. 4, Nov 1997.

[11] Schauder, C.D., et al., “Operation of the Unified Power Flow Controller (UPFC) Under Practical Constrains,” IEEE Trans. on Power Delivery, Vol. 13, No. 2, April 1998.

[12] Sen, K.K., and Stacey, E.J., “UPFC - Unified Power Flow Controller: Theory, Modeling, and Applications,” IEEE Trans. on Power Delivery, Vol. 13, No. 4, October 1998.

[13] Edris, A., et al., “Controlling the Flow of Real and Reactive Power,” IEEE Computer Applications in Power, Vol. 11, No. 1, January 1998.

[14] Renz, B.A., et al., ”World’s First Unified Power Flow Controller on the AEP System,” CIGRE Paper No. 14-107, 1998.

[15] Renz, B.A., et al., “AEP Unified Power Flow Controller Performance” IEEE/PES Winter Meeting, Paper No. PE-042-PWRD-0-12-1998.

[16] Gyugyi, L., et al., “The Interline Power Flow Controller Concept: A New Approach to Power Flow Management in Transmission Systems,” IEEE/PES Summer Meeting, Paper No. PE-316-PWRD-0-07-1998, San Diego, July 1998.

[17] Fardanesh, B., et al., “Convertible Static Compensator Application to the New York Transmission System”, CIGRE Paper 14-103, 1998.

[18] Hingorani, N., G., and Gyugyi, L., “Understanding FACTS – Concepts and Technology of Flexible AC Transmission Systems”, IEEE Press, 2000.

[19] Fardanesh, B., et al., “Mercy Convertible Static Compensator Validation of Controls and Steady State Characteristics”, CIGRE Paper, 2002.

9-1

9 CONVERTIBLE AND EXPANDABLE STATIC COMPENSATORS (OUTLINE)

9.1 Need for Convertibility in Evolving Utility Environment

9.2 Concept of Convertible Static Compensator (CSC)

9.2.1 Converter-Based Functional Building Block Structure

9.2.2 Structure for Selectable Application Function and Operating Mode Control

9.3 Application Example: NYPA CSC Installation at Marcy Substation

9.3.1 Application Requirements and Expected Benefits

9.3.2 Power Converter Topology and Structure

9.3.3 Functional Capabilities

9.3.4 Control System

9.3.5 Installation

9.3.6 Operating experience

9.4 Functional and Rating Expandability via Converter-Based Building Block Structure

9.4.1 Functional Expansion from Single-Line to Multi-Line Transmission Control

9.4.2 Rating Expansion by Modular Approach: Issues and Trade-offs


10-1

10 VOLTAGE-SOURCED CONVERTER-BASED BACK-TO-BACK (BTB) SYSTEM INTERTIES (OUTLINE)

10.1 Application Motivation (As Contrasted to Line-Commutated Approach)

10.1.1 Capability to Provide Capacitive and Inductive Vars for Terminal Voltage Regulation

10.1.2 Capability to Supply Weak Systems and Passw bive Loads

10.1.3 Capability to Operate it Both as an Asynchronous Link or a Synchronous Tie with Controllable Synchronizing Torque

10.2 Operating Principles and Characteristics

10.2.1 Conceptual Representation

10.2.2 Equivalent Circuit

10.2.3 Practical Implementation

10.3 Functional Characteristics

10.3.1 Transmittable Power vs. Reactive Power Generation/Absorption

10.3.2 Transmittable Power vs. Frequency

10.3.3 Synchronizing Torque vs. Frequency Deviation

10.4 Control Principles and Structures

10.4.1 Functional Control Schemes to Meet Application Requirements and

Limitations

10.4.1 Active Power Transmission, Reactive Power and Voltage

Regulation

10.4.2 Torque Control

10.4.3 Power Oscillation Damping

10.4.4 Achievable Dynamic Performance

10-2

10.5 Converter Power Circuit Topology and Output Control Options

10.5.1 Requirements and Trade-offs

10.5.1.1 Voltage vs. Current

10.5.1.2 Structure vs. MVA Rating

10.5.1.3 Modularity and Redundancy vs. Availability and Cost

10.5.2 Harmonic Generation and Filtering

10.5.3 Losses

10.6 Rating and Protection Considerations

10.6.1 Surge Rating Requirements

10.6.2 System Faults

10.6.3 Internal Converter Faults

10.7 Performance During and Following Line Faults

10.8 Installation and Insulation Coordination Considerations

10.9 Installation and Performance Comparison to Conventional Line-

Commutated BtB Power System Ties

10.10 BtB Application Examples

10.10.1 Application Background and Requirements

10.10.2 Power Converter Topology and Structure

10.10.3 Functional Capabilities

10.10.4 Control System

10.10.5 Test Results

10.10.6 Operating Experience


11-1

11 VOLTAGE-SOURCED CONVERTER-BASED DC TRANSMISSION SYSTEMS (DRAFT)

11.0 Introduction

HVDC Transmission was first put into commercial service in 1954 and has since been used extensively for the interconnection of asynchronous ac networks and for the transmission of power over long distances. The technology used for the initial HVDC transmission schemes relied on the presence of an ac voltage in the network for its correct operation, and is known as line commutated converter (LCC) HVDC technology. In this chapter, HVDC schemes using this technology will be referred to as LCC HVDC. This technology is still used extensively for HVDC transmission. The total rating of LCC HVDC schemes installed by the end of 2005 was in excess of 60 GW, the largest scheme having a rating of 6300MW and operating at ±600 kVdc.

The use of voltage sourced converters for dc power transmission (VSC Transmission) was introduced with the commissioning in 1997 of the 3MW, ±10kVdc technology demonstrator at Hellsjön, Sweden. CIGRÉ has given this new type of dc transmission the name VSC Transmission, and this term will be used in this chapter [1]. Figure 11-1 shows a simplified diagram of a VSC Transmission scheme connecting two ac grids.

Figure 11-1 Simplified diagram of VSC Transmission scheme

The demonstrator proved the feasibility of the technology and subsequently, a number of commercial schemes using this technology have been installed. At the end of 2005 ABB was the only manufacturer supplying VSC Transmission schemes. The total rating of VSC Transmission schemes in service at the end of 2005 was 890MVA, a growth rate significantly exceeding that of the early LCC HVDC technology. The largest VSC Transmission scheme in service at the end of 2005 had a rating of 330 MW and operated at ±150 kVdc.

~=

+~ ~=+

-~UdBUdA

Idc

XConv AZLA

ULA

ZLBX Conv B

ULB

ac grid A VSC A DC transmission lineVSC B

ac grid B

Rdc

Sending End Receiving End

-

11-2

11.1 Differences in Features and Characteristics from Conventional HVDC

The following subsections provide a brief overview of the differences in features and characteristics between a VSC Transmission scheme and a LCC HVDC scheme. Further descriptions and details will be provided about VSC Transmission in subsequent sections.

11.1.1 Technology

VSC Transmission is, as its name implies, based on the use of Voltage Sourced Converters (VSC) [2][3]. An ideal VSC has a constant dc voltage at its dc terminals, and creates an alternating voltage at its ac terminals by the switching of the converter. The Line Commutated Converter as used in LCC HVDC is normally a Current Sourced Converter (CSC). An ideal CSC has a constant direct current at its dc terminals and creates an alternating current at its ac terminals by the switching of the converter. In practice the supposedly constant voltage or current will in both cases have some variation during operation, as the voltage/current source will have finite impedance. The variation in voltage/current has to be taken into account in the determination of the actual wave shapes.

VSCs have been used for many years in motor drives, and the development of powerful semi-conductors, with turn-off as well as turn-on capability, and with high speed switching capability, has enabled modern motor drives to become compact and versatile. The semi-conductor used most frequently today in Voltage Sourced Converter applications is the Insulated Gate Bipolar Transistor (IGBT), but several other promising devices are also available or under development [4]. The maximum device rating commercially available at the end of 2005 is 6500V and a current capability of up to 2kA. The IGBT is a controlled device, which means that the impedance between its collector and emitter is determined by the control signal applied to its gate. This feature is used to enable the device to turn off at times other than those corresponding to a naturally occurring current zero, in particular, a VSC can deliver power to a passive network. The IGBT is capable of blocking voltage only in the forward direction. Therefore, a diode is used in parallel to the IGBT. This diode prevents the dc voltage from reversing polarity, and enables current to flow in the opposite direction. When changing the direction of power flow between the two terminals in a VSC Transmission scheme the direction of the direct current flow is reversed. The direction of power flow can be smoothly changed through zero, and there is no constraint on the duration of operation at any power within the rating of the converters.

The power semi-conductor used in today’s LCC HVDC schemes is the thyristor, which can be triggered into conduction by a low power gate signal when the device is forward biased, and which will turn off when the current through the device attempts to reverse. Since the thyristor cannot be turned off on command, a LCC HVDC scheme requires a rotating EMF in the network for satisfactory operation. The thyristor needs a short time interval (a fraction of a millisecond) from the instant it has turned off until it can withstand forward voltage again. Devices capable of being turned on by a light signal are now available, in addition to devices requiring an electrical signal. The power rating of thyristors available for LCC HVDC schemes is very large, with individual devices having a capability of 8.5kV withstand voltage and 4kA current capability. Devices with voltage capability up to 12kV have been developed, but the current capability is then significantly reduced. Since the thyristor can withstand voltage in both directions, the dc voltage can be reversed. When changing the direction of power flow between the two terminals in a LCC HVDC scheme the polarity of the direct voltage is changed, since the thyristors can conduct current only in one direction. An LCC HVDC scheme is not normally able to operate

11-3

continuously at a direct current below about 5% of the rated current, because the harmonics results in discontinuous current operation.

As stated above, the semi-conductors in the VSC can be turned on and off irrespective of the state of the current flow through the converter. Therefore, it is possible to control the converter such that both the active and the reactive power flow at the converter ac terminals is determined in accordance with operational requirements. This means that the VSC can be operated in all 4 quadrants of the P and Q diagram. In other words, the VSC can be operated as a rectifier or an inverter, and as a generator or absorber of reactive power, and in any combination of active and reactive power, within the rating of the converter. Naturally, the active power flow at the terminals of a VSC Transmission scheme must be coordinated, such that the active power is balanced.

The converters in a LCC HVDC scheme operate with a lagging power factor, typically absorbing reactive power equivalent to approximately 50% of the active power. Thus, the converters in a LCC HVDC scheme can only be operated in 2 quadrants of the PQ diagram. AC harmonic filters and shunt capacitors are used to offset the converter’s reactive power absorption, such that the overall power factor is acceptable to the ac network. The ac harmonic filters and shunt capacitors are normally switched in and out by means of circuit breakers as the dc load varies. It is possible to use a capacitor in series with the ac terminals of the converter, and this can significantly reduce both the magnitude and variability of the reactive power absorbed by an LCC HVDC converter. However, unless the value of the series capacitor can be varied, the converter’s reactive power absorption cannot be changed without also changing the active power, unless ac harmonic filters or shunt capacitors are switched.

Different topologies have been used or proposed for the converters in a VSC Transmission scheme. The topology for the first schemes was a 2-level converter, in which the ac terminals are switched between the two dc terminals. Some later schemes use a 3-level converter, in which the ac terminals are also switched to the potential at the midpoint of the dc capacitor. The 3-level topology reduces the voltage across each valve, and also reduces the base harmonic content in the output voltage.

All LCC HVDC schemes are based on the use of a 3-phase, 2-level converter, which is known as a 6-pulse converter. Modern LCC HVDC schemes use basic building blocks of two such 6-pulse converters, combined in series on the dc side and in parallel on the ac side. Converter transformers providing a 30 degree phase shift between the two basic 6-pulse groups can be used to provide harmonic cancellation, leaving only main ac side harmonics of the order 12n±1, where n is an integer.

The turn-off and high switching speed capability of the VSC valves in a VSC makes it possible to turn each valve on and off several times in each power frequency cycle. This feature can be used to eliminate or significantly reduce the low order harmonic content of the voltage on the ac and the dc side of the VSC. A number of different control methods can be used for the reduction of lower frequency harmonics, leaving primarily the higher frequency harmonics, which can be reduced to acceptable levels more easily by passive filters than lower frequency harmonics.

Due to the controllability of reactive power a VSC Transmission scheme does not require switchable ac harmonic filters. This means that the site area of a VSC Transmission terminal is dictated largely by the size of the converter, which tends to be of similar size to that of an LCC HVDC converter. Accordingly, the overall site area of a VSC Transmission terminal is

11-4

significantly smaller than the site area for a LCC HVDC terminal, typically occupying 40% or less of the area of a LCC HVDC terminal.

11.1.2 Operating Features

In order for a LCC HVDC scheme to operate satisfactorily the short circuit level of the ac system at the point of connection needs to have a minimum level. The actual minimum level depends on the characteristics of the converter control and on the design of the converter itself. Effectively, the issue is one of ac network voltage stability during system dynamics. For a conventional LCC HVDC transmission scheme to operate satisfactorily the minimum short circuit level typically needs to be about 2.5 times the rating of the HVDC scheme. A low commutating impedance for the LCC HVDC converter, e.g. by using low transformer impedance and/or a series capacitor, enables satisfactory operation at a lower short circuit level at the point of connection. Similarly, arranging the converter control characteristics such that a drop in ac voltage results in a reduction in reactive power absorption, enable the converter to operate satisfactorily at lower short circuit level at one of the terminals.

A VSC Transmission scheme can provide reactive power support at both terminals independently of the active power transmitted between the terminals. Therefore, it can stabilize the ac network voltage and a minimum short circuit level at the point of connection is not required. As a consequence the terminals of the VSC Transmission scheme can be located at weak point in the ac network, e.g. remote from the core of the network. Naturally, system studies must be performed to determine the acceptability of the resulting load flows and voltage profiles in the ac networks.

As the power flow on a LCC HVDC scheme varies from minimum power to maximum power and back again, ac harmonic filters and shunt capacitor banks are switched in and out. The switching control should at all times and power flows keep the harmonic distortion at an acceptable level, and the steady state ac voltage at the point of connection at an appropriate level for the power flow in the ac network. The size of ac harmonic filters and shunt capacitors must be selected such that the voltage steps caused by filter switching do not cause unacceptable flicker in the ac network. If the power flow on the HVDC scheme changes very frequently, then the number of operations of the circuit breakers can be relatively large, resulting in the need for frequent inspection and maintenance of the circuit breakers.

A VSC Transmission scheme typically uses a fixed ac harmonic filter, and the converter is rated to provide the varying reactive power requirements for the envisaged operating conditions. In this case circuit breaker operation is required only at start-up, shut-down, and fault clearance, and the inspection and maintenance needs will typically be less than for a normal ac substation circuit breaker.

The breaker switched ac harmonic filters and shunt capacitor banks required for steady state ac voltage control in an LCC HVDC scheme can result in large temporary over-voltages if power transmission suddenly stops, e.g. during dc line faults or converter faults. For some events the converters can continue to circulate direct current to absorb reactive power, thereby reducing the amplitude of the over-voltage, but for other events the current must be stopped immediately, e.g. to prevent equipment damage or remove dangerous conditions. The over-voltage could also be reduced by disconnection of the filters and capacitor banks, but this would delay resumption of power transmission until the capacitors had been discharged, typically several minutes.

11-5

Since a VSC Transmission scheme uses only a relatively small ac harmonic filter, and the healthy converter can continue to control reactive power without circulation of direct current. Faults within the converter would normally result in tripping of the main circuit breaker, thereby removing not only the converter but also the ac harmonic filter. Therefore, the ac over-voltage caused in the event of a sudden stop in power transmission is typically very small.

In the event of a sudden drop or phase shift in the ac voltage amplitude, the LCC converter may not be able to complete the commutation of the inverter before the voltage across the outgoing valve becomes forward biased, and therefore the valve will continue to conduct. As the valves continue to be fired, a short circuit between the dc terminals of the 6-pulse bridge will be caused. This event is called a commutation failure, and results in a dip in the ac voltage and temporary dc over-current, both of which are quickly reduced by control at the rectifier. Typically normal commutation and power transmission can be restored within two power frequency cycles, and the event has no significant impact on the operation of the ac system.

Since the VSC valves can be turned off as well as being turned on, the operation of a VSC Transmission scheme is not significantly influenced by voltage dips or other transient ac disturbances. In particular, a VSC Transmission scheme does not suffer commutation failures.

In the event of a short circuit on the dc side of a LCC HVDC scheme, the converters can quickly stop the current flowing into the fault, merely by stopping the firing of the thyristor valves. The quick reaction of the converter control and protection will minimize any damage caused by the fault. If the fault is on a dc overhead line, then the power transmission is stopped for a sufficient period of time to allow de-ionization of the arc, typically 200-300ms, and transmission is then resumed. Several re-start attempts, with increasing de-ionization periods, can be used if desired.

The diodes used in a VSC would continue to conduct current into a fault on the dc side, even when the IGBTs are turned off. A direct current circuit breaker could be used to stop the current flow, but such breakers are expensive and not generally used. Therefore, the ac breakers at both ends of the VSC Transmission scheme must be opened in the event of a fault on the dc line. The scheme then has to be re-started once the fault has been removed. All commercial VSC Transmission schemes have so far used cables rather than overhead lines, which reduces the number of faults on the dc side.

11.1.3 Application Benefits and Disadvantages

The technical and operational differences described above in 11.1.1 and 11.1.2 results in a VSC Transmission scheme having a number of application benefits compared with a LCC HVDC scheme. These application benefits are briefly described as follows:

• The ability of the VSC to control reactive power independently of the active power flow between terminals means that it can support ac voltage at its connection point. This can provide significant benefits to the ac network, e.g. improving voltage stability and increasing load-ability of the network.[5] [6].

• The rapid speed of control of active and reactive power resulting in part from PWM operation can be used to improve the power quality at the point of connection, e.g. the reduction of flicker from industrial plant. [7]

• The self commutation and reactive power control means that a VSC Transmission scheme can be connected at very weak points in the ac network, and can be the sole supply of

11-6

power to a network which does not have synchronous generation or compensation. This capability can be useful also in a conventional ac network, where a VSC Transmission terminal can provide “black start” (passive network start) capability for the restoration of the network after a major blackout.

• The land area occupied by the VSC Transmission terminals are considerably smaller than that occupied by the LCC HVDC converter stations of equivalent rating.

• The constant polarity of the direct voltage means that multi-terminal VSC Transmission schemes can be designed with greater operational flexibility than when using LCC HVDC, where the polarity of a converter would have to be changed if its role were changed from a rectifier to an inverter, or vice versa.

• The constant polarity of the direct voltage means that dc cables do not have to be capable of withstanding polarity reversal. This enables a variety of polymeric cables to be used. Polymeric cables are smaller and lighter than oil impregnated cables, particularly when heavy mechanical protection or armoring is not necessary [8].

These advantages will be further elaborated in section 11.2.

The main application disadvantage of a VSC Transmission scheme compared with a LCC HVDC scheme is the higher power loss in the converter (see section 11.7.3). However, the VSC Transmission scheme’s reactive power control capability may reduce the reactive power flow in the ac network, thereby reducing the overall power loss.

11.2 Application Areas

HVDC transmission has been used since its first introduction in 1954 to provide interconnections between ac networks and the transmission of bulk energy over long distances. The benefits of interconnections include:

• Better utilization of the installed generation in the two networks by taking advantage of the diversity in loads and the characteristics of generation.

• Reduction in overall spinning reserve.

• Emergency power exchanges between the networks.

The benefits arising from the use of HVDC as an interconnector include:

• The power flow on the interconnector is fully controlled. This feature means that loop power flows can be avoided.

• It is possible to interconnect asynchronous networks.

• When the interconnector electrical distance is very long, e.g. >800km overland or >70km submarine, HVDC transmission may provide a more economic solution, both in terms of capital cost and power loss.

VSC Transmission can in principle be used in all the applications for which LCC HVDC are currently used. However, unless the VSC Transmission technology develops further to enable the use of dc overhead lines and higher transmission voltage, the long distance overland bulk power transmission market, e.g. 1000MW+ and 800km+, will continue to be provided only by LCC

11-7

HVDC schemes. Some application areas, in which the technical characteristics of VSC Transmission may be of particular benefit, are described further in the following subsections.

11.2.1 Interconnection of Geographically or Otherwise Isolated Networks

Communities in sparsely populated areas, such as the north of Canada or Alaska, or on small islands, may rely on diesel generation for their electricity supply. The cost of electrical energy in such locations may be very high, because advantage cannot be taken of the benefits of large-scale generation available in major networks. Additionally, generation of the energy using small diesel generators is more damaging to the environment, than using large-scale generation in a major network, with its more efficient generators and operating modes. Therefore, transmission of energy from the large network may be economically and environmentally attractive.

VSC Transmission may offer a good solution to such power transmission if ac overhead line or submarine cable transmission is not feasible, e.g. if the distance is large, or if an underground cable solution is considered advantageous. An underground cable solution rather than an overhead line, may be advantageous for several reasons:

• It is less intrusive on the landscape than an overhead line, particularly if the cable can be installed along the route of an existing road, or the land above the cable can be returned to its original use after installation of the cables (e.g. farming).

• The lower visibility of a cable installation is likely to be more acceptable to the public, thus resulting in planning approval being granted more quickly.

• It is not subject to flashover due to pollution problems, e.g. due to salt fog near to the coast, or to fertilizer application in farming areas.

• It is likely to be more reliable than an overhead line, because it has fewer exposed parts.

• It does not produce electric fields or varying magnetic fields, and the magnetic field produced by the direct current typically results in a steady field with a magnitude significantly less than the Earth’s magnetic field.

Installation of the polymeric cables can be achieved very quickly if the ground conditions are such that the cables can be ploughed directly into the ground using special machinery. The installation of the two 70km long cables for the Gotland project was achieved in 90days. Figure 11-2 shows the cable laying operation for the Murray Link VSC Transmission scheme in Australia.

11-8

Figure 11-2 Cable Laying for the Murray Link VSC Transmission scheme

VSC Transmission is advantageous compared to a LCC HVDC solution for the supply of power to an isolated/passive network, because synchronous compensators are not necessary for the operation of the VSC. This means that significantly less maintenance will be required, not only because the VSC solution has fewer components than a LCC HVDC scheme, but also because of the omission of a rotating machine. Additionally, whilst the power loss of a VSC Transmission scheme can be 2 or 3 times higher than that of a LCC HVDC scheme, the difference would be reduced in this application by the relatively large power loss associated with the operation of a synchronous compensator.

It should be noted that the frequency variation in the network may be significantly larger than when the network is fed from its own diesel generation. This is because a fault causing a large reduction in the ac voltage in the main network will reduce the interconnector’s transmission capability. The VSC scheme can temporarily reduce the ac voltage to reduce the power flow, but if large synchronous or asynchronous machines are used in the network, these may slow down very rapidly, as they start to act as generators. The re-acceleration of the machines may impose a large duty on the VSC, and may need to be taken into account in the rating of the scheme. Appropriate control and protection strategies need to be determined taking into account the duration of faults and the characteristics of the loads in the network.

11.2.2 Interconnections Between Weak Power Systems

For weak ac network applications the control of the ac network voltage is particularly important. When a LCC HVDC scheme is used, the large ac harmonic filters and shunt capacitors used to provide reactive power compensation actually reduce the ac voltage stability of the system, since the reactive power support reduces as the ac voltage reduces. Therefore, the application of LCC HVDC schemes is typically limited to systems where the short circuit power at the point of connection of the LCC HVDC terminal is at least 2.5 times the rating of the scheme. When the

11-9

short circuit level is lower than this ratio special control methods can be used, but these typically require a higher rating of the converter station equipment. Alternatively, additional reactive power support, in the form of SVCs or STATCOMs, can be applied at the point of connection, to improve the ac voltage stability.

As a VSC Transmission scheme is capable of controlling the reactive power at its ac terminals independently of the active power flow between terminals (subject to overall rating limitations), it provides a good solution for the interconnection of networks, where the ac network at one or both terminals is weak.

11.2.3 Reinforcement of Weak AC Tie-Lines for Stability Improvement

When an ac transmission line is imbedded in a geographically large ac network with major load and generation centers at its extremities, a fault within either of the centers can cause major power oscillations on the ac tie-line. If the tie line is weak (e.g. because of its length or because of its rating relative to the network), then the power oscillations may exceed the over-current setting of the line protection, resulting in a trip of the line. The power oscillations may also have a significant impact on the ac voltage along the line and at the termination points.

The power oscillations can be damped by the insertion of controlled series compensation, or, less effectively, by the use of controlled shunt reactive power compensation. A more effective solution would be the installation of a parallel VSC Transmission scheme.

Studies have shown that a parallel VSC Transmission scheme can increase the stability limit on a weak ac tie line by more than the rating of the VSC Transmission scheme. This is achieved by suitable modulation of the active power transfer of the VSC Transmission scheme AND the reactive power at its terminals. [5]

11.2.4 Connection of Distant Loads (Off-Shore Oil and Gas Platforms)

As an oil or gas reservoir is emptied, the power required to extract and transport the oil or gas increases. Where the production platform is located off shore, and at significant distance from the coast, diesel or gas turbine generators located on the production platform have typically produced the electrical power required for the operation. However, as the power requirement increases, additional generating plant becomes necessary and consideration may be given to obtaining the electrical power from the on-shore ac network. The advantages of obtaining the power from the shore include:

• The maintenance intensive generation plant on the platform can be removed.

• CO2 emissions are reduced since the efficiency of on-shore generation plant is more efficient, and may include renewable generation plant. This may result in substantial tax benefits.

When the distance to the shore is large, transmission by means of ac may not be feasible, and HVDC transmission may be considered. The use of VSC Transmission [9] provides the following advantages:

• The VSC Transmission equipment is compact, making the installation on existing platforms feasible.

11-10

• There is no need for synchronous compensators since the VSC Transmission scheme is self-commutating, and controlled reactive power can be provided, independently of the active power being transmitted.

• The VSC Transmission scheme can provide variable frequency power supply to the offshore motors, acting as a variable speed motor drive thereby improving the performance of the offshore equipment.

11.2.5 Connection of Remote Wind-Parks

Wind generation is currently one of the most economical sources of renewable generation, and is the fastest growing sector of electrical power generation. As more and more wind generators are installed there is a tendency for the wind farms to be located further away from the population centers, partly because of better wind conditions and partly to reduce the visibility and environmental impact of the wind turbines. Wind conditions are often best in offshore locations and several offshore wind farms with ratings of 500MW and more are currently under consideration.

When the wind farm can be connected relatively easily to a strong point in the ac network, then an ac interconnection is often the most appropriate solution. However, when the transmission distance to the main network connection point becomes large electrically, or the ac connection point is in a weak part of the ac network, then the most economic solution may be to use VSC Transmission [10][11].

Whilst a HVDC terminal occupies more space and costs significantly more than an equivalently rated ac substation, HVDC Transmission provides the following advantages:

• The transmission distance can be much larger, enabling connection at a more suitable point in the ac network.

• The HVDC scheme will provide de-coupling between the main ac network and the wind farm network, thereby improving the ride through capability of the wind farm during faults in the mainland ac network.

• The frequency of the offshore network can be allowed to vary with the speed of the wind, increasing the efficiency of the wind farm.

In addition to these general advantages, a VSC Transmission scheme provides the following advantages compared with a LCC HVDC scheme

• The VSC Transmission scheme can provide reactive power support of the wind farm ac network, improving the power quality and stability of the wind farm ac network.

• Flicker caused by operation of the wind farm will not be transmitted to the mainland ac network.

• Power can be transmitted to the wind farm ac network during conditions of little or no wind. This is useful, since power is required at all times to satisfy the auxiliary power requirements for both the VSC Transmission scheme and for the wind generators, e.g. control and protection, safety systems, navigation lighting, cubicle heating, etc.

The optimum overall design of the wind farm and interconnector would be achieved when taking into account the technical capability of the wind generators and the VSC Transmission scheme. For example, doubly fed induction generators enable variable turbine speed and are capable of

11-11

providing some element of reactive power compensation, and to some extent these characteristics are a duplication of those of a VSC Transmission scheme. Cost savings and satisfactory technical performance might be achieved by combining a VSC Transmission scheme with induction generators, or by using lower ratings of the converters in the doubly fed generators. Naturally, system studies would be required to determine the overall dynamic and transient performance, before finalizing the design and rating of components.

11.3 Basic Operating Principles and Characteristics

11.3.1 Conceptual Representation

Conceptually a VSC Transmission scheme is similar to a zero inertia synchronous machine/generator A coupled through an infinitely variable gear/coupling to a zero inertia synchronous generator/machine B. The machine and the generator can transfer active power from the network at A to the network at B or vice versa. In steady state operation the output power at the generator ac terminals will be equal to the input power at the machine ac terminals, less the power loss in the machine and generator and the transmission line. The machine and the generator are each able to provide controlled reactive power independently of the power transmitted between the two terminals.

In an actual implementation, the dc capacitors and the cable capacitance provide a small amount of energy storage, typically equivalent to the rated energy for one power frequency cycles or less. This energy provides some smoothing of the power flow, and helps the control and stability of the VSC Transmission scheme. However, the energy storage in the dc capacitor is only a small fraction of the kinetic energy in a typical generator.

11.3.2 Equivalent Circuit of a VSC Transmission scheme

Figure 11-3 shows a simplified diagram of a VSC Transmission scheme connecting two ac networks A and B. The essential elements of the VSC Transmission scheme are:

• The converters VSC A and VSC B, which perform the conversion between the dc voltage and the ac voltage.

• The dc capacitors, which provide the voltage source, from which the ac voltage is created by switching of the converter.

• The transmission link, which enables energy to be exchanged between the dc capacitors at VSC A and VSC B.

• The inductive impedance between the converter and the ac network, which can be found at both VSC A and VSC B, and which is essential to stable operation of the system.

11-12

Figure 11-3 Simplified diagram of VSC Transmission scheme

Most simulation programs (e.g. PSCAD/EMTDC, PSS/E, DigSilent) provide models of VSC Transmission schemes, and can be adapted to provide the relevant control characteristics. Such models are fine for preliminary studies. However, generally speaking, care should be taken when using such models, as they may not include all features of a real control and protection system, and therefore the response during transients and system dynamics may not reflect the actual performance of a real VSC Transmission scheme.

11.3.3 Steady State Operating Characteristics of a VSC Transmission scheme

As is the case for a synchronous machine/generator the current flowing at the ac terminals of the VSC depends on the voltage amplitude and phase angle relative to the network EMF, as well as on the impedance between the VSC and the network EMF.

Figure 11-4 illustrates the control of active power through the converter line inductance by the variation of the phase angle. When the phase angle of the converter ac voltage Uconv leads the ac system system voltage UL the VSC injects active power in the ac system. Conversely, when the converter ac voltage lags behind the ac system voltage the VSC absorbs active power from the ac system.

Figure 11-4 Active Power Control

~=

+~ ~=+

-~UdBUdA

Idc

XConv AZLA

ULA

ZLBX Conv B

ULB

ac grid A VSC A DC transmission lineVSC B

ac grid B

Rdc

Sending End Receiving End

-

=Iconv

~

Pconv

Uconv Xconv UL+-Us

Id

UdR

DCresistor

DCcapacitor

d ~

U = U conv L

P = 0 P < 0

UL

Iconv

LU

ϕ ϕ

Uconv

Iconv

conv conv

Uconv-

Rectifieroperation

Inverteroperation

P > 0conv

11-13

Similarly, Figure 11-5 illustrates the control of reactive power. When the amplitude of the converter ac voltage Uconv is larger than the ac system voltage UL the VSC injects reactive power in the ac system, i.e. it acts as a shunt capacitor. Conversely, when the converter ac voltage is lower than the ac system voltage the VSC absorbs reactive power from the ac system, i.e. it acts as a shunt inductor.

Figure 11-5 Reactive Power control

As the amplitude and phase angle of the VSC ac voltages can be controlled independently of each other, subject to possible rating limitation in the converter and the balance of active power of the converters at the two ends, the active and reactive power output of the converter can be controlled to provide active power transmission as well as independent reactive power control at the terminals.

The voltage and current capability of the converter components determine the operating range of the converters. Different components determine the limit at different operating points, and therefore, the operating range can be tailored to meet specific requirements by enhancing some of the components. However, enhancement of some components could make the VSC deviate from the manufacturers standard module. More details of the PQ capability of the VSC Transmission scheme will be given in section 11.5.

11.3.4 Dynamic Capabilities and Application Potential

The speed of response of a VSC Transmission scheme can be very fast, with the possibility of reversal of power from 100% import to 100% export, and/or a reactive power change from generation to absorption within 1 power frequency cycle. Such fast changes may however not be acceptable to the ac network, and it is then necessary to slow down the speed of response to ensure that the overall performance is acceptable.

The overload capability of most VSCs is relatively limited, and thus the fast speed of response of the controls is important in order to keep the VSC within its acceptable duty during disturbances and transients in the ac network.

=Iconv

~

Pconv

Uconv Xconv UL+-

Us

Id

UdR

DCresistor

DCcapacitor

d ~

U = U conv L

Q = 0

FloatOperation

conv Q < 0

ULUconv

conv

I

Inductiveoperation

Q > 0

U conv

LU

convconv I

Capacitiveoperation

conv

11-14

The ability of the VSC Transmission scheme to modulate both the active and the reactive power means that it can provide superior assistance to an ac network, compared with a LCC HVDC scheme or a shunt reactive power device, such as a SVC or STATCOM.

The VSC Transmission controls typically include a number of control loops aimed at enhancing the performance of the ac network. When the system disturbance, to which the controller is designed to respond, can be measured at the location of the VSC Transmission terminal, it may not be necessary to use telecommunication in order for the converter to provide correct performance. However, when the VSC Transmission scheme can only provide an indirect attenuation of the disturbance, the use of telecommunication may enable a better response of the VSC Transmission terminal to the disturbance.

A. First Swing Stability

Potential first swing stability problems usually involve one or more generators, which are connected to a remote ac network. The instability is caused when a remote generator is isolated e.g. due to the tripping of the last remaining ac line to the generator. The remote generator will accelerate, as it cannot evacuate the power still being supplied by the prime mover, and re-synchronization may not be possible when the line is re-connected. The loss of the large generator may also cause problems in the remaining ac network, particularly if the remaining local generation is much smaller than the local loads, and the shortfall has to be supplied from the main network via a long ac line. In this case, power swings between the local area and the main network may be so large that the ac line trips due to over-current, leaving the local area isolated, with a resulting rapid drop in frequency.

The high speed at which the active and reactive power can be controlled by a VSC Transmission scheme may help avoid a first swing stability problem. However, if the ac system voltage at the VSC Transmission terminal has been reduced to near zero during the short circuit to ground, the scheme cannot influence the generator or the ac network until the fault has been cleared.

It is necessary to perform detailed studies to determine the control transfer functions that will best suit the situation. The VSC Transmission scheme may be able to assist the ac network by one or more of the following actions:

• Rapid reversal of the direction of power flow.

• Rapid restoration of the ac voltage after fault clearance.

• Providing an asynchronous connection between the affected part of the network and the main network.

One way to provide maximum benefit to the overall network may be to temporarily sever the synchronous links between the parts of the network, leaving only the VSC Transmission link between the two parts. This would allow the frequency in each to swing freely until the networks settle down and the generator governors and VSC Transmission can return the frequencies to the nominal values and the systems can be re-synchronized. Some load shedding may be necessary to achieve power balance, but a total system collapse will be avoided, and the duration of load interruption will be minimized.

11-15

B. Rotational Oscillation Damping

A sudden change in a power system, e.g. a short circuit to ground, can result in a temporary unbalance between the electrical power output of generators and their mechanical power input from the prime mover. This is because the time constant associated with a change in mechanical power input is considerably longer than the electrical power output time constant. The result is an acceleration of the mechanical speed of the generators. Depending on the location of generators relative to the short circuit, and their mechanical and electrical characteristics, their speed will change at different rates. The changes to the generators (and loads) cause different power flow in the network, and results in voltage oscillations and power oscillations between the generators. Typically, electrically close generators will exchange electrical energy, some acting as generators and others as machines, and can be considered as one generator, referred to as a generator area.

The power swings between the generators cause mechanical wear in the power plants and reduce the power quality in the network, e.g. causing flicker. The power oscillations in the ac network may also cause overloading of transmission lines and equipment, and if the oscillations persist or becomes of too high amplitude, the protection may trip lines, causing further problems in the network.

The power and voltage oscillations must be stopped as quickly as possible to minimize wear and the risk of further problems. The generator voltage regulators and turbine governors can be controlled to assist with the damping of the oscillations. However, if a VSC Transmission scheme has one terminal in one generator area subject to inter area oscillations and another in another area, or in another network, it may be able to provide substantial damping of the oscillation. The VSC Transmission scheme can provide damping by:

• Modulating active power into the oscillating area and keeping the voltage at the connection point constant.

• Modulating reactive power at the point of connection and keeping the injected active power constant. (SVC type damping).

• Modulating active and reactive power independently of each other.

The control algorithm must be determined based on a full understanding of the dynamics of the network, generators and loads. This may require extensive dynamic system studies, with and without the simulation of the VSC Transmission scheme. In some cases the dynamics will include several oscillation modes, making the determination of the optimum algorithm more difficult. Since the VSC Transmission scheme can provide two independent degrees of freedom (active and reactive power) in its operation it can provide superior damping compared to a LCC HVDC scheme and/or to a SVC.

C. Voltage Stability and Prevention of Voltage Collapse

A satisfactory operating voltage profile in a network is typically determined by automatic voltage regulators on machines, operation of tapchangers on transformers, and switching of reactive power elements (reactors and capacitors). As the network loading increases the reactive power generated by the transmission lines and cables reduce, and more and more of the voltage

11-16

supporting equipment will be brought into use. In extreme circumstances equipment and line trips can cause a deficit of reactive power, leading to a reduction in the ac voltage. As the ac voltage falls, the shunt capacitors used for reactive power support becomes less and less effective (their output drops with the square of the ac voltage). If the under-voltage persists for a long time, some equipment will eventually be tripped by under-voltage protection, as a sudden collapse in ac voltage could occur if under-voltage load shedding is not performed.

A VSC Transmission scheme can help improve the voltage stability of an area in different ways:

• By providing reactive power support to the ac network. The reactive power support available from a VSC Transmission scheme is superior to that provided by shunt capacitors, as the reactive power current can be kept at the same amplitude independently of the ac voltage.

• If the VSC Transmission scheme is in parallel to an ac line it may be possible to divert power flow from the ac line to the VSC Transmission line, thereby reducing the reactive power absorption of the ac line.

D. Power Quality

Power Quality issues includes voltage sags and spikes, power interruptions, harmonics, flicker and voltage oscillations. Certain loads, e.g. arc furnaces and rolling mills may cause flicker and other power quality problems for other customers on the network. Faults in the ac network cause voltage sags and phase shifts, the magnitude being larger the closer the fault, and can cause problems for sensitive industrial processes, such as paper mills, semi-conductor processing plants, etc. LCC HVDC schemes can also cause disturbances due to commutation failures, which locally aggravate the disturbance due to remote faults in the ac network.

A VSC Transmission scheme can respond very rapidly to voltage sags, increasing its reactive power generation to the maximum permissible level in less than a power frequency cycle. Similarly, the VSC Transmission scheme can respond very quickly to the over-voltage, which typically follows the clearing of a network fault, by increasing its reactive power absorption thereby reducing the temporary over-voltage. The VSC Transmission scheme can also respond and suppress voltage oscillations caused by power swings in the ac network. The ability of the VSC Transmission scheme to reduce voltage sags, temporary over-voltages and to respond to voltage oscillations, can be increased by ensuring that the scheme is normally operated with low reactive power output, such that maximum dynamic capability is available for this duty.

To improve the power quality for sensitive industrial processes, consideration can be given to supplying such loads via a VSC Transmission scheme. This could eliminate the voltage sags and over-voltages caused by remote ac system faults. However, in the event of faults close to the VSC Transmission rectifier terminal, the scheme would not be able to maintain power supply, unless significant energy storage were provided on the dc side of the scheme.

11.4 Practical Implementation

A simplified diagram of one terminal of a VSC Transmission scheme is shown in Figure 11-6. A similar terminal would be located at the end of the dc transmission line (in case of a multi-terminal scheme, 2 or more terminals would be connected to the dc transmission line). The VSC has a capacitor connected directly across its dc terminals, with no intervening impedance apart

11-17

from the stray inductance of the connection conductors. The different parts of the terminal equipment are described in more detail in the following sub-sections.

Figure 11-6 Basic Diagram of one terminal of a VSC Transmission scheme

11.4.1 Two- and Three-Level Converter Structure

The Voltage Sourced Converter performs the conversion between ac and dc and vice versa. For short-time transients the dc capacitor at the dc terminals of the converter can be regarded as a constant voltage source. The switching devices interconnect the dc terminals and the ac terminals using a control sequence resulting in an alternating ac waveform at the ac terminals, with a desired fundamental frequency amplitude and phase angle.

11.4.1.1 Two-Level Converter

The simplest converter implementation is a 2 level converter as shown in Figure 11-7.

dU~=

Interface Transformer Phase reactor

lL

ldconvU

convl

AC Filter

ConverterUL

11-18

Figure 11-7 Single Phase 2-level converter with switching output shown

The switching of the VSC valve at time t1 changes the ac output from +U/2 to –U/2, and the switching of the valve at time t2 changes the ac output from -U/2 to +U/2.

The ac output voltage can only attain two different amplitudes, namely +U/2 or –U/2. The diode connected in parallel to the IGBT prevents the direct voltage from changing polarity, since the diode would enter into conduction, if it were to become forward biased, thereby discharging the dc circuit. The current can flow in both directions through the VSC valve, passing through either the IGBT or the diode. This characteristic is one of the main differences between VSC Transmission and LCC HVDC.

In case of a short circuit on the dc side the diodes will conduct current into the fault, until the ac voltage at the ac terminals is removed. In order to re-establish power transmission the VSC Transmission scheme would need to be re-started again.

In principle it would be possible to switch each valve on and off just once per power frequency cycle to perform the conversion between dc and ac. However, whilst it would be possible to control the phase angle of the fundamental frequency component of the resulting waveform, there would be a fixed ratio between the dc and the ac converter voltage. Thus, in order to control the reactive power, without the use of a tapchanger, it would be necessary to change the dc voltage, which is not desirable on a transmission scheme. Additionally, the ac harmonics would be low frequency and very large.

By switching the valves on and off several times each power frequency cycle, additional degrees of freedom are available for the controls. Thus it becomes possible to control both the amplitude and the phase angle of the fundamental frequency component of the ac voltage, independently of the dc voltage, subject to limitation caused by the dc voltage. This method of operation is explained in more detail in section 11.4.2 below.

The following paragraphs describe in more detail the basic operation of the 2-level VSC shown in Figure 11-7.

Assuming both IGBTs are blocked (i.e. the IGBTs are in the high impedance state) the freewheeling diodes form an uncontrolled rectifier. An external ac voltage source applied as Uac would charge the two dc capacitors via this rectifier to the peak value of the ac voltage Udc/2 =

t0 t1 t2 t

-Uac

+

~

Udc

2

Uac

- Udc

2

Udc

2

Udc

2

11-19

Uac peak across the upper and lower dc capacitor, with polarities as shown in Figure 11-7. With the dc capacitors charged and the external source connected, the VSC is ready for operation.

The IGBTs can be switched on and off in a desired pattern via the gate signals. During operation, only one of the two IGBTs is in the conductive state and the other is turned off. Turning on both IGBTs at the same time would create a short circuit of the dc capacitors, and must be prevented. In order to avoid such a short circuit, the IGBT in the off-state is turned on shortly after the IGBT in the on-state is turned off. The short instant (about 10 microseconds), during which neither of the IGBTs are on, is called the “blanking time”, and during this time the current path is provided by the freewheeling diodes.

Assuming the upper IGBT is turned on at t = 0, as shown in Figure 11-7, the ac terminal would be connected to the positive terminal of the storage capacitor, resulting in current flow through the upper IGBT. Depending on the pulse pattern applied, the upper IGBT will be turned off after a short time span at t = t1. However, the impedance of the interface transformer and/or converter reactor (Figure 11-6), will maintain the current. Thus, the diode connected in parallel to the lower switch turns on. As a consequence, the voltage Uac changes from plus to minus U/2. The polarity reversal is initiated by turning off the IGBT actively, i.e., while it carries current.

The lower IGBT is required to be on whenever the current reverses, to maintain the voltage polarity, and therefore it will get a gate signal following the blanking time. The IGBT will then take over current as soon as the current reverses direction, the timing of which depends on the impedance of the interface transformer and converter reactor, and on the ac system voltage.

If current is still flowing in the diode parallel to the lower switch, when at t = t2 it is required to connect the converter ac terminal to the positive terminal of the VSC dc capacitor again, then it is necessary to first ensure that the lower IGBT is turned off. Therefore, the lower IGBT is turned off and the blanking time is allowed to pass, before the upper IGBT is turned on. During the blanking period the current continues to flow through the diode. When the upper IGBT is switched on, a temporary current path from the VSC dc capacitors through the upper IGBT and the lower diode back to the VSC dc capacitors is created. The current flowing in this circuit will extinguish the current in the diode, which will automatically turn off, leaving the upper IGBT as the only current-carrying device, and causing the desired voltage polarity reversal at Uac.

The three-phase implementation of a 2-level converter is shown in Figure 11-8.

11-20

Figure 11-8 Three phase implementation of a 2-level converter

Figure 11-9 shows the terminal voltages on the valve side of the phase reactor for a three phase 2-level VSC in square wave operation.

Figure 11-9 VSC voltages converter side—square wave operation

The first three sets of curves are the voltages between phase terminals and the midpoint “m” of the dc capacitors. The next three are the line-to-line voltages between the ac terminals. The last curve is the phase to neutral (n) voltage in phase a.

Uam

Ubm

Ucm

Uab

Ubc

Uca

Uan

+Ud/2

+Ud

+1/3Ud+2/3Ud

-Ud

-Ud/2t

1 fundamental cycle

Ud2

~

~

~

Uconv-ph

UL

m n

a

b

c

Ud2

11-21

With full wave, square pulse operation, i.e. each valve turning on and off once each cycle of the ac voltage, the waveform is the dual of the converter current for a six pulses LCC converter with zero commutating conductance.

11.4.1.2 Three-Level Converter

In the 2-level VSC the ac voltage at the converter ac terminals can attain only the voltage at the two dc terminals. Furthermore, the VSC valves are exposed to the full amplitude of the voltage between the two dc terminals.

By subdividing the dc capacitor and the VSC valves it is possible to arrange for the ac voltage at the VSC terminals to move not only to the voltage at the two dc terminals but also to intermediate levels. The number of voltage levels to which the ac terminal voltage can be switched will depend on the number of valves and the number of dc capacitor subdivisions or additional dc capacitors. These arrangements are known as 3-level or multi-level converters, depending on the number of voltage levels that can be achieved. The term multi-level refers to a converter topology where the ac bus can be switched to attain more than three different voltage levels. This section will focus on 3-level converters. Higher level converters can be implemented, but so far only 2-level and 3-level converters have been used for VSC Transmission.

In 3-level converters the VSC valves do not normally have to be designed for the full dc terminal-to-terminal voltage. For example, in normal operation each valve in a 3-level converter experiences only 50% of the terminal-to-terminal dc voltage.

Figure 11-10 shows a three-phase, 3-level VSC implemented with a Neutral Point Clamped (NPC) topology.

Figure 11-10 Three-phase, 3-level converter with full wave switching

The converter has three dc terminals to connect to a split or center-tapped dc source. There are twice as many valves as in the 2-level VSC, and additional diodes are required to connect to the dc supply center-tap, which is the reference zero potential. With identical valve terminal-to-

Ud

U aUb

+Ud

-Uda

Ud

Neutral

mid- pointU c

+

-

11-22

terminal voltage rating, the total dc supply voltage can be doubled so that the output voltage per valve remains the same.

The ac waveform shown in Figure 11-10 is the phase-to-neutral voltage, assuming fundamental frequency switching of the valves. The neutral voltage is the voltage at the midpoint of the dc capacitor. The output voltage of the 3-level phase unit can be positive, negative, or zero. Turning on both upper valves in a phase unit produces positive output, while turning on both lower valves produces negative output. Zero voltage is produced when one of the upper and lower middle valves are turned on, and the upper and lower outer valves are turned off, the connection to the center tap of the dc supply being via one of the two diodes. At zero voltage, positive current is conducted by the upper-middle IGBT and the upper NPC diode, and negative current by the lower-middle IGBT and the lower NPC diode.

As indicated in Figure 11-10 the relative duration of the positive (and negative) output voltage with respect to the duration of the zero output is a function of control parameter , which defines the conduction interval of the top upper, and the bottom lower valves. The magnitude of the fundamental frequency component of the output voltage produced by the phase unit is a function of . When equals zero degrees it is maximum, while at equals 90 degrees it is zero. Thus, one advantage of the 3-level phase unit is that it has an internal capability to control the magnitude of the output voltage without changing the number of valve switching operations per cycle.

The operating advantages of the 3-level phase unit can only be fully realized with some increase in circuit complexity, as well as more rigorous requirements for managing the proper operation of the converter circuit. These requirements are related to executing the current transfers (commutation) between the four (physically large) valves, with well-constrained voltage overshoot, while maintaining the required di/dt and dv/dt for the semiconductors without excessive power loss.

An additional requirement is to accommodate the increased ac ripple current with a generally high triplen harmonic content flowing through the mid-point of the dc supply. This may necessitate the use of a larger dc storage capacitor or the employment of other means to minimize the fluctuation of the mid-point voltage. However, once these problems are solved, the 3-level phase unit provides a useful building block to structure high power converters.

The conduction periods for the inner and the outer valves is different, and therefore it is possible to use two different designs of VSC valve for the two positions.

11-23

Figure 11-11 Waveshape of full wave switched 3-phase, 3-level NPC converter

Figure 11-11 shows the waveshape of a three-phase, 3-level NPC converter with full wave switching. The waveshapes clearly show that the harmonic performance of the 3-level converter is superior to that of a 2-level converter.

11.4.1.3 Other Converter Arrangements

The main incentive for using a converter topology different from the 2-level converter is the achievement of lower power loss and reduction of harmonics. Harmonics can also be reduced by switching the converter valves more than once per power frequency cycle, and this will be discussed below in Section 11.4.2.

As mentioned above, the 2-level and the 3-level NPC converter topology are just two of several different topologies that could be used for VSC Transmission. Other converter topologies proposed for VSC Transmission have included floating capacitor solutions, which may be suitable for implementation of 4 or 5-level converters [12]. Another arrangement involves the use of an interface transformer with star and delta connected valve winding voltages, to which the VSCs can be connected, and which would provide harmonic cancellation [13]. However, at the end of 2005, none of these alternative topologies had been implemented on commercial VSC Transmission schemes.

As the complexity of the converter increases, the capital cost also increases. Additionally, the space occupied by the VSC Transmission station will increase. However, the power loss for the scheme will decrease, assuming that the basic harmonic performance of the overall scheme is maintained.

U am

Ubm

U cm

U ab

U bc

U ca

+U d-Ud

+2Ud Ud

+2/3Ud+4/3Ud

U an

1 fundamental cycle

11-24

11.4.2 High-Frequency Pulse-Width-Modulation

As mentioned above, full-wave switching limits the number of degrees of freedom for a VSC converter. Since the IGBT is capable of repetitive high speed switching, modern motor drives now use VSCs with IGBTs switched at a frequency of several kHz. The advantage of such fast switching of the IGBTs is that the harmonics produced by the VSC are pushed towards higher frequency, which are easier to limit by means of shunt harmonic filters.

However, the high voltage IGBTs used in VSC Transmission have higher switching loss than the lower voltage devices used in motor drives. Therefore, the power capability of the power device decreases with increasing switching frequency. Furthermore, the power loss is an important element in the economic assessment of a VSC Transmission scheme, so a careful optimization process has to be carried out to determined the optimum switching frequency for a given application. Typically, the switching frequency used in a VSC Transmission scheme will be less than 2000Hz.

Many different control strategies can be used for converters switched more frequently than once per cycle [14] [15]. Three different strategies will be described in this section:

• Carrier-modulated method with pure sinusoidal ac voltage reference

• Carrier-modulated method with 3rd harmonic modulated ac voltage reference

• Selective harmonic elimination method (SHEM).

Figure 11-12 PWM controlled VSC with pure sinewave ac voltage reference and triangular carrier

Figure 11-12 shows the derivation of the switching instants for a control methodology using a triangular carrier wave (Vcarrier) and a pure sine wave ac voltage reference (Vcontrol). ). The figure shows the derivation of the control signals (a), and the resulting waveshape (b).

Vam

+Ud/2

-Ud/2

(a)

(b)

a b c

11-25

The intersection between the carrier triangles and the sine wave determines the switching time for the VSC valves. When Vcontrol is greater than Vcarrier the converter output is positive, and when Vcontrol is smaller than Vcarrier the converter output is negative. The ratio between Vcontrol and Vcarrier is known as the modulation index m.

The value of the modulation index can be any value between 0 and 1. The fundamental frequency component of the VSC ac voltage varies linearly with the modulation index, with a maximum voltage equal to the dc voltage.

In the example shown in Figure 11-12, the frequency of the carrier wave is 9 times the power frequency. If the carrier frequency is an odd integer multiple of the fundamental frequency the ac waveform does not contain any even harmonics. In a three-phase VSC all of the triplen harmonics, i.e. 3rd, 9th, etc are eliminated in the phase to phase voltages. Both of these statements assume that the ac network is balanced and free of background harmonic distortion.

Figure 11-13 PWM switched 2-level VSC with carrier frequency at 21st harmonic.

Figure 11-13 shows a typical voltage harmonic amplitude at the ac terminals of the VSC, when using PWM switching with a carrier frequency of 21 times the fundamental frequency. The

VcarrierVcontrolm =

(b)

(c)

(a)

11-26

spectrum changes with operating conditions. The figure shows the stepped ac waveshape, and in dotted line the fundamental frequency component of this waveshape (a), the harmonic voltage for phase to neutral (b), as well as the harmonic voltage phase to phase (c).

Figure 11-14 PWM switched VSC using sinusoidal PWM and 3rd harmonic injection.

Figure 11-14 shows the voltage harmonic spectra using a sinusoidal control voltage with 3rd harmonic injection and triangular carrier. The 3rd harmonic injection enables an increase of the fundamental frequency component of the VSC voltage, for the same dc voltage, giving greater utilization of the converter equipment. The 3 sub-figures are as described for Figure 11-13.

As an alternative to the use of the simple triangular carrier wave for the determination of the VSC Valve switching instants, it is possible to use pre-calculated switching instants, which are determined such that specific harmonics are eliminated. This control method is called the Selective Harmonic Elimination Method, SHEM for short, or Optimised Pulse Width Modulation, OPWM for short. Since the switching instant required in order to eliminate specific harmonics depends on the operating conditions, it is either necessary to pre-determine the instants for the full range of normal operating conditions, or they must be determined on-line by the algorithm.

Half-wave and quarter-wave symmetries have to be preserved in the waveform, to ensure that even harmonics are eliminated. If there are say 4 switching instants in a quarter cycle, then one

(a)

(b)

(c)

11-27

of these is normally required for the control of the fundamental frequency component of the ac waveform. The other 3 can be used for the elimination of 3 harmonics, say the 5th, 7th and 11th harmonic.

Figure 11-15 shows the PWM waveshape and the voltage harmonic spectra for a VSC using SHEM or OPWM. The 3 sub-figures are as described for Figure 11-13.

Figure 11-15 PWM waveforms and harmonics for VSC operated with SHEM modulation

It can be seen that even with the same number of VSC Valve switch operations in a power frequency cycle, the harmonic content in the ac waveshape is considerably reduced, when using the SHEM control method. This means that either smaller ac harmonic filters can be used, reducing the site area required and capital cost, or fewer VSC Valve switch operations can be used, thereby reducing the power loss. The switching strategy can also be optimized to reduce the number of operations at high current, thereby further reducing the power loss.

11.4.3 High Voltage Turn-Off Valves

The VSC Valves must be capable of turning on and off in response to control signals, irrespective of the current flowing though the valve, or of the voltage across the valve, at the time of the valve operation. This characteristic makes VSC valves fundamentally different from the thyristor valves used in a LCC HVDC scheme.

(a)

(b)

(c)

11-28

Several different semiconductor devices capable of actively turning off the flow of current are now available [4]. For a VSC Transmission application many devices must be connected in series in order to achieve the required voltage withstand, and this imposes additional requirements on the semiconductor. At present (2005) the Insulated Gate Bipolar Transistor (IGBT) is the device of choice for the following reasons:

• It has transistor action, which enables precise control of the device in a manner that is not possible with latching alternatives. For instance:

o The converter can be turned off even in short circuit conditions.

o It is possible to perform active voltage control, which is useful for the sharing of the voltage between a large number of series connected devices.

• Control of the device is achieved using low power, since it is a MOS-controlled device. This is advantageous when operating at very high voltage levels, where auxiliary power has to be obtained either from the power circuit itself or from ground level across the insulation barrier.

• It is capable of high switching speed, thus making high switching frequency feasible.

New semi-conductor devices are being developed all the time, and future schemes may use different devices. At the end of 2005 the maximum voltage withstand capability of an IGBT was about 6.5kV, but such devices have considerably higher power loss than lower voltage devices. The high power loss reduces the current capability, particularly if the devices are switched at high frequency. Therefore, lower voltage devices are currently used for VSC Transmission.

The IGBT is designed to have forward voltage blocking capability only, since a parallel diode provides current capability in the reverse direction, providing protection against reverse voltage. High blocking speed diodes, with low recovery charge, are used in order to minimize power losses. Silicon Carbide diodes have been tested in VSC Transmission schemes, but the benefits obtained because of their low recovery charge did not justify the additional costs. Both the IGBT and diode are usually implemented as chips, and are mounted in the same integral housing.

In order to increase the reliability of the VSC Valves, the series strings of IGBTs typically include a number of redundant devices, such that operation can continue with satisfactory insulation margins, even when one or more IGBT devices have failed. Presspack mounting, as used for thyristors for LCC HVDC schemes, is also used for the IGBT in a VSC Transmission scheme in order to ensure that a failed device can continue to provide a safe and reliable current path until the next maintenance outage. However, because of the use of numerous chips, the construction is more complex than for a thyristor device, and presspack IGBTs are therefore considerably more expensive than the more usual IGBT modules.

The voltage sharing across the series connected devices in a VSC Valve during turn off and in the off state can be controlled by using the IGBT’s transistor action. Snubber circuits could also be used, but whilst these would reduce the power loss in the device, they would also create their own power loss, since the energy absorbed during turn off has to be released when the IGBT turns on again.

To reduce the VSC Valve stress and the power loss, the inductance in the loop formed by the VSC Valve and the dc capacitor must be kept as small as possible. This consideration imposes strict requirements on the arrangement of the VSC Valve, the dc capacitor and the overall circuit.

11-29

The current capability of the IGBT must be determined taking into account not only the steady state operating conditions, but also the transient conditions encountered during various faults and abnormal operating conditions. Significant power loss is caused in the device during turn-off, since high voltage and high current are present simultaneously, as shown in Figure 11-16. Effectively, the current in the device is driven to zero by the voltage developed across the IGBT.

Figure 11-16 Voltage and Current across an IGBT during switching.

During faults in the ac network or close to the converter, the IGBTs must protect themselves by interrupting the fault current before too much heat has been developed in the device. In order to achieve this, protective action must be taken before the current has risen too high. As the current through the device increases, the IGBT de-saturates, thereby increasing the resistance to the fault current. Gate control further enhances this natural response to stop the current quickly and completely. This requirement imposes constraints on the design of the VSC control and protection system.

The diodes in parallel with the IGBT form an uncontrolled rectifier in the event of a short ciruit on the dc side of the VSC, and IGBT action cannot stop the current. Therefore, the diodes must be rated to cope with the short circuit current, until the ac side of the VSC has been disconnected from the ac network by circuit breaker action.

The power loss caused in the devices during the operation of the VSC must be removed using a cooling system, to ensure that the device temperature is kept in the acceptable design range. The design of the cooling system is very similar to that used in a LCC HVDC scheme. Typically, the valves will be water-cooled and the on-valve pipe work must be designed taking into account the dc stress within the valve structure and between metallic components. To achieve high reliability, similar to that of the VSC Valves, redundant cooling pumps, fans, instrumentation and controls are typically used.

The mechanical implementation of the VSC valve must take into account the requirements for low circuit inductance, electrical clearances, space for cooling, access for maintenance etc.

Voltage

Time

Current

Voltage & Current

Time

Power Loss

Power Loss = U(t)*I(t)

11-30

Figure 11-17 shows VSC valves inside a valve enclosure. A single valve for a ±150kVdc, 2-level, VSC Transmission scheme, can contain more than 300 series connected IGBTs. The enclosure is made of steel and aluminium, and acts to contain the electromagnetic interference caused by the switching of the VCS Valves. For some applications the enclosures can be implemented as transportable containers, such that a significant part of the checking and testing can be done in the factory, before the valves are sent to site.

Figure 11-17 VSC valves inside Valve Enclosure.

11.4.4 Harmonic Filtering

The ac voltage waveshape produced by the VSC is far from sinusoidal, as can be seen in the Figures in section 11.4.2. Therefore, filtering is required to make the ac voltage acceptable to the ac network [16][17][18]. The phase reactor and ac harmonic filters perform the required filtering. The filtering required depends on:

• The amplitude and frequency of the harmonics produced by the converter.

• The characteristics of the ac network.

• The permissible distortion limits.

It is also necessary to consider the harmonic performance on the dc side, even if the connection between VSC Transmission terminals is by means of cables. This is because harmonic resonance could cause magnification of the harmonics, resulting in localized overloading and premature aging. Furthermore, if the dc cable has limited screening, interference could also be

11-31

injected into adjacent communication circuits. Finally, if an overhead dc line were used between the terminals, interference could be injected in nearby telecommunication circuits.

The determination of the harmonics in the ac and dc output is beyond the scope of this chapter. The harmonics depends on many factors, including:

• The operating conditions.

• The converter control strategy.

• The converter topology.

• The impedance of the converter components.

The harmonics can be determined by means of Fourier analysis of the wave-shapes, or by mathematical analysis. It should be noted that harmonics are also influenced by the characteristics of the ac and dc circuit to which they are connected, in particular the balance of these circuits, and the presence of any pre-existing harmonics

On the ac side the phase reactor plays a significant role in ac harmonic filtering by acting as a low pass filter. The phase reactor and the interface transformer (if used) will change the appearance of the VSC from a voltage source towards that of a current source. One or more ac harmonic filters are connected on the ac system side of the phase reactor, as shown in Figure 11-18.

Figure 11-18 Simplified diagram of one terminal of a VSC Transmission scheme.

The ac harmonic filter at each terminal will typically have a total rating of 10-20% of the converter rating. The design will typically consist of 2 or 3 filter branches, one or more of which being in ungrounded star or delta configuration. Typically the filter branches will include damped high pass filters, to provide damping and attenuation at high frequency. Tuned filters may also be used, e.g. at the PWM carrier frequency.

The ac harmonic filters tend to be easier to design than equivalent filters for a LCC HVDC scheme, since the harmonics to be filtered are at higher frequency. However, as for a LCC HVDC scheme, care need to taken to avoid resonance between the filters and the ac network at frequencies where significant pre-existing harmonics may exist in the ac network.

dU~=

Interface Transformer Phase reactor

lL

ldconvU

convl

AC Filter

ConverterULDC Reactor

11-32

For a dc cable scheme the dc capacitor and dc smoothing reactor normally provides adequate harmonic filtering. However, as mentioned above, if sensitive telecommunication circuits are located close to the cable route, then it may be necessary to provide additional filtering. This can be provided by additional series reactance, and/or by a shunt dc filter.

11.4.5 Phase Reactor

The phase reactor performs the following important role:

• It provides constant fundamental frequency impedance for the control of the VSC active and reactive power output.

• It provides a high frequency blocking filter between the VSC and the ac network.

• It limits short circuit currents.

Typically, the phase reactor has a short circuit impedance of about 15%, and it is implemented as an air-cored and air-insulated design. The reactor can be very large, and is typically enclosed in a metallic shield, to prevent the escape of high frequency electric and magnetic fields. Forced air circulation is used to cool the reactor.

Because of the duty as a high frequency blocking filter the phase reactor is continuously exposed to the voltage steps in the ac output voltage. The reactor design must take this stress into account, to ensure that the phase reactor will perform satisfactorily throughout the life of the VSC Transmission scheme.

The phase reactor must also have very low stray capacitance between its terminals, since high frequency current might otherwise bypass the reactor. The need for low stray capacitance makes the air-cored and air-insulated design particularly attractive.

In principle, the stepdown transformer could perform the phase reactor’s role. However, the transformer would then be exposed to repetitive high frequency transients. Additionally, the stray capacitance in an oil immersed transformer is likely to be significantly higher than for an air-cooled, air-insulated reactor, so it would not perform as effectively as a high frequency blocking filter. Therefore, the use of a transformer only solution is likely to be limited to low power and low voltage applications, where a dry type transformer can be considered.

11-33

Figure 11-19 Phase Reactor for a 65MVA, ±80kVdc VSC Transmission scheme

11.4.6 DC Capacitor

The dc capacitor provides the voltage source for the operation of the VSC. The dc capacitor also plays an important part in the harmonic filtering on the dc side. The dc capacitor is connected directly across the converter dc terminals, and recent VSC Transmission schemes have used dry type dc capacitors, to eliminate the risk of fire.

High frequency current flows in the dc capacitor during converter operation. It is important to keep the inductance of the loop formed by the VSC valves and the dc capacitor as small as possible, in order to minimize the VSC Valve stress and the power loss resulting from the stored energy in the current when the valves are switched off. Additionally, the design of the dc capacitor itself must have as low stray inductance as possible.

The high frequency current causes electromagnetic noise, which must be contained as it cannot be eliminated at source (the high frequency current is necessary for the operation of the converter). Typically, the dc capacitor is mounted inside the shielded valve enclosure as close as possible to the VSC valves.

11-34

The flow of current causes a change of the voltage across the dc capacitor, and this change is known as the voltage ripple. The amplitude of the voltage ripple depends on:

• The capacitance of the dc capacitor - the larger the capacitance the smaller the ripple.

• The VSC Valve switching strategy - long conduction periods at high current cause larger ripple.

The dc capacitor ripple is also affected by unbalance and pre-existing harmonics in the ac network.

The dc capacitor ripple can have an impact on the generation of harmonics on the ac side, and therefore the ripple is typically limited to less than 5% of the rated direct voltage.

The dc capacitor must also be rated such that the dc voltage is kept reasonably constant during fast dynamic changes to the VSC Transmission scheme operating conditions. The voltage change during transient conditions, e.g., a short circuit to ground in the ac network, must also be taken into account in the control and protection strategy for the converter.

11.4.6 Interface transformer

A transformer may be used to couple the converter ac filter bus and the ac network. The use of a transformer fulfills the following tasks:

• It enables the VSC to be designed independently of the ac network voltage constraints. If the transformer has an online tapchanger, it is possible to adjust the voltage on the converter side of the transformer independently of the ac network voltage, to achieve optimum converter performance and reduction of the steady state power loss.

• It blocks the flow of zero sequence current between the ac network and the converter.

• It provides additional series impedance on the ac side, which may be beneficial for the harmonic performance of the converter, and may enable a reduction in the rating of the phase reactor.

Since the transformer is not exposed to direct voltage, and the harmonic filtering is performed on the converter side of the transformer, a conventional station transformer can be used. A tertiary winding can be provided for the connection of the auxiliary supply system.

In some applications a transformer may not be necessary, e.g. if the ac network voltage level is such that no voltage transformation is necessary.

11.4.7 Other VSC Transmission sub-station equipment

A typical VSC Transmission substation includes a number of items in addition to the main equipment described above. These include:

• AC circuit breakers. Typically a circuit breaker is only used at the connection to the ac network, since the ac harmonic filter does not need to be switched during the operation or

11-35

starting of the station. The ac circuit breaker may include a pre-insertion resistor, in order to limit the dc capacitor voltage overshoot during first energisation. In some projects the energy absorption requirements for the inrush limiting resistor is so large that a separate resistor with a parallel circuit breaker has been connected in series with the normal breaker to fulfill this role.

• RFI and PLC filters. A small high frequency blocking reactor, and a shunt capacitor is used to limit the injection of electromagnetic disturbances into the ac network to an acceptable level.

• Voltage and Current measuring transducers are required at a number of locations in order to provide information for the control and protection systems. The transducers may be duplicated, unless backup control and protection use different inputs.

• Surge arresters are used for over-voltage protection. Arresters are used on either side of the converter.

• Disconnectors and earth switches are used at the ac and dc terminals to enable the equipment to be isolated and made safe for maintenance work.

• Auxiliary power supplies. Typically, these will be taken from two independent sources, one of which may be a tertiary winding on the interphase transformer. The auxiliary supply system will typically provide duplication of the supply to all critical equipment, with automatic changeover in the event of a failure of the primary supply.

• Fire Protection. The requirements are similar to those for a LCC HVDC scheme, but the equipment can be smaller, since the VSC Transmission station and the equipment within it is smaller and more compact.

• Civil works. The design of the civil works must take into account the requirements for housing and environmental control for the power equipment, protection and control, as well as of any operator and maintenance facilities. Additionally, the civil works must be designed to limit the audible and electromagnetic noise from the operating equipment.

Figure 11-20 shows the layout of the equipment for one terminal of the Gotland VSC Transmission scheme, which has a rating of 50MW and operates at ±80kVdc.

11-36

Figure 11-20 Layout of 50MW, ±80kVdc VSC Transmission substation

11.4.9 DC Cables

The transmission of energy between the terminals in a VSC Transmission scheme is achieved by dc cables. In principle, when the transmission is over land it could also be achieved using an overhead dc line. However, as discussed above, a fault on the dc side impose additional stress on the diodes and converter equipment, and requires tripping of the ac circuit breakers at all terminals, to clear the fault. Therefore, at present (end 2005) all commercial VSC Transmission schemes use dc cables.

The use of dc cables has a number of advantages compared to the use of dc overhead lines:

• Cables often have less environmental impact than an overhead line.

• The right of way required for a cable installation is considerably less than that for an overhead line, and the land above the installed cable can often be returned to its previous use.

• Cables can be run next to roads and through existing tunnels.

• Cables are much less prone to faults than overhead lines. In particular, lightning strikes or pollution do not affect a cable.

The voltage polarity on a dc cable in a VSC Transmission scheme is fixed and independent of the power direction. This enables the use of XLPE type polymeric cables for VSC Transmission. Polymeric type cables cannot be used for LCC HVDC schemes, because the high resistivity and

Building45 x 18 m

AC Yard & Harmonic filters

Phase Reactors

Phase A, B and C valve compartments

Auxiliary Power System& Cooling Control

Cooling towers

DC Yard Equipment

11-37

low mobility of trapped charges in the material would cause unacceptable and potentially damaging voltage distribution within the dc cable immediately after a voltage reversal[19].

Polymeric cables can be lighter and have smaller bending radii than the conventional oil-impregnated or oil-filled cables, which are used with LCC HVDC schemes. Therefore, polymeric land cables are easier to handle during installation, and special direct ploughing and laying techniques can be used where the ground conditions are favourable, making installation less time consuming. An additional advantage of the use of polymeric cables is that the environmental risks are lower than with an oil-impregnated or oil-filled cable.

Transport logistics limits the length of each section of land cables to 1 – 2 km. This means that land cables have many field joints. The joint of a cable tends to be less reliable that the cable itself, and it is important to ensure that adequate type and life testing has been performed on the joints before the cable is installed. Furthermore, the field joints must be installed under clean and controlled conditions. Figure 11-20 shows a mobile jointing container in which the work can be performed. Generally, the performance record of cables for VSC Transmission schemes has been very good, one exception being the Direct Link scheme in Australia, where many joint failures occurred in the early years after the scheme entered service, due to a design defect.

Figure 11-21 Jointing Container used for the Murray Link scheme in Australia

Submarine cables can also be provided as polymeric cables. Submarine cables require armoring, to enable the cable to withstand the mechanical stresses imposed during the laying operation. Submarine cables are normally installed from specially equipped ships with the cable coiled on a large turntable. Continuous cable lengths of up to 100km can be laid, minimizing the number of cable joints required.

11-38

11.5 Basic Operational Limits

A simplified steady state operating range of one of the terminals in a VSC Transmission scheme, expressed as the VA capability at its ac connection point, is illustrated in Figure 11-22. The figure has been drawn for 3 different ac network voltages, to illustrate the dependence of the capability on the ac voltage. The use of a tapchanger on the interface transformer can remove this dependence for steady state operation, thereby enabling the capability of the converter to be utilized more fully.

Figure 11-22 Simplified PQ characteristic of a VSC Transmission terminal

The maximum current capability of the VSC Valve dictates the MVA capability at a given ac voltage. Assuming that the IGBT and the diode have the same current capability, the MVA capability at a given ac voltage can be assumed to be described by a circle. The current capability of the IGBT and the diode do not necessarily have to be identical. The thermal duty on the IGBT consists of the switching loss and the conduction loss. The switching loss exceeds the conduction loss when the IGBT is operated at a relatively high switching frequency. Therefore the difference in the thermal duty is not dramatically different, whether the converter operates as a rectifier or as an inverter. For the diode the switching loss tends to be much smaller than the conduction loss. Since the conduction duty on the diode is greater when the converter acts as a rectifier than when it operates as an inverter, it would in principle be possible to design the scheme such that it has a higher transfer capacity in one direction than in the other. However, the actual saving has to be considered in the context of the overall solution, and must take into account the benefits resulting from the use of standard building blocks. Therefore, the converters at the two ends of a VSC Transmission scheme are normally identical, and provide the same power flow capability in both directions.

In the example shown in Figure 11-22 the active power capability of the VSC Transmission scheme exceeds the desired capability at all ac voltages within the range of Umin to Umax.

Rectifier Mode

Inverter Mode

CapacitiveInductive

Pconv

Qconv

Desired Active Power

Desired Reactive

Power

Uac = MaxUac = NomUac = Min

DC cable Thermal

limit

11-39

In addition to the limit on operation caused by the thermal duty on the semi-conductors, another number of issues must also be considered:

• The voltage on the dc capacitors and dc cables must be kept low enough to provide safe operation for both the VSC valve and the dc capacitor and cable themselves. The maximum ac voltage produced by the converter depends on the direct voltage on the dc capacitor. This then results in a limitation on the capacitive reactive power, which the converter can produce. Since the generation of reactive power requires the converter voltage amplitude to be higher than the ac network voltage, the capacitive power capability falls with increasing ac voltage. In the example shown in Figure 11-22 the capacitive power capability is lower than the desired output at the maximum ac network voltage. In practice this is unlikely to be of any consequence, since generation of reactive power is unlikely to be required when the ac network voltage is already higher than the nominal value. When specifying the required reactive power capability, it is necessary to state explicitly the ac voltage and active power exchange at which the reactive power capability is required.

• The dc cable may impose a restriction on the maximum dc current allowed to flow between the VSC Transmission terminals. This limitation will typically be just above the desired power transfer capability, and the VSC control system will ensure that this limit is not exceeded. For clarity the dc cable limit in Figure 11-22 is shown significantly higher than the desired power transfer capability.

• In addition to the limitation on the capacitive power capability at high ac system voltage, there is also a minimum dc voltage limit for the steady state operation of the VSC Transmission converter. This limits the reactive power that the converter can absorb.

11.6 Converter Rating and Protection Considerations

The rating of the VSC Transmission substation must take into account a number of dynamic and transient conditions in addition to the steady state design considerations as outlined in section 11.5 above. The voltage withstand requirement for components and insulation must be determined based on the over-voltages that may be experienced due to any credible event. Similarly, the equipment must be capable of withstanding the transient and dynamic current surges resulting from any credible event outside and within the VSC Transmission station.

11.6.1 Over-voltage Stresses and Protection

When the first VSC Transmission substation is energized by the closing of its ac circuit breaker the converter diodes acts as an uncontrolled rectifier charging the local and remote dc capacitors and the dc cable. Under some conditions, the combined impedance of the ac network and interface transformer and phase reactor will react with the total dc system capacitance such that an over-voltage is created. The over-voltage can be reduced by control action as soon as the IGBTs can be switched. If the initial over-voltage is unacceptable, then it may be necessary to use a pre-insertion resistor on the ac circuit breaker, or to use other techniques to limit the over-voltage. When the interface transformer has a tapchanger, it is normally placed in the position giving the lowest converter side ac voltage, prior to energisation of the transformer.

11-40

In some multi-level converter circuits, uncontrolled rectifier charging of the dc circuit will result in unequal voltage distribution on the dc capacitors. The rate of charging can be slowed down by the use of a pre-insertion resistor, such that the IGBTs have time to control the dc capacitor voltages.

Figure 11-23 DC over-voltage on rejection of 330MW

During a sudden load rejection at the inverter a significant over-voltage may be experienced on the dc side. The over-voltage will be reduced very quickly by the converter control, as can be seen in Figure 11-23, which has been taken from an ABB site measurement and digital simulation. The figure shows a load rejection of 330MW at 0.47seconds, which results in an increase of the dc voltage from about 147kVdc to a peak value of 205kV within 10ms. The voltage is reduced to less than 160kV within a further 20ms.

Just like any ac substation, a VSC Transmission substation will be subjected to transient voltages caused by lightning strikes on ac overhead lines, or even within the substation itself. Overvoltages caused by switching events, e.g., the clearance of line faults or equipment failures in the ac network, and within the VSC Transmission substation itself, may also be experienced.

In the VSC Transmission substation surge arresters are typically connected close to the interface transformer, to limit the over-voltage amplitude and to protect the transformer and other power equipment. Surge arresters are typically also provided on the converter side of the interface transformer in order limit any transferred over-voltages. There is no significant difference between the application of surge arresters to protect against lightning, and switching surge over-voltages and the insulation co-ordination process for a conventional ac substation and a VSC Transmission substation in this respect. The interface transformer will act as a barrier to the transfer of over-voltages from the HV bus to the converter ac harmonic filter bus. However, care must be taken to determine the transfer of over-voltages through the transformer, taking into account the stray capacitance across and between windings, and the impedance of the components on the converter side of the interface transformer.

11-41

It should be noted that if pre-insertion resistors were used to limit the over-voltage on first energisation, these pre-insertion resistors would not be in the circuit when the ac voltage suddenly returns after the clearing of a local three-phase fault to ground. Typically the control and protection system would prevent the dc circuit from fully discharging in the event of a major fault in the ac network. Therefore, energisation of a discharged dc circuit without pre-insertion resistors can be considered to be a double contingency failure.

Whilst the ac harmonic filters are smaller than those used on a LCC HVDC scheme, it is still important to check that there is no resonance between the filter capacitance and the combined impedance of the interface transformer and ac network. If a resonance is possible, then it may cause magnification of the switching surge type over-voltages appearing during clearance of faults in the ac network. The surge arresters may limit the over-voltages to acceptable levels, but in case of a resonance the energy absorbed may be large, and need to be taken into account in the rating of the arrester.

In a 2-level converter the VSC valves are rated to withstand the full voltage between the dc terminals, therefore major over-voltages across components will not be caused by faults within the converter.

In multi-level converters internal faults in the converter or a control malfunction could potentially cause the build up of significant unbalance of the dc capacitor voltages. Surge arresters will prevent these over-voltages from causing damage to other converter equipment. However, if the condition were allowed to continue the surge arresters would be overloaded and suffer a catastrophic thermal run-away. Therefore, voltage monitoring and protection is necessary to detect the onset of such a situation.

A breakdown to ground internally in a multi-level converter can result in an instant and large over-voltage across VSC valves and other components. Surge arresters will limit the over-voltage and prevent equipment damage but, to prevent the over-voltage from continuing and overload the surge arrester, it is necessary to trip the converters at both ends of the VSC Transmission scheme.

When available, communication between the terminals can be used to speed up the protective action. However, the protection should be designed such that equipment is safe even without inter-station communication, and such that it can detect that a fault has occurred at the remote terminal.

11.6.2 Over-current stresses and protection

During start up of the VSC Transmission substation the diodes in the first terminal to be energized will act as an uncontrolled rectifier and charge the local and remote dc capacitors and the dc cable. The diodes must be rated for the current duty associated with this event.

The interface transformer and phase reactor will block lightning strikes in the ac network, and the lightning current would be diverted to ground by the surge arresters. The design of the VSC Transmission substation in respect of over-current stresses due to lightning strikes is therefore no different from the design of a conventional ac substation.

11-42

In the event of a short circuit to ground in the ac network close to the VSC Transmission terminal operating as an inverter, there would be a rapid increase in the converter ac current. The combined impedance of the phase reactor and the interface transformer will limit the rate of rise of the current. As the current increases, the IGBTs will de-saturate, increasing the voltage drop across the device, and the protection will take action to actively reduce the over-current, by turning off the IGBT. The design of the converter and its protection must be such that the IGBT stress during this event remains within the IGBT’s safe operating area, and such that the converter can immediately resume operation, as soon as the fault in the ac network has been cleared.

If the short circuit to ground appears within the VSC Transmission terminal, the rate of rise of the fault current will be higher than when the fault appears in the ac network, since the impedance in the fault current path will be lower. The protection system will still limit and stop the fault current, but the ac circuit breaker must also be tripped, to prevent further damage.

The operation of the VSC includes a “blanking period” (see section 11.4.1.1) to prevent the IGBTs in the upper and the lower valve groups from conducting at the same time. If both IGBTs were to conduct, a so-called fire-through would be caused, creating a short circuit of the dc capacitor. Since the inductance in the VSC Valve and dc capacitor loop is kept very small by design, the rate of rise of the fault current would be extremely high, and the risk of IGBT failure would be substantial. The risk of a control failure causing such a short circuit to occur across the dc capacitor must be made very small by design.

During a fault on the dc side of the VSC, the diodes at both VSC Transmission terminals will act as uncontrolled rectifiers, resulting in large current flowing into the fault. The current will continue to flow until the ac circuit breakers at the VSC Transmission terminals have been opened. The fault will be detected by the converter protection, and immediate tripping of the circuit breaker is initiated, in order to limit the thermal stress on the diodes, and to remove the voltage stress. Whilst faults are rare on a cable transmission system, a short circuit on the dc side must be taken into account when rating the diodes.

11.7 Practical Operating Characteristics

11.7.3 Terminal Voltage and Current relationships

When the VSC Transmission substation is connected to an ac network which has other sources of generation and loads, the relationship between the VSC Transmission ac voltage and current depends on:

• the amplitude and phase angle of the converter voltage (or more precisely, the ac filter bus voltage) relative to that of the ac system EMF, and

• the impedance between the converter EMF and the ac system EMF.

This relationship was discussed in Section 11.3 for normal operating conditions, and it was assumed that the impedance between the two EMFs was largely inductive. In this case the active power flow into the ac network depends on the phase angle between the two EMFs, and the

11-43

reactive power flow depends on the relative amplitude of the two EMFs. The VSC Transmission terminal therefore operates electrically similarly to a generator, and the active and the reactive power can be controlled independently of each other. In this operating mode the frequency used for the control of the VSC is determined from an electrical measurement of the ac network voltage.

If the VSC Transmission terminal is connected to an ac network which does not have its own generation, then the absolute phase angle of the converter voltage is of no significance. The frequency used for the control of the VSC is determined purely by the control system, in accordance with settings from the network control center. The current flowing into the ac network is determined purely by the amplitude of the converter voltage, and the combined impedance as seen from the VSC Transmission ac terminal plus the impedance of the terminal itself. As the load in the ac network increases the converter EMF must be increased to maintain the ac voltage at the connection point. The load current on the ac network transmission lines may change the reactive power flow in the ac network, and therefore the ac voltage profile in the ac network. The VSC Transmission scheme is able to adjust/control the voltage amplitude at its connection point, but in doing so it will also change the active power delivered to the loads. Thus, when feeding a passive load there is not true independence between the active and the reactive power at the VSC Transmission ac terminals.

The maximum amplitude of ac current that the VSC Transmission scheme can inject into a short circuit ac network will be close to the maximum current capability of the scheme. Therefore, special consideration needs to be given to the ac network protection, when a VSC Transmission scheme is the sole supply of power (see section 11.8.3).

When the ac network is extremely weak at the point of connection of a VSC Transmission terminal it becomes more difficult to control the VSC. The frequency used for the control of the VSC has to be derived from the ac network voltage, in order to prevent unacceptable active power flows in the ac network. It is also desirable to control the ac voltage at the connection point to a “constant” value, the “constant” value being changed in accordance with a pre-set strategy e.g. a slope setting, as the active load changes. In order to avoid power swings in the ac network, it may be desirable to control the amplitude of the active current flowing into the network to a “constant” value.

11.7.2 Active Power vs. Reactive Power Generation/Absorption Capability

The PQ capability of a VSC as described in section 11.5 did not take into account the presence of the interface transformer impedance or the ac harmonic filter impedance. These impedances modify the capability of the VSC, shifting the characteristics towards more reactive power absorption or generation capability, depending on the relative impedance of the two elements.

When required the capability of the VSC Transmission terminal can be deliberately shifted further towards absorption or generation, by the connection of a shunt reactor or a shunt capacitor, either on the VSC side or the network side of the interface transformer. When the reactive power elements are connected on the VSC side the rating of the interface transformer may have to be increased. By connecting any additional reactive power elements by means of circuit breakers, the capability could be shifted as necessary to suit the particular transmission

11-44

and loading conditions, thereby optimizing the converter operating conditions and reducing the power loss.

11.7.3 Operating Losses vs. Transmitted Power

Significant reductions of the power loss, relative to the first generation of VSC Transmission schemes, have been achieved by the development of the technology. The first generation of VSC Transmission used 2-level converters with triangular carrier wave PWM control. Subsequently 3-level converters were used, and the latest large scheme (2005) will use 2-level converters with SHEM or OPWM control. Additionally, improvements have also been made in the design of VSC Valves and other equipment. As a result of these developments the power loss at rated power has been reduced from approximately 3% to approximately 1.9% per terminal. The power loss is still considerably larger than the power loss for a LCC HVDC converter station, which would typically be about 0.8% at rated power. Whilst future development and optimization is likely to result in a further fall in the power loss, it is unlikely that it will get as low as for a LCC HVDC scheme, since the conduction loss of the silicon based IGBT is significantly larger than that for a thyristor. Furthermore, the VSC Transmission scheme will have higher power loss because of the IGBT switching loss and the higher frequency switching of the VSC.

A formal methodology has not yet been established for the determination of the power loss of a VSC Transmission scheme. The major contributions to the power loss of a VSC Transmission scheme are as follows:

• The VSC Valves

• The interface transformer

• The phase reactor

• The ac harmonic filter

• Auxiliary/service loads

• The dc smoothing reactor Simplistically, the power loss in the VSC Valve is caused partly by the conduction loss and partly by the switching loss. The switching loss is caused primarily by the simultaneous presence of high voltage and high current during the turn-off process. Figure 11-16 in section 11.4.3 illustrated the turn-off process for an IGBT, and the power loss caused during this process. Whilst the IGBT power loss due to conduction varies almost linearly with the converter current, the switching power loss has a substantial value, even at virtually zero converter current.

The conduction and the switching power loss in the IGBTs and the diodes have different values. Therefore, the VSC Valve power loss depends on the direction of power flow, the terminal acting as rectifier tending to have slightly lower power loss than the inverter terminal.

The VSC Valve power loss also varies with the reactive power absorbed or generated by the converter. The total power loss for any given conditions must be determined based on the detailed current flowing through the different valve components during a power frequency cycle.

11-45

The power loss in the interface transformer is determined as for a conventional transformer, since the harmonic current flowing through the transformer is negligible, compared with the fundamental frequency component. The transformer power loss includes a no-load loss and a load loss.

The power loss in the phase reactor is largely due to the flow of fundamental frequency current. As for the interface transformer, the harmonic current flowing through the reactor is negligible, compared with the fundamental frequency component.

The power loss in the ac harmonic filter is dominated by the harmonic contributions.

The auxiliary power required for the operation of the VSC Transmission substation tends to have a large fixed component (no-load loss), and a smaller component that varies with the operating condition of the converter, e.g. cooling plant losses etc.

Figure 11-24 shows the approximate variation of the total power loss for the two terminals of a VSC Transmission scheme. The power loss shown does not include the power loss in the dc cable, which will naturally depend on the length and design of the cable.

00.5

11.5

22.5

33.5

4

0 20 40 60 80 100VSC Converter Load (%)

Pow

er L

oss

(% o

f rat

ed p

ower

)

Figure 11-24 Variation of power loss with load for the two substations in a VSC Transmission scheme.

In the absence of a standard way of determining the power loss for a VSC Transmission scheme, consideration may be given to the measurement of the power loss. Such an approach would not be easy, and would require very accurate measuring transducers at the ac terminals of both VSC Transmission stations, assuming that the total power loss including the dc cable is to be determined. The measurements would have to be performed at all operating conditions at which performance has been guaranteed. Unless the auxiliary power for the operation of the scheme has been derived from tertiary windings on the interface transformers, it would be necessary also to measure the auxiliary power consumed at each station. It may also be necessary to correct the power loss for ambient temperature.

11-46

11.8 Control Principles and Structures

The control and protection system is an extremely important element of a VSC Transmission scheme. An ideal control and protection system enables the full use of the capability of the equipment in accordance with operator commands and pre-set strategies for transient and dynamic conditions, and it will take action to prevent damage to the equipment in case of failures or unforeseen circumstances. Modern control and protection equipment is close to providing an ideal system.

Many elements of the control and protection system for a VSC Transmission scheme are similar to those used for a LCC HVDC scheme, and their implementation use many identical components. Naturally, some control algorithms are different, e.g. the determination of the turn-on and turn-off instants for the IGBTs and for the thyristors.

This section will describe a typical control and protection structure used for a VSC Transmission scheme, and will describe how some of the control objectives are achieved. The description has been based on an ABB diagram, but other implementations are of course possible[20]. The section also provides a description of a typical protection system for a VSC Transmission scheme. It should be noted that the control and protection systems are not completely separate, as the control system plays a significant part in keeping component stresses within acceptable limits, even during dynamic and transient conditions.

11.8.1 Control Structure

Figure 11-25 shows a simplified single line diagram of one terminal of a VSC Transmission scheme and a typical control structure. The control system includes the following parts:

• I/O Interface

• PQU Order Control

• Current Order Control

• Current Control

• Valve Control

11-47

Figure 11-25 Typical control structure of a VSC Transmission scheme (courtesy of ABB).

The structure of a VSC Transmission system can be different, and the functions may be implemented differently than shown in Figure 11-25. However, all of the functions outlined below must be implemented in the control system. The following descriptions relate to the implementation shown in Figure 11-25.

11.8.1.1 I/O Interface

The I/O Interface takes measurements from the VSC Transmission substation and processes these for use in Current Control. The processing includes Analogue to Digital conversion, filtering and normalization. The data is passed directly to Current Control.

The input data includes:

• AC voltage and current at the point of common coupling,

• AC voltage at the filter bus,

• Positive and negative direct voltage on the dc capacitors,

11-48

• Direct current at the positive and the negative dc terminals.

11.8.1.2 PQU Order Control

The PQU Order Control receives setting information from the operator, internal and external system status information, and data from Current Control describing the operation of the system. PQU Order Control acts directly on the interface transformer tapchanger, to optimize the operating conditions for the VSC. PQU Order Control also provides reference values for the use of Current Order Control, the reference values having been derived to achieve the settings from the system operator, subject to operational limits, e.g. temperature and equipment limitations. PQU Order Control also takes into account any high-level automatic control algorithms, e.g. power system frequency control, damping control, runback control etc, when determining the reference values passed to Current Order Control.

The setting references are the conditions to be achieved at the VSC terminals, and are dc or rms quantities. The data communicated to Current Order Control include:

• Active Power reference

• Reactive Power reference

• Direct Voltage reference

• Block (digital signal to turn off IGBTs)

• Runback (digital signal)

• DC Voltage Control enable (digital signal)

The last three signals over-ride the setting references.

11.8.1.3 Current Order Control

Current Order Control receives the reference values from PQU Order Control and data from Current Control, which gives the following data for the VSC:

• Positive Sequence dq voltage at the ac terminals

• Positive and negative direct voltage at the dc terminals

• Direct current at the positive and negative dc terminals

The signals provided to Current Control are instantaneous values and include:

• d axis ac current reference

• q axis ac current reference

• Block signal

These values are determined by the following control loops:

• PQ2IDQ, which calculates the ac d and q current orders with respect to the provided positive sequence voltage.

11-49

• DCVC, which calculates the current order required to keep the direct voltage at the desired level.

Typically, one VSC Transmission terminal is assigned to steady state control of the active power flow between stations, and the other is assigned to steady state control of the direct voltage. At the station operating in active power control, the direct voltage reference depends on the active power exchange, but during transient or dynamic conditions dc voltage limits may come into force in order to avoid over or under-voltage on the dc side. If the ac voltage falls below the normal operating range for the ac network the control will change from power control to current control.

Current Order Control also includes a number of pre-set characteristics with the objectives of keeping the operation of the VSC within safe limits during dynamic and transient conditions. The objectives are:

• Keeping the dc voltage within safe limits (UDCCOL – dc voltage dependent current order limit).

• Keeping the ac voltage at the filter bus within safe limits (UACCOL – ac voltage dependent current order limit).

• Keeping the device current within the safe operating area.

11.8.1.4 Current Control

Current Control processes the data from the I/O Interface and determines the control instants for switching VSC valves on and off, in order to satisfy the Id and Iq reference values sent by Current Order Control.

The de-composition of the ac voltage measurement into positive and negative voltage sequence components must be done very quickly, in order to provide accurate and independent control of active and reactive power. In order to achieve de-coupling between the active and reactive power, a quasi-positive sequence voltage at the point of common coupling can be derived and used together with the true negative sequence voltage for the ac current control, effectively providing a feed forward loop.

A Phase Locked Loop (PLL) is used to achieve synchronization between the converter voltage and the ac voltage at the point of common coupling. An additional Phase Locked Loop (DQPLL) operates on the positive sequence voltage derived from the ac filter bus voltage, and provides the sampling instants.

The objective of the AC Current Control is to control the current through the converter reactors. Symmetrical three-phase currents are provided by the converter irrespective of whether the ac network voltage is symmetric or not. The current order is calculated from the power order.

The PWM function, which may be a SHEM PWM or OPWM strategy, is used in conjunction with the DQPLL, and determines the switching instants for the VSC Valves.

11-50

11.8.1.5 Firing Control

Firing control checks that it is safe for the VSC Valve to turn on and off, before issuing the control signal to the valve level. The blanking time interval is implemented and checked in the Firing Control. Firing Control also checks that it is safe to turn on the valve, e.g. that the valve redundancy has not been exceeded by IGBT or diode failures. The Firing Control communicates with the valve level by means of a dedicated fiber optic cable. Firing Control also receives status information from the valve level, e.g. healthy/failed status and other information.

11.8.2 Control Strategies for Different Applications

Figure 11-25 shows a number of reference signals as input to PQU Order Control. For some applications some of these signals may in fact be derived automatically based on other local information. The processing of the local information, and the derivation of the reference signals may be done in a separate control block, but can also be derived in the PQU Order Control block.

The following sub-sections describes the higher level controls for a number of different applications. The descriptions are for the steady state and dynamic control of the VSC Transmission scheme. In all cases the control limits within Current Order Control will also be available, and will act as required during abnormal conditions in the ac network in order to:

• Limit the stresses on the converter equipment to within acceptable limits.

• Ensure that normal operation can be resumed as quickly as possible when normal ac conditions are resumed.

11.8.2.1 Power Transmission between weak ac networks

One station of the VSC Transmission scheme will normally be assigned to dc voltage control, whilst the other will be assigned to active power control.

If the dc voltage at the inverter terminal is controlled and the transmission distance is large, it may be desirable to have a slope setting on the reference voltage. This would enable the mean dc voltage to be increased at part load, thereby decreasing the power loss in the cable. When the station in dc voltage control needs to increase the dc voltage, e.g. because it has dropped below the reference value, the station has to increase the active power injected if it is a rectifier, or decrease the power extracted if it is operated as an inverter.

If one of the stations is connected to a relatively weaker ac network than the other, then it may be desirable for that station to be in power control. Similarly, if one ac network has a requirement for frequency and/or damping control, then it may be preferable for that station to be in power control.

If both ac networks are weak it is normally desirable for both stations to be in ac voltage control. The ac voltage setting may have a slope with active power, such that an appropriate ac voltage profile is established in the ac network.

11-51

If one or both terminals are in a strong ac network, the system operators may prefer to operate that terminal or those terminals in reactive power control mode. Again the reactive power reference may have a slope with active power, to suit the particular ac network conditions.

11.8.2.2 Feeding of Passive Network

When a VSC Transmission scheme feeds a passive load, the rectifier (i.e. the terminal in the main ac network) will typically be in dc voltage control and either ac voltage control or reactive power control, dependent on the system operators preference. The PLL will lock to the ac system voltage.

The terminal in the passive network will be in frequency and ac voltage control mode. As the ac system does not directly create its own ac voltage EMF, the PLL operates from its own internal oscillator. If synchronous generator(s) may occasionally operate in the passive ac network, then the control system must be capable of switching between an external ac system voltage and its own internal oscillator. In this case, the VSC Transmission scheme will be in active power control mode. The active power will then typically be determined by a frequency control loop. If the generators supply only a small fraction of the load, it may be difficult to detect their presence, and signalling between the generators and the VSC Transmission scheme may be necessary.

11.8.2.3 Connection of a Wind Farm

The connection of a large, remote wind farm is a special condition of the example described in 11.8.2.2 above. The VSC Transmission scheme may both have to operate as the single power supply to a passive network, e.g. during conditions or either no wind or extreme wind, and as an exporter of power, when wind conditions are suitable for generation. Typically, the mainland terminal will be in dc voltage control and either ac voltage or reactive power control, as desired by the system operator or appropriate to the ac system conditions. The wind farm terminal will typically be in frequency control (with internal oscillator) when the scheme is importing power, and in active power or frequency control (using system frequency or internal oscillator, depending on generator type) when exporting power. The control of the wind generators and the VSC Transmission scheme should be co-ordinated to achieve optimum operation of the overall remote system.

11.8.2.4 Multi-terminal schemes

If the dc capacitors of three or more VSC Transmission terminals are connected in parallel by dc cables, then the scheme will operate as a multi-terminal scheme. Typically, one terminal would be assigned to dc voltage control, with the other terminals in active power control mode. If required, the dc voltage can be controlled with a slope, to optimize the voltage profile on the dc network.

The power input and output must be balanced in order to maintain a stable dc voltage. The terminal operating in dc voltage control will provide or absorb active power, as required to maintain the dc voltage at the desired level. When a rating limit is exceeded, the terminal will no

11-52

longer be able to maintain dc voltage control. Therefore, the other terminals must also have a dc voltage control mode, which can over-ride the active power control, when the dc voltage deviates outside set limits. These separate dc voltage control modes must be co-ordinated, e.g. by the use of different dead-bands, such that a stable and robust control system is achieved.[21].

11.8.3 Performance During and After System Faults

Aspects of the performance of the VSC Transmission scheme during faults in the ac or dc system have already been discussed above in section 11.6. This section focuses on the transient and dynamic responses.

During a fault within the converter or on the dc side, the IGBTs are immediately turned off and a trip signal is sent to the ac circuit breakers at all VSC Transmission terminals. To minimize the stress caused by the over-current, the circuit breakers normally have fast operation and clearance time of less than 3 power frequency cycles. Due to the severity of the fault, an inspection of the converter and the dc system is typically undertaken before the VSC Transmission system is energized again.

During a fault in the ac system the IGBTs may be temporarily turned off, if the ac voltage is very low (typically below 0.2pu), to prevent excessive over-current. The IGBT is turned on again when the current has been controlled down to an acceptable level.

The performance of a VSC Transmission scheme during a remote single phase fault in the ac grid is illustrated Figure 11-26.

During the recovery from a fault in the ac network, the VSC Transmission scheme would in principle be able to increase the power back to the pre-fault level within one power frequency cycle. However, depending on the ac network characteristics, such quick resumption of active power may not be the optimum action, particularly not if the ac network is weak at the point of connection. In such a case, a better strategy may be to give preference to the control of the ac system voltage, with a slower return of the active power to the pre-fault level. The optimum post fault strategy is best determined by system studies.

The traces have been provided by ABB and are for the Cross Sound Cable project (see section 11.12.5 for further detail of this project). The upper trace shows the ac voltage at the point of common coupling (PCC). The fault appears just before time 0 and results in a significant reduction in ac voltage, which lasts until fault clearance at about 75ms. The lower trace shows the transmitted power, which can be seen to reduce during the fault (because of the current limit for the IGBTs), but recovers very quickly after the clearance of the fault. Full power transmission has been restored within 80ms of fault clearance.

11-53

Figure 11-26 Dynamic response to a remote ac fault

During the recovery from a fault in the ac network, the VSC Transmission scheme would in principle be able to increase the power back to the pre-fault level within 1 power frequency cycle. However, depending on the ac network characteristics, such quick resumption of active power may not be the optimum action, particularly not if the ac network is weak at the point of connection. In such a case, a better strategy may be to give preference to the control of the ac system voltage, with a slower return of the active power to the pre-fault level. The optimum post fault strategy is best determined by system studies.

If the VSC Transmission scheme is the sole power supply to the ac network, or the ac link to other parts of the ac network is extremely weak, then the protection in the ac network needs special review. The fault current from the VSC Transmission scheme is typically only 20% of that of a comparable generator. Whilst this reduces the system stress during a fault, and therefore gives more time for fault clearance, some protection systems may not function correctly at these low currents. For example, conventional over-current relays may not be suitable for the protection of ac lines and major equipment, since the fault current would be much smaller than in an ac network powered by conventional generators. The VSC Transmission scheme will normally be arranged to feed as much current into the network as possible, to facilitate the operation of protection in the ac network. However, prolonged operation of the VSC Transmission scheme may not be possible, if the fault is very close such that the fault impedance as seen from the VSC Transmission substation is very low.

11-54

11.8.4 VSC Transmission Protection

The purpose of the VSC Transmission protection system is to detect equipment failures and mal-operation of the converter control. Figure 11-27 shows a typical protection block diagram for a VSC substation.

Figure 11-27 Typical Protection diagram for VSC Transmission substation (Courtesy of ABB)

In the event of a fault the following actions may be taken:

• Temporary blocking (turning off) of the converter.

• Single phase transient current limitation.

• Permanent blocking of the converter.

• Trip of ac circuit breakers.

The action taken depends on the severity and type of the fault. Note that for a VSC blocking action means that an off pulse is sent to the IGBT, thereby stopping the conduction of the IGBT.

11-55

Since a control system failure could be responsible for protection actions, those protection indications that do not require immediate action may first initiate a changeover between the active control channel and the stand-by channel. If the problem disappears after the changeover, the previous control channel is locked out, an alarm is raised and operation continues with the new control channel.

The protection of the interface transformer is basically that of a conventional sub-station transformer:

• Differential protection

• Over-current protection

• Restricted earth fault protection

• Transformer oil & winding temperature, gas detector and oil level

The main ac bus typically uses differential protection and ac circuit breaker failure protection.

The converter bus, i.e. the zone between the converter and the interface transformer, typically has differential and abnormal ac voltage protection. Over-current protection is not typically provided for the phase reactor, since the reactor has considerably higher over current capability than the semi-conductor devices.

The ac harmonic filters have capacitor unbalance and resistor/reactor over-load protection similar to that used on a LCC HVDC scheme. The protection has to take into account the higher frequencies flowing in the filter components, compared with those in a LCC HVDC scheme.

The converter valve protection typically includes:

• Converter over-current protection. This protection uses temporary blocking to limit transient currents, and uses blocking and tripping in the event of persistent over-current

• AC terminal short circuit current protection

• VSC Valve short circuit protection

• High dI/dt IGBT current. This is also used for short circuit current detection

• IGBT monitoring. To ensure that the converter is tripped in the event of more IGBT failures than the number of redundant levels.

The converter dc protection includes abnormal dc voltage protection, which protects the converter from dc over-voltage and unbalanced dc voltage. An under-voltage protection is also included, to protect the VSC valves from operating when the dc voltage is inadequate for the gate unit power supply.

The VSC Transmission scheme typically also has a number of other protections and indications, including VSC Valve cooling protection and alarms, dc cable fault location, auxiliary power supply protection and alarms and fire protection.

11-56

11.9 Installation Considerations

Since a VSC Transmission scheme does not require breaker switched ac harmonic filters, the land areas required for its substations are considerably less than those required for a LCC HVDC scheme. The actual area required depends on the ac voltage level as well as on the power capability and the dc voltage. Typically, a VSC Transmission substation will occupy 40% or less of the space required for a LCC HVDC installation of the same power rating and ac connection voltage.

Because the VSC Transmission substation is relatively compact, it becomes reasonable to enclose all the equipment between the secondary side of the interface transformer and the dc cable connection in a warehouse type building. Enclosing the electrical equipment in a building provides a number of advantages:

• It provides containment for the electromagnetic noise generated by the high frequency switching of the converter. The floor, ceiling and walls include an electric screen, which functions similar to a faraday cage. This is a similar requirement to that for a LCC HVDC scheme, which requires the thyristor valves to be located in an electrically screened valve hall.

• It provides protection against adverse weather conditions. Depending on the voltage used for the dc transmission and the pollution level at the place of installation, the building can be either fully enclosed and pressurized, with a closed cooling system for the valve hall environment, or can be a relatively open building, which allows air to enter and escape freely.

• It provides attenuation for the audible noise generated by the converter operation. The audible noise from the electrical equipment tends to be at higher frequency than for a LCC HVDC scheme, which makes it somewhat easier to attenuate the noise.

• It makes it unnecessary to mount equipment on steel structures, as access can be restricted to the equipment to times when the scheme is out of service and safely earthed down. This means that the overall building height can be kept to a minimum.

The compactness of the VSC Transmission substation is of considerable benefit if it is intended to install one of the terminals off shore (e.g. for a wind farm application, or for the feeding of power to an off shore oil or gas production platform) or in a downtown city area.

The VSC Valves may be mounted in special enclosures at the factory, and shipped to site after extensive testing. The control and protection equipment is similarly housed in a special enclosure, and extensively tested prior to shipment to site. The enclosures are placed on relatively simple foundations within the warehouse type building.

The interface transformer, ac circuit breaker, ac isolator and PLC/RFI are typically located outdoors, as are the cooling fans required for the VSC Valves and other equipment.

In order to satisfy local requirements in terms of audible noise, the overall design of the VSC Transmission substation must take into account all sources of audible noise. The main sources include the interface transformer and the VSC Valve cooling equipment, but attention must also

11-57

be paid to the noise from the ac harmonic filters, HF filters and DC capacitors. As the latter components may be inside a warehouse style building, incorporation of noise attenuation in the roof and walls of the building can reduce the noise from these components to an acceptable level. Additionally, consideration may be given to the orientation of the substation, such that maximum distances are developed between sensitive locations external to the station and the noise producing components.

The use of underground dc cables, which are much less visually intrusive than overhead lines, for the connection of the VSC Transmission substations makes it easier to find suitable locations for the substations close to the desired point of connection.

11.10 Modularity

Whilst a large number of LCC HVDC installations are in operation worldwide, each installation tends to be tailor made for the application. “Standard ratings” and dc voltage levels tend to be brought about because of technical limitations, e.g. 500kVdc is considered a proven technology, which provides reliable and economic transmission at high power. Similarly, 3000MW is considered the maximum economic rating for a bipolar LCC HVDC scheme operating at ±500kVdc, partly because the associated direct current is close to the maximum available current capability for high voltage thyristors, and partly because the impact that a failure of a larger rated scheme would have on the ac network performance. However, for smaller LCC HVDC schemes a large range of direct voltages and converter ratings are in use.

Other issues which makes it more natural to tailor-make a LCC HVDC scheme are the reactive power absorption by the converter, and the generation of relatively large amplitudes of harmonic current at low harmonic orders. The harmonic filters supplied to limit the harmonic interaction with the ac network also serve the dual purpose of providing compensation for the converter’s reactive power absorption, and need to be specially designed for the particular ac network. Additionally, the reactive power banks are a potential source of large ac over-voltages, during fault recovery and malfunction of the converter. Therefore, insulation co-ordination for a LCC HVDC station has to take into account the specific ac network conditions, and the reactive power compensation equipment installed.

Naturally, within the tailor made LCC HVDC solution, each manufacturer uses a number of standard building blocks from which the converter valves, control and protection system, circuit breakers, and other switchgear etc, are constructed. Nevertheless, the engineering of an LCC HVDC scheme can account for up to 15% of the total cost of converter stations with a rating of 500MW, the percentage decreasing as the rating increases.

A VSC Transmission substation does not usually have large switchable reactive power bank, but has the ability to control its reactive power by converter action. Therefore, ac overvoltage issues are less onerous than with a LCC HVDC scheme. Similarly, as the harmonics generated are at relatively high harmonic orders, the interface transformer typically dominates the harmonic impedance as seen from the converter, making the design of the ac harmonic filters easier. The application of a VSC Transmission scheme in an ac network becomes much more like the installation of a generator (or large machine), than the installation of a complex power electronic

11-58

system. Therefore, the likelihood of being able to standardize on solutions is greater for a VSC Transmission, than for a LCC HVDC.

Nevertheless, there are a number of engineering issues that need to be addressed for specific VSC Transmission substation sites, just as they have to be addressed for a LCC HVDC scheme. In addition to determination of the basic design parameters for the converter, converter reactor, interface transformer and ac switchgear, the following issues need to be addressed:

• Harmonics

• Audible Noise

• Radio Frequency Interference

• Power Line Carrier Noise

• Cooling plant design

In many cases the action may be limited to confirmation that the performance is met with the manufacturer’s standard solution, thereby minimizing the design work.

At the end of 2005 ABB was the only manufacturer offering commercial VSC Transmission solutions. ABB had completed 5 commercial VSC Transmission schemes, with a further scheme in progress, in addition to two demonstrators (Helsjon, 3MW and Tjæreborg, 7MW) and a back to back scheme (Eagle Pass, 36MVA). The rating of the 6 commercial schemes were:

• Gotland, 50MW, ±80kVdc

• Direct Link, 3 x 60MW, ±80kVdc

• Murray Link, 200MW, ±150kVdc

• Cross Sound, 330MW, ±150kVdc

• Troll, 2 x 40MW, ±60kVdc

• Estlink, 350MW, ±150kVdc

More information about these schemes will be provided in section 11.12.

ABB documentation describing the HVDC Light® technology [23] sets out 9 standard solutions for VSC Transmission. These use dc voltages of ±80kVdc, ±150kVdc or ±300kVdc and direct current of 627, 1233 or 1881Adc, providing sending end power levels between 101MW and 1140MW. The interface transformer and ac switchgear is designed according to the ac connection voltage.

The advantages of standard solutions include:

• Cost reduction because of reduced engineering

• Lower project risks because of use of proven solutions

• Shorter project delivery times because of reduced engineering and setup times

11-59

• Immediate support from manufacturer in case of problems because a larger knowledge base will be available

For a potential customer the perceived disadvantage of using a standard solution may be that the standard rating is higher than that actually required. However, the cost differential, between the larger standard solution and a tailor made solution may be minimal. Additionally, the larger rating may provide better future proofing of the transmission solution.

11.11 Reliability, Availability and Maintainability

Reliable performance and good availability is achieved by good design practice that takes into account the steady state, dynamic and transient stresses likely to occur during the service life of the equipment, and which incorporate equipment margins where necessary. A protection system, which quickly detects the onset of a fault and takes the correct action to minimize any damage caused, is another essential part of good design. When the VSC Transmission substation is unmanned, as is typically the case, the clear communication and presentation of indications, alarms and protection signals to the system operator is another essential requirement for good availability performance.

The reliability of equipment or sub-systems, which may be subject to random failures of components, can be improved significantly by the use of on-line redundancy. In a VSC Transmission scheme, redundancy is typically provided in the following equipment and sub-systems:

• VSC Valves – additional series connected IGBT levels

• Cooling plant – additional pumps, cooling units and fans

• Control system – complete duplication, one system in hot stand-by

• AC and DC capacitors – additional series and parallel connected elements

• Auxiliary power supplies – two independent supplies and two fully rated systems, each one capable of supplying all station loads, with automatic changeover between the two.

The Availability of a VSC Transmission system can be increased by:

• The provision of adequate spares such that failed equipment can be replaced and the system returned to operation as quickly as possible. Some expensive spare components could be shared between the substations, provided that suitable transport infrastructure exists to quickly move the spare component to the other station, in case of need.

• Availability of well-trained maintenance and repair personnel, who can quickly attend the VSC Transmission substation, thereby cutting down the outage time and reducing unavailability.

• The ability to access the information from Digital Transient Recorders and Sequence of Events Recorder from a remote location, where specialists can interpret the data in order to determine the cause of failure and the appropriate actions to be taken.

11-60

Prior to operation of the VSC Transmission scheme, the only way to determine the Reliability and Availability performance is through statistical calculations. The liquidated damages should reflect the potential loss of revenue resulting from poorer than guaranteed performance, as well as the manner in which the performance is to be measured. Typically, manufacturers will wish to limit their financial risks, e.g. by placing an overall limit on liquidated damages. In addition to any financial damages, consideration could be given to specifying that the manufacturer shall rectify all problems during the Reliability and Availability period, shall report on each problem and, where relevant, shall make design changes to prevent future occurrences.

The maintenance activities for a VSC Transmission scheme are very similar to those required for a LCC HVDC scheme, but can be completed in a shorter time since less equipment is provided, particularly on the ac side of the converter (ac filters and ac switchgear).

Certain on-line redundant equipment in a typical VSC Transmission substation can be maintained with the system live, but the scheme would be at higher risk of trip during such maintenance work. The equipment may include:

• Cooling pumps and fans

• Control and protection system

• Auxiliary power supply equipment (e.g. batteries and battery chargers, and ac switchboards)

Maintenance of other equipment requires a complete shut down of the VCS Transmission substation. Typically, all substations in a VSC Transmission system are maintained simultaneously during the shutdown, to minimize the overall shutdown time. This approach requires a larger maintenance team, than if the substations were maintained sequentially.

During the shutdown of the VSC Transmission system for maintenance, power supply to any island loads must be provided by alternative means. If two parallel VSC Transmission schemes are used then scheduled maintenance work is typically done during periods of low load in the ac network.

As the VSC Transmission technology has now been in service in commercial applications for several years, the early learning experiences gained have been incorporated in the design of later schemes, resulting in continuing reliability and availabilityperformance improvements. Commercial projects are now being quoted with guaranteed Reliability and Availability performance similar to that of LCC HVDC schemes, i.e. in the range of 98 to 99%.

Formal records of the Reliability and Availability performance of have been collected by CIGRE for LCC HVDC schemes for many years [24]. Similar records have not been collected for VSC Transmission schemes. The only published information describing the actual Reliability and Availability performance for VSC Transmission projects identified during the writing of this chapter can be summarized as follows:

• The IGBT failure rate on the Cross Sound Cables scheme during the period 1st July 2004 and 30th June 2005 was 0.38%, compared with the guaranteed rate of 0.5% per annum (Presentation at 8th EPRI FACTS Meeting, Stamford, CT, 17-19 August 2005.)

11-61

• Long Island Power Authority has expressed satisfaction with the availability and reliability of the Cross Sound scheme in a press release issued on 14th July 2005.

• During the first year of operation the scheme transmitted 1,195,895 MWh of electricity to Long Island, a utilization of 41% (LIPA Press statement mentioned above). Apparently, the scheme has seen higher utilization since this report, with the power flow regularly being 330MW in the period between 9am and 10pm. The daily power flow on the scheme can be seen on the New England ISO website.

• There were no emergency or Forced Outage Applications for the Cross Sound project during the first 6 months of 2005 (presentation by New England ISO).

• The Annual Report provided by the Technical Regulator of South Australia [25] shows that the total outage time for the Murray Link scheme during the period 1st July 2003 to 30th June 2004 was 131hrs, equivalent to an availability of 98.5%, which is better than the target availability of 98% for the project.

11.12 VSC Transmission Application Examples

This section provides information about the 7 VSC Transmission systems in service or under construction at the end of 2005. Since ABB is the only supplier to have delivered commercial VSC Transmission schemes, all illustrations in this section have been obtained from ABB, and their kind assistance is appreciated.

11.12.1 Gotland, Sweden

The purpose of the Gotland VSC Transmission scheme [26] was to provide a 70km long transmission link for wind generation at the south of Gotland to the more populated North, from where energy can also be exported to mainland Sweden when the generation exceeds the load requirement on Gotland. Transmission of the power by underground cable was more acceptable to the population than an additional 130kVac overhead line. The VSC Transmission scheme has also resulted in power quality improvements on the island, in particular voltage flicker at the terminals of a large industrial processing plant on the East Coast has been significantly reduced. The power quality of the power from the wind generators is also significantly better than with the use of an ac overhead or cable line connection between the North and the South.

The scheme operates at ±80kVdc and has a rating of 50MW. A simplified single line diagram of the transmission scheme is shown in Figure 11-28. The VSC Transmission scheme operates in parallel with an existing 70kVac line, not shown in the diagram.

11-62

Figure 11-28 Single Line diagram of the Gotland VSC Transmission scheme.

The scheme uses a 2-level converter with triangular carrier PWM switching at 1950Hz. Both terminals use ac voltage control, and are capable of either dc voltage control or active power control. Normally, dc voltage control is at the inverter end. The scheme operates unmanned and is controlled from the control room on Gotland. The scheme does not require telecommunication between the terminals for its operation.

Figure 11-29 shows the VSC Transmission substation at Bäcks, near Visby in the North of the island. The VSC Valve cooling plant can be seen on the side of the building.

Figure 11-29 The VSC Transmission substation at Bäcks.

The Gotland scheme entered service in 1999. The scheme is reported to have met all the specified requirements in terms of power transmission capability, and to have provided significant improvement of power quality in the Northern part of the ac network.

11-63

11.12.2 Tjæreborg, Denmark

The Tjæreborg VSC Transmission scheme is a wind power transmission demonstration project, built by ELTRA, Denmark in anticipation of the construction of several large offshore wind farms [27]. The scheme has a relatively small rating of 7MW, and operates at ±9kVdc with a dc cable length of only 4.5km. However, the low rating was considered sufficient for the purpose of investigating the benefits which could be derived from the use of a variable frequency supply to test a number of different wind generators.

A simplified single line diagram of the transmission scheme is shown in Figure 11-30. The VSC Transmission scheme operates in parallel with an existing 10.5kVac line as shown in the diagram.

Figure 11-30 Single Line diagram of the Tjæreborg VSC Transmission scheme.

The scheme uses a 2-level converter with triangular carrier PWM switching at 1950Hz. Both terminals use ac voltage control, and are capable of either dc voltage control or active power control. Normally, dc voltage control is at the inverter end. The interface transformers are dry type transformers. The scheme operates unmanned and is controlled from the control room at Tjæreborg Enge. The scheme does not require telecommunication between the terminals for its operation.

Figure 11-31 shows the VSC Transmission substation at Tjæreborg Enge.

11-64

Figure 11-31 The VSC Transmission substation at Tjæreborg Enge

The Tjæreborg scheme entered service in 2000. The scheme has been used as intended, and a number of papers have been published to present the results of the studies[28]. There have been no reports of any major problems.

11.12.3 Direct Link, Australia

The Direct Link VSC Transmission scheme was built to provide a non-regulated power link for power trading between New South Wales and Queensland Australia [29]. The use of underground cables enabled planning permission to be obtained much earlier than a competing regulated ac overhead line, thereby enabling opportunity to be taken from a significant energy price differential between the two states. The scheme consists of 3 identical parallel VSC Transmission schemes, providing a total transmission capability of 180MW, 195MVA. The ac connection points in both states are relatively weak, making the controllability of reactive power particularly useful in this application. Revenues can be earned not only when transmitting active power, but also by providing a reactive power or ac voltage control service. The use of virtually the same converter rating as for the Gotland scheme meant that the equipment could be manufactured very quickly. The three cable systems operate at ±80kVdc and have a dc cable length of 65km.

A simplified single line diagram of the transmission scheme is shown in Figure 11-32.

11-65

Figure 11-32 Single Line diagram of the Direct Link VSC Transmission scheme.

The scheme uses 2-level converters with triangular carrier PWM switching at 1950Hz. Both terminals can use ac voltage or reactive power control, and are capable of either dc voltage control or active power control. Normally, dc voltage control is at the inverter end. The scheme operates unmanned and is controlled from the grid control center. The scheme does not require telecommunication between the terminals for its operation. Figure 11-33 shows the VSC Transmission substation at Mullumbimby.

Figure 11-33 The VSC Transmission substation at Mullumbimby

11-66

The Direct Link scheme entered service in 2000. Soon after entering service the scheme suffered numerous failures of the joints between cable segments. The faults were traced to a manufacturing error, which necessitated gradual replacement of a large number of cable joints.

As a non-regulated trading link the Direct Link VSC Transmission scheme can be used in many different ways, including bulk power transmission and power balancing between networks. Figure 11-34 shows the variation of active power flow during one summer week, where rapid variation of the power between import and export can be seen. For a LCC HVDC scheme such rapid and frequent changes in power flow would have necessitated numerous switching operations of the ac harmonic filters, reducing the time interval between maintenance and adjustment of the breakers substantially.

Figure 11-34 Power flow on the Direct Link scheme during one week in summer

Apart from the cable joint failures, the scheme has been reported to meet all requirements and there have been no reports of any major problems.

11.12.4 Murray Link

The Murray Link VSC Transmission scheme was the second non-regulated power link for power trading in Australia, and provides a link between the power networks in the states of South Australia and Victoria [30]. A simplified single line diagram of the transmission scheme is shown in Figure 11-35.

11-67

Figure 11-35 Single Line diagram of the Murray Link VSC Transmission scheme

As for the Direct Link project, the use of underground cables enabled planning permission relatively quickly. The scheme is rated at 200MW, ±150kVdc and the cable length is 180km, making this the longest overland cable scheme in the world. The ac network connection points at both ends of the scheme are relatively weak. As with the Direct Link scheme, revenue can be earned not only when transmitting active power, but also by providing a reactive power or ac voltage control service.

This scheme marked the second generation of VSC Transmission schemes using a transmission voltage of ±150kVdc as well as 3-level Neutral Point Clamped converters with triangular carrier PWM switching at 1350Hz. The lower PWM frequency and the use of 3-level converters resulted in a significant decrease in the power loss, compared with the earlier VSC Transmission schemes. Figure 11-36 shows the VSC Transmission substation at Berri.

Figure 11-36 The VSC Transmission substation at Berri.

Both terminals can use ac voltage or reactive power control, and are capable of either dc voltage control or active power control. Normally, dc voltage control is at the inverter end. The scheme operates unmanned and is controlled from the grid control center. The scheme does not require telecommunication between the terminals for its operation.

11-68

The Murray Link scheme entered service in 2002. There have been no reports of any major problems.

11.12.5 Cross Sound, USA

The Cross Sound VSC Transmission scheme provides a non-regulated energy trading link between New Haven, Connecticut and Shoreham on Long Island, Massachusetts. A 40km submarine cable across the Long Island Sound connects the two terminals. The scheme was designed and built almost at the same time as the Murray Link, and uses the same technology as that scheme, operating at ±150kVdc, but with a higher power rating of 330MW, 346MVA [31]. A simplified single line diagram of the transmission scheme is shown in Figure 11-37.

Figure 11-37 Single Line diagram of the Cross Sound VSC Transmission scheme

The submarine cables and fiber-optic cables were bundled together to minimize the impact on protected shellfish. The cable bundle was buried 6 feet below the seabed, to achieve protection against damage from fishing vessels and anchors, and to minimize the long-term impact on shellfish living on the seabed. Hard rock below the seabed made it impossible to achieve a burial depth of 6 feet in some areas, and environmentalists objected to the potential impact of the cables on creatures on the seabed. Therefore, it was not possible to enter commercial service on conclusion of commissioning in 2002. The dispute over the environmental issue resulting from the lower burial depth of the cable continued until the Big Black Out in North East USA on the 14th August 2003, when the Cross Sound scheme played a major role in restoring power supply to Long Island. The scheme continued in service until April 2004, when transmission again had to be stopped due to the environmental protests. A settlement was reached in June 2004, resulting in immediate re-energisation and operation of the link.

This scheme use second generation VSC Transmission technology with 3-level neutral Point Clamped converters with triangular carrier PWM switching at 1260Hz. The lower PWM frequency and the use of 3-level converters resulted in a significant decrease in the power loss, compared with the first generation VSC Transmission schemes. Figure 11-38 shows the VSC Transmission substation at Shoreham. The interface transformers can be seen in the lower left corner, RFI and PLC filters can be seen between the transformers and the cooling units for the VSC Valves. The building behind the coolers houses the ac harmonic filters, and the larger

11-69

building in the lower right hand corner houses the phase reactors, VSC Valves, dc smoothing reactors etc.

Figure 11-38 Cross Sound VSC Transmission substation at Shoreham.

Both terminals can use ac voltage or reactive power control, and are capable of either dc voltage control or active power control. Normally, dc voltage control is at the inverter end. The scheme operates unmanned and is controlled from the grid control center. The scheme does not require telecommunication between the terminals for its operation. There have been no reports of any major problems.

11.12.6 Troll, Norway

The Troll VSC Transmission scheme provides the first HVDC power connection to an offshore gas production platform[9]. The Troll platform has 2 large compressors to enable more gas to be extracted from the sub-sea reservoirs, and to transport the gas through the pipeline to the processing station on mainland Norway. As the gas reservoir is being emptied, more and more power is needed for the extraction, and it was decided to import the power from the mainland grid, rather than to provide large on-platform generators. The main reasons for the choice were financial, e.g., because of lower taxes due to reduced CO2 emissions and lower maintenance and operation costs.

Two parallel systems are being provided, each rated at 40MW and operating at ±60kVdc. A simplified single line diagram of one of the transmission schemes is shown in Figure 11-39. The cable distance is 70km, and each scheme uses two cables. The VSC Transmission substation at Kolness uses an interface transformer for the connection to the converter to the mainland 132kV ac network. The converter on the platform connects directly through the phase reactor to a high voltage motor, which is also supplied by ABB. The high voltage motor drives the pre-compressor.

11-70

Figure 11-39 Single Line diagram of one of the Troll VSC Transmission schemes

This scheme uses first generation VSC Transmission technology with 2-level converters with triangular carrier PWM switching at 1950Hz. The main reason for this choice was the smaller space and lower weight occupied when compared with a 3-level NPC converter, space and weight being an extremely significant consideration for an offshore installation. The safety requirements placed on equipment in a petrochemical environment meant that many of the normal protection and control principles had to be revised. In particular, continuity of compression was less important than safety, resulting in the normal temporary blocking actions being redesigned to provide block and trip actions.

The offshore environment also placed stringent requirement on the design of equipment. Virtually all the equipment is housed within an enclosure, which was constructed on the mainland and shipped to the platform, where it was raised into position on the reconstructed platform using a large crane. Figure 11-40 shows the enclosure in position on the Troll A platform. The enclosure houses the two separate VSC Transmission substations.

11-71

Figure 11-40 Troll A VSC Transmission enclosure.

The mainland terminal is in dc voltage control and can use ac voltage or reactive power control, as required by the network operator. The offshore converter is controlled to function as a varible speed drive, capable of providing a frequency between 0 and 63Hz and an ac voltage between 0 and 56kVac. The scheme operates unmanned and is normally controlled from the onshore control center. The scheme does not require telecommunication between the terminals for its operation.

The enclosure was installed in 2004, and the first scheme was commissioned late 2004/early 2005, the second following during the spring and summer of 2005. The scheme entered service in the autumn of 2005, after extensive testing. There have been no reports of any major problems.

11.12.7 Estlink, Estonia-Finland

The Estlink project was ordered early 2005, and will provide an interconnection between Estonia and Finland, crossing the Gulf of Finland. The primary objective of the link will be to provide the Nordic electricity market with electricity generated in the Baltic States. The European Union

11-72

has approved the link as a non-regulated trading link. The rating of the interconnector is 350MW, ±150kVdc.

The VSC Transmission substation will be located at Harku, near Tallinn in Estonia and Espoo near Helsinki in Finland. The cable between the two VSC Transmission substations will include 74km submarine cable and 31km land cable. No information is yet available describing the implementation of the VSC Transmission substations.

11.13 Acknowledgement

The author acknowledges with thanks ABB’s kind permission to use some illustrations from their book, It’s Time to Connect, and other publications.


11.14.1 General and Applications

[1] “VSC Transmission”, CIGRE Working Group B4.37, CIGRE Brochure No 269, April 2005.

[2] “Feasibility of Gate Turnoff Thyristor in a High Voltage Direct Current Transmission System.” EPRI Research Project 2443-5, EL5332, Electric Power Research Institute (EPRI), Palo Alto, California , August 1987.

[3] “Power Electronics: Converters, Applications and Design.” Mohan, N., Undeland, T.M., Robbins, W.P., John Wiley & Sons, Hoboken, NJ, 2003.

[4] “Power Semiconductors in Transmission and Distribution Applications.” Chokhawala, R., Danielsson, B., Ängquist, L. Proceedings of the 2001 International Symposium on Power Semiconductor Devices & ICs (ISPSD), Osaka, Japan, pp. 3-10.

[5] “Power system stability benefits with VSC DC-transmission systems.” Johansson SG, Asplund G, Jansson E & Rudervall R, CIGRE 2004 Session, Paper B4-204.

[6] "Co-ordination of Parallel AC-DC Systems for Optimum Performance." Castro, A.D., Ellström, R., Häffner, Y.J., Liljegren, C., Power Delivery Conference, Madrid, September 1999.

[7] “Enhancing of power quality and availability in distribution systems by means of Voltage Source Converters.” Grunbaum, R., IEE 16th International Conference and Exhibition on Electricity Distribution, CIRED2001, Conference Publication No. 482, Amsterdam, June 2001.

11-73

[8] “Cost-efficient XLPE cable system solutions.” Karlstrand, J., Bergman, G., Jönsson, H-Å., IEE 7th International Conference on AC-DC Power Transmission, Conference Publication No. 485, pp. 33-38, November 2001.

[9] “ New Application of Voltage source converter (VSC) HVDC to be installed on the gas platform Troll A”, Hyttinen, M., Lamell, J.O., Nestli, T.F., CIGRE session paper B4-210, Paris, 2004.

[10] “The challenge of integrating large-scale offshore wind farms into power systems”. Belhomme, R., Joncquel, E., Søbrink, K., Abildgaard, H., Woodford, D., CIGRE session paper B4-204, Paris, 2002.

[11] “HVDC Light experiences for power transmission from offshore wind power parks”. Eriksson, K., Liljegren, C., K. Søbrink, K., ASME/AIAA, Wind Energy Symposium, Reno NV, USA, January 2004.

11.14.2 Implementation of VSC Transmission schemes

[12] “Topologies for VSC Transmission”. Andersen B.R., Xu L., Wong K.T.G., 7th International Conference on AC-DC Power Transmission, IEE Conf. Publ. No.485, pp. 298-304, IEE 2001.

[13] “HVDC Transmission Systems using Voltage Sourced Converters - Design and Applications”. Schettler F., Huang H., Christl N., IEEE PES Summer Meeting 2000, Seattle, Washington, USA,

[14] “PWM and Control of Two and Three-level High Power Voltage Source Converters.” Lindberg, A., ISSN-1100.1616. TRITA-EHE 9501, Royal Institute of Technology, Stockholm, Sweden. 1995.

[15] “Programmed PWM Techniques to eliminate harmonics: a critical evaluation.” Enjeti, P.N., Ziogas, P.D., Lindsay, J.F., IEEE Transactions on Industry Applications. Vol. 26, No. 2, pp. 302-316, March/April 1990.

[16] “Recommended practices and requirements for harmonic control in electrical power systems”, IEEE Standard 519-1992 (replacing IEEE Standard 519-1981).

[17] “Connection of harmonic producing installations in AC high voltage networks with particular reference to HVDC. Guide for limiting interference caused by harmonic currents with special attention for telecommunication systems.” JTF 02 of CIGRÉ WG 14.03 and CCO2 (CIGRÉ 36.05/CIRED 2), Electra reference No. 159, 1995.

[18] “Guide to the Specification and Design Evaluation of AC Filters for HVDC Schemes.” CIGRE WG 14.30, CIGRE Publication No. 139.

11-74

[19] “Cost-efficient XLPE cable system solutions.” Karlstrand, J., Bergman, G., Jönsson, H-Å., IEE 7th International Conference on AC-DC Power Transmission, Conference Publication No. 485, pp. 33-38, November 2001.

[20] “A control system for HVDC transmission by voltage sourced converters.” , Nakajima, T., Irokawa, S., Proceedings, IEEE Power Engineering Society Summer Meeting 1999, Vol. 2, pp. 1113-1119, 1999.

[21] “DC Overvoltage control during loss of converter in Multi-terminal Voltage Sourced Converter Based HVDC (M-VSC-HVDC)”, Lu W, Ooi B.T., IEEE Transactions on Power Delivery.

[22] “VSC transmission operating under unbalanced AC conditions - analysis and control design”. Xu L., Andersen B.R., Cartwright P.J., IEEE Transactions on Power Delivery, Volume 20, Issue 1, Jan 2005.

[23] “Its Time to Connect, Technical description of HVDC Light® technology,” ABB, available from www.abb.com/hvdc.

[24] A Survey of the Reliability of HVDC Systems throughout the World during 2001-2002, Vancers I., Christofersen D.J., Leirbukt A., Bennett M.G., CIGRE WG B4.04, CIGRE 2004 Session, paper B4-201.

[25] Annual Report by the Technical Regulator 2003/04, Electricity, Government of South Australia, available from www.technicalregulator.sa.gov.au

11.14.3 Description of VSC Transmission schemes

[26] “Gotland HVDC Light Transmission - worlds first commercial small scale dc transmission.” Axelsson U., Holm A., Liljegren C., Eriksson K., Weimers L., CIRED, Nice, France, 1999.

[27] “DC Feeder for Connection of a Wind Farm”, Søbrink K., Sørensen P.L., Christensen P., Andersen N., Eriksson K., CIGRÉ Symposium Paper 500-06, Kuala Lumpur, Malaysia, 1999.

[28] “Development and testing of ride-through capability solutions for a wind turbine with doubly fed induction generator using VSC transmission”. Eek, J., Pedersen, K.O.H., Søbrink, K., CIGRE session paper No. B4-302, Paris, 2004.

[29] “The Direct Link VSC-based HVDC project and its commissioning”, Railing B.D., Moreau G., Wasborg J., Stanley D., Miller J.J., Jiang-Hafner Y., CIGRÉ session paper No B4-108, Paris, 2002.

[30] “Murray Link, The Longest Underground HVDC Cable in the World. Mattsson W.I, Ericson A., Railing B.D., Miller J.J., Williams B., Moreau G., Clarke C.D., CIGRÉ session paper No B4-103, Paris, 2004.

http://www.abb.com/hvdc

11-75

[31] “Cross Sound Cable Project, second Generation VSC Technology or HVDC”, Railing, B.D., Miller, J.J., Steckley, P., Moreau, G., Bard, P., Ronstrom, L., Lindberg, J., CIGRÉ session paper No B4-102 , 2004.

12-1

12 AC TO DC CIRCUIT CONVERSION (DRAFT)

12.1 Background

Ever since commissioning of the first successful commercial high voltage dc (HVDC) link from Sweden to the Island of Gotland in 1954, engineers have speculated over the conversion of ac circuits to dc. Though obviously feasible technically, the idea was easily dismissed on the basis of cost. One had to buy converter capacity equal to the full rating of the converted circuit while gaining only incremental capability over the ac rating. Furthermore, operating an ac circuit with dc meant idling one phase position—a third of the circuit’s investment. Even if those barriers were overcome, making dc replicate the synchronizing function served by ac circuits would have demanded diagnostic accuracy and speed for which the industry was then ill prepared.

Speculation over the prospect of ac to dc conversion increased as dc became more widely applied. [1, 2] In 1998 a 138 kV circuit in India was actually converted to dc operation, though primarily as a test bed to demonstrate domestic dc equipment manufacturing capability. [3] Past failures to justify conversion notwithstanding, recent changes have greatly enhanced the feasibility of conversion.

a. System Context

Until the mid 1970s, transmission system growth closely paralleled growth in generation and in load. In the mid 1970’s environmental sensitivity made transmission expansion increasingly difficult and expensive—in some areas virtually impossible. While load in the US and other developed countries continued to grow, transmission investment declined steadily and, in a deregulated environment, congestion-based revenues increased. [4]. In that climate it is not surprising that transmission research focused heavily on two areas: (a) The introduction of new equipment, much of it based on rapidly evolving power-electronics technology, to allow greater pre- and post-contingency loading of existing transmission assets considering both n-1 constraints and each circuit’s thermal capability and (b) means to extend that thermal capability, e.g. by dynamic conductor ratings, sag monitoring, and sag mitigation. While both measures are critical to better utilization of existing systems, their potential for enhancing flows is limited and carries with it a disproportionate increase in the cost of system losses. [5]

b. Transmission Costs

In many parts of the US the cost of new circuit construction, where construction is possible at all, has risen to unprecedented levels. Permitting, now time-consuming and uncertain, has become a major element of that cost. Higher costs for new circuits increase the viability of any option that will make existing circuits work harder.

12-2

c. Diagnostics, Communication and Control

DC’s full advantage as an integral part of an ac network can be achieved largely by use of information available at sending and receiving terminals. Although a circuit which has been converted to dc can be made to mimic the ac function it displaces, it can do substantially better than that, both by virtue of its unique overload capability, redundancy, and controllability as well as its ability to draw on the same diagnostic logic and breadth of data inputs as do FACTS devices. Growing acceptance of remedial action schemes (RAS), Special Protection Schemes (SPS), FACTS Controllers and the ultimate prospect of Wide Area Managements Systems (WAMS) to facilitate a “smart grid” operation will all improve the case for ac to dc line conversion.

d. Regulation and Planning Standards

Reliability standards, derived largely from ac system operating experience, have been slow to adapt to dc’s capabilities and characteristics. It was only recently that NERC and most other planning groups accepted the fact that ac system n-1 criteria, which presumed total loss of an ac circuit, should apply not to an entire dc circuit but to one of its poles; the surviving pole(s) assumed to remaining in operation. [6] The same standard accepts continuously acting control capability as a means to accommodate a first contingency event (e.g. dc controls) while disallowing post-contingency switching schemes as a means of n-1 accommodation

e. Introduction of Tripole DC

Prior to 2004, conversion of single circuit ac transmission circuits to dc meant that one conductor was left idle, presumably to serve as an emergency return path. This reduced by one third the potential “per phase position” advantage of dc conversion. Double circuit lines, having an even number of phase positions, could be converted either to one bipole, each circuit serving as a pole, or three bipoles, each using two conductor positions. The former scheme is often awkward since many double-circuit lines serve intermediate loads, making conversion of just one of the circuits to dc a more attractive option. The latter suffers the same disadvantage, plus the added cost of subdividing the total MW rating into three dc bipoles.

At CIGRE 2004 a scheme was introduced which makes use of the full thermal capability of all three phase positions of an ac circuit, and does so without introducing ground current. This “tripole” scheme, described in detail in the following paragraphs, changes the economics of conversion and brings its prospect much closer to economic feasibility. [7]

12.2 Characteristics of Tripole DC

Figure 12-1 shows a bipole and monopole systems fed from the same bus and supplying a common receiving-end bus. However the earth return current normally associated with a monopole is eliminated by two modifications - both using standard dc system components: (a) The monopole is equipped with an additional bridge connected in parallel with the first but in the reverse direction and (b) all thyristors and their heat sinks are rated higher than the average

12-3

power level per pole. Transformers and other station equipment are standard, both in design and rating. [8]

Pole 3

Pole 1

Pole 2

A B

_

+

+/_

Figure 12-1 Parallel bipole and monopole systems configured to eliminate earth return current

Poles 1 and 2 of the tripole configuration of fig. 12-1 are unidirectional but of opposite polarity. Pole 3 is capable of reversing both voltage and current. At intermediate levels of power the three poles would operate together as a composite bipole in which pole 1 carries full positive current and poles 2 and 3 share the negative return current. That mode of operation will minimize transmission losses. As power is increased, pole 1 will reach its thermal rating first - the rating which poles 1 and 2 could sustain by themselves as a normal bipole. That power level will correspond to 1.0 pu current and 1.0 pu power in subsequent analyses.

Power can be increased above 1.0 pu, sustaining the split return operation described above, by periodically interchanging high and low current duty as shown in fig. 12-2. In that figure Pole 3 periodically relieves a portion of the current carried by pole 2 - then a portion of the current carried by pole 1. In that way the ratio Imax/Imin = 2 can be sustained while power is increased until poles 1 and 2 reach full thermal loading, i.e.:

0.15.1 222max =+I and 265.1max =I (12-1)

When the current magnitude in pole 1 is 1.265, it is half that value in poles 2 and 3. Total pu power is then 2x1.265 compared to 2.0 in the bipole case or 1.265 times the bipole level. It can be shown that it if the ratio Imax/Imin is then increased from 2.0 to 3.73, power is equally divided among all three poles and the total is 1.37 times the level achievable with just two conductors and a bipole converter. Ideally one would expect three poles to yield 1.5 times the power on two poles but the form factor of the tripole current is less than 1.0; hence the ratio 1.37:1.

The period of the current cycle in fig. 12-2 should accommodate the thermal time constant of the line conductors. If made the order of four or five minutes, conductor temperature excursions will normally be a few degrees C above and below normal - roughly the same as would be seen by normal changes in wind or cloud cover.

12-4

Imax

Imin

V =+1

- Imin

- ImaxV = - 1

0

0

0

- (Imax-Imin)

(Imax-Imin)

V = - 1

V =+1

Pole 1

Pole 2

Pole 3

T

I3 = 0

Figure 12-2 Current-time diagram of poles 1, 2, and 3

The inset in fig. 12-2 shows that current ramps can be modest in slope, e.g. one or two seconds and that ample time can be allowed to reverse voltage on pole 3 – all with continuous total power.

Fig. 12-3 shows a detailed PSCAD simulation of voltage and current on all three poles during pole 3 reversal of voltage and current. Fig. 12-4, from the same simulation, shows total tripole power and the ac voltage at the receiving terminal. The interval where voltage flicker might be of concern is circled. However for the case simulated flicker was shown to be within tolerances specified by IEEE standard 141 for voltage dip magnitude and frequency of dip occurrences.

The tripole configuration differs from the bipole configuration in several other respects:

a. Overload Characteristics

DC converter stations are specified to have a short term (e.g. 30 minute) overload rating. Modest overload ratings are inexpensive and impose a minimal loss-of-life on terminal equipment. Higher overload ratings are achieved with additional cooling provisions.

12-5

Figure 12-3 PSCAD simulation of tripole current and voltage transitions

Figure 12-4 Total power and ac voltage during tripole transition

Fig. 12-5 assumes that both bipole and tripole configurations have an inherent thirty minute overload capability of 15%. Thus in response to a call for emergency support the bipole profile (Black curve) goes from 1.0 pu to 1.15 in a ramp time of, for instance, 1 second. The tripole begins at 1.37 pu and, benefiting from the same overload assumption, has an emergency rating of 1.37x1.15 = 1.57. (Blue curve) The tripole can gain an additional 9% if it brings current form factor to 1.0 by reverting to normal, constant current bipole and monopole operation with short term monopole return current either in the earth or in insulated shield wires. Since that recourse will yield a power rating of 1.5, emergency power increases to 1.5x1.15=1.73 as shown by the red curve in fig. 12-5. The additional emergency pick-up capability will increase the allowable n-1 constrained loading on parallel ac circuits. The ramp time from normal to emergency loading, arbitrarily set at 1.0 in Figure 12-5, could be made faster or slower depending on available reactive support. In any case the response is clearly fast enough to satisfy thermally-based transfer limits. No ac line overheating or excess sag will occur in the second or two involved.

12-6

For new dc circuits where overload capability is important, one should compare a bipole scheme deliberately designed for high overload capability with a tripole scheme of the same overload capability. But for conversion of ac circuits to dc, the conductors themselves are likely to limit usable bipole overload capability.

~ 1 Sec,

PowerBipole

Tripole w/o ground return

Tripole with ground return

Temporary ground or shield wire return

1.15

1.501.73

2.0

1.5

1.0

0.5

0.0

-0.5

-1.0

Pow

er a

nd C

urre

nt -

pu

Figure 12-5 Short term (e.g. 30 minute) overload rating of various dc configurations

b. Internal Redundancy

The continuous rating of an ac transmission circuit converted for conventional bipole operation drops by 50% after loss of a pole – by 43% if a 15% overload capability is assumed. Mechanical switching can restore the original full bipole capability assuming the outage is due to the line, not the converter. While this can be achieved without interruption of power, it is of limited value since (a) the prospect of loosing a bridge will already limit n-1 transfer and (b) switching operations are normally not allowed as a means of satisfying n-1 limits. [9]

If a tripole configuration looses either a bridge or a line conductor, power will, in the worst case, drop momentarily from 1.37 to half value or .685, then ramp to 1.15, assuming 15% overload capability. That excursion will take the order of 1 second during which earth or shield wire return current will flow. This results in a redundancy of 1.15/1.37 = 0.84. In other words the loss of either a conductor or converter pole with the tripole system results in a loss of 16% of the power if the transmission circuit is operating at maximum continuous capacity. If the circuit is operating at 84% of peak capacity or less, no power is lost upon either outage.

If the tripole station is given an overload rating of 37% instead of 15%, its redundancy is 100%. It can then accommodate the loss of a pole or conductor at maximum continuous load without reducing power transfer. Paragraph 12-2a pointed out that this overload capability, for bipole operation of a tripole system, is already inherent in the valves and their cooling systems. Thus achieving 100% redundancy may require no more than additional cooling provisions for transformers. A bipole scheme would need 100% overload rating for all equipment to achieve the same objective.

The high redundancy of a tripole configuration means (a) higher allowable loading on parallel ac where loss of a dc pole is the limiting contingency and (b) a higher allowable dc terminal rating for the same ac system impact upon loss of a pole.

12-7

c. Valve Considerations

Pole 3, the “modulating” pole of a tripole system must reverse voltage and current simultaneously in order that power continues in the same direction. This can be done with commercially available conventional valves connected in back-to-back or with bi-directional valves. The former would have the advantage of commercial readiness and fewer spare parts since all valves will be identical. The latter, while not yet commercially available, have been shown to be achievable with existing thyristor valve technology and to offer some reduction in cost. [10]

Tripole dc will require special attention to mechanical design of thyristor valve housings due to thermal cycling resulting from changing current level. While frequent changes in current level are common in industrial applications, the latter are not designed for the operational life of electric power applications. Where current is modulated only during periods of maximum continuous power operation, as would be the case in minimum loss operation (See paragraph 12-2), thermal cycling would apply to just a fraction of the tripole’s operating time.

d. VSC Systems

VSC-based bipole systems have, within their rating capability, proven to have certain economic and technical advantages over thyristor-based systems. Their ability to generate rather than consume reactive power makes them particularly well suited to ac systems with very low short circuit ratio. Because of simpler filtering requirements, they can be designed for more compact converter terminals. Apart from somewhat higher losses, they have two significant disadvantages, the first of which is the absence of redundancy. Because commutation of either pole involves both poles, loss of either will cause loss of the entire system. Secondly, because of their dependence on diodes, they cannot clear a dc fault without intervention of the circuit breakers on the ac system. This makes them unsuitable for overhead transmission where lightning flashovers are reasonably common.

It has been shown that VSC-based systems can also be configured to operate as a tripole, providing that the third pole is thyristor based and capable of reversing both current and voltage. [11] Fig. 12-6 shows the schematic of such a system.

As with thyristor-based bipoles, power capability is increased by 37% by adding the third pole. In addition the redundancy is increased from 0 to at least 0.84 as discussed in paragraph 12-2b, above. Redundancy in this case requires more careful evaluation. If VSC-based systems are applied to underground or undersea cables, as has been the case to date, redundancy in the cable circuit itself is of little value where all poles are bundled or buried in close proximity. The large majority of outages are due to dig-ins or anchor fouling which, in either case will invariably fault all of the dc poles. The tripole redundancy will apply to terminal outages however, and will

apply to cable outages where cables are deliberately buried some distance apart to provide redundancy equivalent to or better than an overhead circuit.

12-8

Pole 1

Pole 2

Pole 3

ThyristorValves

VSCs

Pole 1

Pole 2

Pole 3

ThyristorValves

VSCs

Figure 12-6 Schematic of a VSC-based configuration modified for tripole operation

Both redundancy and overload ratio can be made substantially higher for VSC-based tripole cable circuits than for overhead lines for another reason - the high temporary overload capacity inherent in cables. The foregoing paragraphs explain that a tripole system, overhead or cable, must be designed for valve and cooling system ratings equivalent to 1.37 times average power. Thus if that system reverts to separate bipole and monopole-earth return systems, valve capacity could support operation at a constant current of 1.37 in each pole, which translates to an overload rating of 1.37x3/1.37x2 = 1.5. This presumes that sufficient additional short term cooling provisions are provided for converter transformers and that all remaining series-path equipment can accommodate that rating. Thyristor-based systems depend on strong reactive power support to achieve that overload rating; VSC systems do not. High overload capability has an important bearing on parallel ac loading as explained later in this chapter.

e. Cost Premium

The tripole system will cost more per kW than a bipole of equal rating since (1) pole 3 requires one extra, reverse-polarity bridge, (2) At full power1.37 pu current will be sustained for some minutes (“continuous” as far as thyristor ratings are concerned) requiring that heat sinks be proportionately oversized1 and (3) its form factor of less than unity results in a 9% reduction in useful MW rating per aggregate rating of station equipment. Initial estimates of this premium are from 15% to 20% of station cost per kW, depending on the extent of permitting and infrastructure costs common to bipole and tripole options.

The above notwithstanding, the tripole configuration will usually result in a lower cost per incremental kW due to the higher dc/ac power ratio achievable with the tripole. That benefit is further increased by higher (n-1)-constrained loading allowed on parallel ac circuits and greater dynamic response capability. The prospect of very high congestion-based revenues may put an additional premium on the capacity advantage of the tripole system.

1 For some ratings that factor may also force a move from 100 mm to 125 mm thyristors

12-9

12.3 Configuration Options for ac to dc Conversion

Fig. 12-7 shows alternate ways in which a bipole can be configured to make use of the three phase positions of an ac circuit. In fig.12-6a, the third conductor is used as a metallic return, allowing 57% redundancy upon loss of either a converter terminal or a transmission line conductor (see paragraph 12-2b). That redundancy, achieved without ground current, could be increased by specifying a higher overload rating for the converters but only to the limit imposed by existing line conductors.

For a transmission line fault, it would also be possible to switch the center conductor to a pole position, without power interruption, thus restoring full capability without overload rating in the converters. Limitations to the usefulness of this option are also cited in paragraph 12-2b.

Fig. 12-7b shows the bipole operating with one conductor carrying full current and the other two sharing the return current. This mode of operation will reduce losses by 25% and will again achieve 57% redundancy on loss of either a bridge or line conductor. However in this case earth current will flow until switching takes place to put one of the conductors into position as a metallic return. As with the metallic return scheme, switching can restore full power for a circuit outage but not for a bridge outage.

Pole 1

Pole 2

+

-

Pole 1

Pole 2

+

-a. Metallic Return b. Split Return

Figure 12-7 Alternative uses of the third conductor in conversion of an ac circuit to a bipole dc configuration

In contrast to the bipole alternatives shown in fig. 12-7, the tripole scheme of fig. 12-1 achieves redundancy without mechanical switching. At any given time the tripole will be operating with two poles of one polarity and the third at another. The worst outage contingency is the one-in-three chance that two poles of the same polarity are left in tact. That implies that pole 3 is one of the surviving poles since poles 1 and 2 are permanently opposite in polarity. The recovery to redundant power level in that case is illustrated in fig. 12-8. As cited in paragraph 12.2b, power drops momentarily from 1.37 to half that value, .685, then recovers to the bipole emergency level of 1.15. Because the system is left momentarily with two poles of the same polarity, it acts as a monopole with ground return current until pole 3 has time to reverse polarity and create a bipole system.

12-10

1.5

1.0

0.5

0.0

-0.5

-1.0

Power

Pow

er a

nd C

urre

nt -

pu

-1.5

Ground Current

~ 1 Sec.

Figure 12-8 Alternative uses of the third conductor in conversion of an ac circuit to a bipole dc configuration

12.4 Power Capability of ac Circuits Converted to dc

a. Power Relationships

In a balance three phase ac system, real power is:

θCosIVP acac lg3= (12-2)

Where Vlg is and Iac are rms line-to-ground voltage and line current respectively and Cos is the ac power factor. If a three phase circuit is converted to bipolar dc using just two conductors and a dc voltage equal to crest line-to-ground ac voltage, the power transmitted for the same conductor heating would be:

dcbipole IVP 22 lg= (12-3)

and the ratio of bipole to ac power will be

ac

dc

ac

bipole

II

CosPP

θ322

= (12-4)

Although maximum allowable dc current may be marginally higher due to lower dc resistance, the advantage to power rating is in the order of 1% and will be neglected here. However the ac operating current, Iac, in (12-4) will be considerably below the conductor’s thermal rating due to ac system limitations; principally those associated with reliability. It will be shown later in this chapter that for converter rating purposes, dc current, Idc, can be assigned a value equal to the conductor’s thermal rating. It will normally operate at a lower value, the difference in MW being picked up by the ac system dispatch, thus achieving the same (n-1)-constrained path loading as would be the case if the terminal were operated at its full rating. If the ac power factor is 0.9 and n-1 rules constrain ac loading to 70% of its thermal rating, Idc/Iac = .7 and (12-4) becomes:

12-11

5.1=ac

bipole

PP

(12-5)

On the same three-conductor system the tripole can achieve 1.37 times the bipole rating, in which case:

1.2=ac

tripole

PP

(12-6)

The actual dc voltage sustainable by an ac circuit converted to dc will usually be higher than ac line-to-ground crest but depends on a variety of factors discussed in the following paragraph.

b. Sustainable dc Voltage

Suspension Insulator Requirements

DC switching surges are relatively modest compared to those on high voltage ac circuits. Furthermore dc voltage control is much closer with dc and disruption due to lightning less serious. On the other hand insulators subject to constant voltage are more prone to accumulate pollution than is the case with ac and the pollution failure mechanism itself is different. Insulators for dc service are specially designed with longer surface “creep” distance per unit length and have a contour that discourages flashover under dc voltage. While in theory it would be possible to use ac insulators with dc if operating voltage were dropped 30% to 40% below ac line-to-ground crest, that move would render most prospective projects uneconomic. DC Insulators capable of accommodating a dc voltage equal to full crest line-to-ground ac voltage are the same length as the ac units they would replace. Insulator change over would normally be done live, with the ac circuit still in service. The cost of re-insulation would not weigh heavily in the economics of conversion feasibility.

If accommodation of longer insulator strings were the only limit to higher dc voltage on a line to be converted from ac, change-over to V-strings or even conversion to dead-end strings at each tower would allow a further increase in allowable dc voltage.

Some increase in allowable insulator length would result from lower mid-span clearance requirements associated with dc. That clearance, set by the National Electric Safety Code, provides for both a fixed clearance and a clearance governed by applied voltage. [12] The latter includes a switching surge factor which would be lower for dc.

Clearance and Air Gaps

For the reasons noted above, clearances and air gaps associated with a particular ac circuit will normally be more than sufficient for dc having the same crest line-to-ground voltage; both because of lower dc switching surge levels and, with intermediate and lower voltage ac circuits, conservative design practices. This relationship is illustrated by fig. 12-9 which illustrates very

12-12

generally, that the dc/ac power boost at intermediate and low voltages can be expected to be much higher, in pu, than at higher voltages. [13]

5

6

7

8

9

10

11

0 100 200 300 400 500 600 700 800

Line Voltage - kV

Rat

io o

f Pha

se S

paci

ng to

Fla

shov

er

Spac

ing

Figure 12-9 Excess gap margins characteristic of intermediate and lower voltage transmission circuits

Gradient and Electric Fields

As with ac, dc voltage must be limited to a magnitude which (1) limits conductor surface gradient to values below those producing excessive corona and (2) limits electric and magnetic fields at ground level to values deemed safe for humans.

Standards governing maximum allowable dc electric field strength at ground level vary from one region to another. [14] CIGRE working group SCB2-05 recommends a maximum of 40 kV/M. [15] Because that field is established both by conductor-determined gradient and by space charge, solutions are best undertaken with the EPRI workstation for dc lines. [16]

Conductor surface gradient also varies considerably from one line design to another. Values typically range from 20 kV/cm to as high as 27.5 kV/cm. Circuits with gradients at the upper end of that range have experienced occasional “anomalous” (unexplained) flashovers, believed by many to result from space charge generation. This suggests a gradient criterion in the range of 23 to 25 kV/cm when selecting maximum dc operating voltage for an ac line converted to dc.

In calculating dc gradients it’s important to note that after conversion, maximum line-to-ground voltage appears as simultaneous and opposite polarities on two of the former phase positions. Thus the maximum phase-to-phase voltage is increased in the ratio 2.0/1.73 = 1.16. That increase is not likely to be limiting from an insulation standpoint since switching surges are reduced by a much greater ratio. The voltage increase between phase positions will cause the surface gradient under dc operation to be higher than with ac, though not proportionately. The actual gradient is a function of the total charge distribution and can be assessed only by gradient calculations specific to the circuit in question.

12-13

Example calculations will illustrate what one might expect. 12-10 shows dimensions of typical 230 kV configurations along with gradient calculations for both pre-conversion ac conditions and post-conversion tripole and bipole operation. In the examples shown, dc voltage is equal to line-to-ground crest ac voltage. All cases assume a 3.0 cm diameter conductor and two 0.95 cm appropriately placed shield wires. Neither assumption is very critical to the conclusions reached.

The upper table in each section of in fig. 12-10 shows, for both ac and dc energization cases, the instantaneous applied voltage producing the maximum surface gradient as well as the gradient on each conductor at that time. The corresponding lower table relates the above maximum values to the maximum pre-conversion ac gradient. In the tripole case the phase position assigned to the modulating pole may be important. The shaded sections in fig. 12-10 represent conductors energized with dc

Column 2 of fig. 12-10a, a delta configuration, shows the tripole dc circuit having a gradient 12.1% greater than the ac circuit while the bipole line produces gradients about the same as the ac line. Changing the position of pole 3 (column 4) does not change the maximum gradient.

In fig 12-10b, a horizontal configuration, tripole conversion increases maximum gradient by 11.2% over the ac case while the bipole leaves it unchanged.

Fig. 12-10c is more complex in that the left hand circuit (A,B,C) of a vertical double circuit line is assumed converted to dc while the right hand circuit (a,b,c) remains in ac service. In this case ac voltage was allowed to vary sinusoidally. Instantaneous ac voltages which correspond to maximum gradient are plotted in the upper table, together with the gradients on other conductors at that instant. The first tripole energization option (Column 2) produces the highest gradient on one of the dc conductors. It is 7.1% above the ac level. The alternative mode tripole energization (column 3) in which one of the outer poles is assumed to be pole 3, produces a higher increase; 15.6%. Bipole energization produces gradients slightly lower than the ac case where the upper and lower conductors are used, and slightly higher where adjacent conductors are used.

Maximum gradient on a particular phase or pole is not a complete indicator of either radio noise or loss, both of which are functions of the gradient pattern on all phases and poles as well as the physical configuration of the circuit itself

Summary

It is apparent from the above that the dc/ac ratio of real power capability of an ac circuit converted to dc will depend primarily on (a) the ac power limit imposed by n-1 and other system-related dispatch restrictions, (b) the ac power factor, (c) the sustainable dc voltage and (d) the dc configuration used. Fig. 12-11, based on equations (12-3) and the corresponding tripole equation, illustrates some reasonable bounds for this conversion ratio both for bipole and tripole configurations assuming an ac power factor of 0.9. Fig. 12-11 relates only to the power carried by the converted circuit. It ignores the effect of dc conversion on the allowable (n-1)-constrained loading of parallel ac circuits. That effect, illustrated in the following section, depends on both overload and redundancy characteristics of the dc scheme used

12-14

b

a 6 m c kV kVp/Cm kV kVp/Cm kV kVp/Cm kV kVp/Cma 187.8 21.1 187.8 23.7 -187.8 -18.5 -187.8 -21.2b -93.9 -10.5 -187.8 -18.5 187.8 23.7 0.0 0.0c -93.9 -10.5 -187.7 -18.5 -187.8 -18.5 187.8 21.2

10 m Percent of maximum ac gradientMax 21.1 23.7 23.7 21.2% 100.0% 112.1% 112.1% 100.4%

6m kV kVp/Cm kV kVp/Cm kV kVp/Cm kV kVp/Cm kV kVp/Cma b c a -93.9 -11.3 187.8 20.1 187.8 20.1 -187.8 -19.3 -187.8 -21.0

b 187.8 21.4 -187.8 -23.8 187.8 -23.8 0.0 0.0 187.8 21.4c -93.9 -11.3 -187.7 20.1 -187.8 20.1 187.8 19.3 0.0 -1.7

10 m Percent of maximum ac gradientMax 21.4 -23.8 -23.8 19.3 21.4% 100.0% 111.5% 111.5% 90.2% 100.0%

A a

4.5 m

B 13 M b kV kVp/Cm kV kVp/Cm kV kVp/Cm kV kVp/Cm kV kVp/CmA 93.9 11.9 -187.8 -16.8 -187.8 -20.0 -187.8 -19.9 -187.8 -22.4

4.5 m B -188 -22.3 -187.8 -18.7 187.8 25.7 0 0.1 187.8 22.2C 93.9 11.9 187.8 23.9 -187.8 -20.9 187.8 20.1 0.0 -1.6

C 11 m c a 93.9 11.9 132.79 14.9 -187.8 -19.8 93.9 12.6 178.6 19.7b -188 -22.3 48.603 6.4 93.9 12.1 -187.8 -22.1 -39.0 -4.7c 93.9 11.9 -181.4 -19.6 93.896 10.1 93.9 11.6 -139.6 -14.7

Percent of maximum ac gradient 10 m Max 22.3 23.9 25.7 -22.1 -22.4

% 100.0% 107.2% 115.6% -99.3% -100.8%

ac tripole tripole bipole

ac tripole tripole bipole1 2 3 4

bipole 5

5bipole

1 2 3 4ac tripole tripole bipole

1 2 3 4

a

b

c

Figure 12-10 Peak conductor gradients for various post-conversion dc energizations

.

12-15

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

1.0 0.9 0.8 0.7 0.6 0.5

Ratio of AC Operating to Maximum MVA

Rat

io o

f DC

MW

to A

C M

W

1.5

1.5 1.0

1.0

Tripole

Bipole

Ratio of dc pole voltage to ac l-gcrest voltage

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

1.0 0.9 0.8 0.7 0.6 0.5


Rat

io o

f DC

MW

to A

C M

W

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

1.0 0.9 0.8 0.7 0.6 0.5


Rat

io o

f DC

MW

to A

C M

W

1.5

1.5 1.0

1.0

Tripole

Bipole

Ratio of dc pole voltage to ac l-gcrest voltage

Figure 12-11 General range of increase achievable through conversion of ac circuits to dc

12.5 The Effect of System Context on Total Path Flow

Both overload capability and redundancy characteristics of a dc configuration will affect the allowable n-1 compliant loading of an ac system operating in parallel with it. That can best be illustrated by several simple examples.

a. Identical Parallel Circuits

The simplest illustration of the influence of system context on total path loading is the case where the nth circuit of n parallel and identical circuits is converted to dc. For that case it can be shown that both

• the ratio of gain in (n-1)-constrained path loading to the gain achieved by adding an additional parallel ac circuit and

• DC Effectiveness (DCE), the ratio of increase in total path loading to dc terminal rating,

are independent of the number of parallel circuits and independent of the relative susceptance of the parallel circuits.

Fig. 12-12 illustrates the first point above by plotting the ratio of gain in path MW achieved by dc conversion of the nth circuit to the gain that would be achieved by construction of an additional similar ac circuit. The ratio is plotted as a function of D, the dc terminal rating in pu of the maximum ac thermal rating. For example if a 230 kV circuit has a maximum thermal limit of 800 MW and could achieve a dc rating of 1,200 MW, D = 1.5. It is apparent from fig. 12-13 that for reasonable dc to ac rating ratios, conversion may achieve at least the same transfer benefit as would construction of another ac circuit; in most cases more than one circuit.

12-16

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

1.0 1.5 2.0 2.5 3.0

DC/AC Rating Ratio - D

DC

Effe

ctiv

enes

s

Tripole

Bipole

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

1 1.5 2.0 2.5 3.0

DC/AC Rating Ratio - D

Conv

ersi

on/N

ew C

ircui

t Rat

io

Tripole

Bipole

Figure 12-12 MW increase achieved by conversion relative to addition of a new circuit

Figure 12-13 DC Effectiveness as a function of relative dc rating

Fig. 12-13 plots DC Effectiveness as a function of D. Both figures 12-12 and 12-13 are somewhat misleading in that a comparison of tripole to bipole alternatives for the same ac circuit should include an increase in D by the factor 1.37, the capability ratio of tripole to bipole systems. A direct comparison is shown in red on both curves for the case where conversion from ac to a bipole would achieve a dc rating 1.25 times the maximum ac rating. For that example DCE would increase from 0.38 to 0.75, possibly making the difference in whether or not conversion is feasible.

Figures 12-12 and 12-13 mask another potential advantage of conversion. Suppose the nth circuit’s ampacity is 20% higher than the parallel circuits. Series compensation of the line as an ac line would increase its loading proportionately while maintaining n-1 compliance. Converting the same circuit to dc multiplies that 20% advantage by D.

Example 1

The four-circuit ac case illustrated in fig. 12-14a has an n-1 compliant transfer of 2,700 MW, assuming a power factor of 0.9. Loading indications in the figure are in the form x/y where x is the maximum pre-contingency loading and y is the maximum thermal rating of the same circuit. In fig. 12-14b one circuit is converted to a bipole having a rating 25% greater than the maximum rating of the ac circuit. The allowable transfer is increased by 537 MW - achieved by adding 1,250 MW bipole converters; a DCE of 537/1,250 = 0.43

The same circuit could be converted to a 1,713 MW tripole system as shown in fig. 12-14c. In this case the increase in flow is 1,259 MW and the DCE is 1,259/1,713 = 0.74. These DCE values are consistent with those shown in figs. 12-12. Note that in this case 21% of the increase in flow is on the ac system, demonstrating the importance of system context in assessing conversion feasibility.

12-17

750 / 1,000750 / 1,000 3,000 MVA750 / 1,000 2,700 MW750 / 1,000

736 / 1,000736 / 1,000 3,458 MVA736 / 1,000 3,237 MW

1,250 / 1,438 DCE=

832 / 1,000832 / 1,000 4,209 MVA832 / 1,000 3,959 MW

1,713 / 2,158 DCE=

0.43

0.74

Bipole

Tripole

a.

b.

c.

Figure 12-14 Example conversion advantage – four identical circuits

Adding additional ac circuits to the system of fig. 12-14a would increase transfer by 900 MW each at the assumed power factor of .9. Thus the bipole conversion option is equivalent to 0.6 new circuits and the tripole option equivalent to 1.4 new circuits.

b. Conversion of an underlying circuit which limits n-1-constrained transfer

The above analysis implies that all circuits have identical susceptance and conductance. It is directly extendable to cases where the converted circuit has higher or lower susceptance (e.g. a lower voltage or higher voltage circuit) and a conductance/susceptance ratio (g/b) other than 1. The incentive for conversion to dc is very large for lower voltage circuits where g<b, i.e. where the prospect of its overloading upon loss of a higher voltage circuit is the n-1 limiting case. Whereas the ac solution to that limitation would be insertion of series reactance or a phase-shifting transformer, restricting flow on the underlying circuit, dc conversion eliminates the bottleneck and increases the capability of that circuit.

Example 2

In this case the underlying circuit is assumed to have a susceptance of 0.2 times the overlying circuit’s (e.g. a 230 kV circuit underlying a 500 kV system) and a conductance 0.16 times that circuit. The initial path flow of 1,581 MW is illustrated in fig. 12-15a, constrained by post-contingency overloading of the underlying circuit. The obvious solution in this case would be addition of series reactance to the underlying circuit, making its susceptance and conductance ratios with respect to the other circuits equal, thereby achieving a large, inexpensive boost in capability as shown in the dispatch of fig. 12-15b.

The bipole conversion alternative is illustrated in fig. 12-15c. Its rating is again assumed to be 1.25 times the maximum thermal rating of the underlying circuit; in this case limited by the lower admittance on that circuit. The conversion boosts the original path loading by 2,030-1581 = 449 MW – more than twice the rating of the dc converter. However it’s more realistic to

12-18

compare the gain achieved by the dc option with the easy ac fix of fig. 12-15b. On that basis the increase is 2,030-1,945 = 85 MW which, achieved with a 200 MW terminal, gives a DCE of .43.

The tripole configuration, illustrated in fig. 12-15d, achieves a total path transfer of 2,145 MW, an increase of 200 MW over the system of fig. 12-15b. In this case DCE is 200/274 = .73, about 70% greater than the bipole case.

549 / 1,000549 / 1,000 1,757 MVA549 / 1,000 1,581 MW110 / 160

684 / 1,000684 / 1,000 2,161 MVA684 / 1,000 1,945 MW109 / 160

678 / 1,000678 / 1,000 2,233 MVA678 / 1,000 2,030 MW200 / 230 DCE=

693 / 1,000693 / 1,000 2,353 MVA693 / 1,000 2,145 MW274 / 345 DCE=

724 / 1,000724 / 1,000 2,362 MVA724 / 1,000 2,145 MW190 / 345 DCE=

0.43

0.73

0.73

Bipole

Tripole

b.

c.

d.

Tripole e.

a.

Figure 12-15 Advantage in converting an example underlying circuit

Figure 12-15e, in which the terminal of fig. 12-15d is operated at a lower level, allows an increase in dispatch of the ac system, bringing the total path transfer to the same level as previously. Figure 12-15e makes three important points:

1. Converting one of several parallel ac circuits to dc can take full advantage of a rating equal to maximum thermal capability of its conductors without the need to operate them at that rating

2. DC allows the dispatcher to allocate load between dc and the ac system to minimize losses

3. The DCE of fig. 12-15d and 12-15e could be increased considerably by purchasing a dc converter lower than 343 MW but specifying an overload capability that would attain the

12-19

432 MW shown. The penalty for doing so would be proportionately less for tripole than bipole terminals because of the inherently higher valve ratings of the tripole.

Each new ac EHV circuit added to the system of fig. 12-15a increases total path transfer by 730 MW. Each intermediate circuit (of full ampacity g=b) would increase transfer by 150 MW. Thus the path flow increase achieved through bipole conversion, 144 MW, would be equivalent to approximately 1 additional underlying circuit or .17 new EHV circuits. The tripole flow enhancement of 288 MW would be equivalent to construction of 1.9 new underlying circuits or 0.4 new EHV circuits.

c. Conversion of an overlying circuit

Suppose that a single EHV circuit overlies a number of intermediate voltage circuits; perhaps because the plan to construct multiple EHV circuits was stalled by permitting issues. Loading on the EHV circuit will be severely limited by prospective overload or reactive shortage in the underlying system upon its loss. Transfer may also be limited where the receiving system has limited means of adjustment to prepare for a second contingency.

If the EHV circuit is converted to dc, its (n-1)-constrained rating will be limited by the lower of: (i) the dc capability of the conductors on the converted circuit and the degree to which they’re used by the dc scheme adopted or (ii) the dc rating beyond which loss of a pole cannot be accommodated by the underlying ac system even if the latter is completely unloaded.

Example 3

Figure 12.16a shows a four circuit example where the total useful transfer is limited to 675 MW by the need to accommodate loss of the EHV circuit; about 40% of the path’s thermal capability. Fig. 12-16b shows the n-1-compliant loading achievable with the EHV circuit converted to a bipole dc scheme. The bipole circuit itself is again assumed capable of carrying 1.25 times the maximum ac MVA rating; 1,250 MW. The dispatch shown in fig. 12-16b represents the loading at which the ac, with a power factor of 0.9, is just capable of absorbing the loss of a dc pole, i.e. 0.43 x 1,250 = 537 MW where 0.43 is the characteristic bipole redundancy discussed in paragraph 12.2b. Bipole conversion increases total path flow by 713 MW with a DCE of 0.57.

If converted to a tripole configuration, the same conductor system would allow a terminal rating of 1,250 x 1.37 = 1,713 MW. The maximum rating from an n-1 standpoint would be that for which 16% equals the MW capability of three unloaded ac circuits, i.e. 675/.16 = 4,219 MW. The lesser of these two values, 1,713 MW, would apply and is shown in fig. 12-16c. In this case the original ac loading is increased by 1,440 MW with a DCE of 0.84.

A dc rating as high as 1,713 would raise questions as to (i) whether n-2 reliability criteria could accommodate the loss of that much power on one structure and (ii) whether the external systems would need or could accommodate what amounts to a 250% increase in power transfer. Lower rating of the tripole system would not change its DCE. Fig. 12-16d selects a dc terminal rating equal to the maximum bipole case. For the same terminal rating the tripole achieves 24% higher (n-1)-compliant transfer than would a bipole configuration.

12-20

107 / 250107 / 250 750 MVA107 / 250 675 MW429 / 1,000

51 / 25051 / 250 1,403 MVA51 / 250 1,388 MW

1,250 / 1,438 DCE=

149 / 250149 / 250 2,160 MVA149 / 250 2,115 MW

1,713 / 2,158 DCE=

176 / 250176 / 250 1,778 MVA176 / 250 1,725 MW

1,250 / 1,575 DCE=

0.57

0.84

0.84

Bipole

Tripole

a.

b.

c.

Tripole d.

Figure 12-16 Advantage in converting an example over-lying circuit

If the ac system itself were expanded, adding each intermediate voltage circuit would increase (n-1)-constrained path flow by 225 MW and each EHV circuit by 900 MW. Thus conversion of the EHV circuit to a bipole configuration would be equivalent to approximately 0.80 new EHV circuits or 3.2 new intermediate voltage circuits. The tripole conversion could increase path capacity by the equivalent of 1.6 EHV circuits or 6.4 intermediate circuits.

12.6 Dynamic Support Capability

The above analysis, confined to thermally limited transfers, treats dc and ac overload capability alike even though ac response is a consequence of Kirchoff’s laws and dc’s is dependent on diagnostic logic and control actions. This distinction is becoming less important as ac systems too become increasingly dependent on dynamic control of system apparatus.

Where power transfer is limited by dynamic considerations, dc links can typically transfer a much higher percentage of their normal rating as synchronizing power than ac and can bring that capacity to bear faster than ac. Thyristor valves normally have a 50% overload capability in the one to two second time frame. Fig. 12-17 illustrates dc’s ability to transfer momentary synchronizing power for several configurations. Both the rate of increase and the magnitude of synchronizing power shown in fig. 12-17 depend on the availability of sufficient transient reactive power and ac voltage support.

12-21

~ 1 Sec,

Power

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0.0

-0.5

-1.0

-1.5

-2.0

Temporary ground or shield wire return current

Pow

er a

nd C

urre

nt -

pu

1.50

3.08

2.05

Bipole

Tripole w/o ground return

Tripole with ground return

Figure 12-17 Advantage in converting an example over-lying circuit

The bipole curve (black) in fig. 12-17 assumes a short term increase of 50% over the nominal rating of 1.0 pu. The tripole system, normally having a capability of 1.37 can supply momentary current of 1.37 x 1.5 = 2.05 without earth or shield wire return current (blue curve).

If momentary earth or shield wire return current is provided for, the tripole system could operate as a parallel bipole and monopole at the full thyristor ratings of 1.37 x 1.5. This would result in a momentary rating of 1.5 x 1.37 x 1.5 = 3.08 as shown in the red curve of fig. 12-17. The shaded red curve shows the return current magnitude for that assumption.

Synchronizing power provided by a dc link will reduce the angular excursion between two synchronous system segments by acting immediately on diagnosis of the need for synchronizing power. AC synchronizing power flows only in proportion to the angular excursion itself, i.e. the growth in angular difference between the systems that occurs over time. This difference is illustrated for a classical first swing stability curve in fig. 12-18.

In fig. 12-18 total power is the sum of the dc power (dark blue) and the sine curve representing ac synchronizing power characteristics. After a fault at angle a, the sending system accelerates with respect to the receiving system until angle b, at which point the fault clears. The area A represents the energy stored in accelerated generators. Without benefit of the dc, the systems would continue to depart in angle according to a new characteristic curve (absent the tripped circuit) shown as the lower sinusoid in 12-18 until area B equals area A, at which time the angle would diminish.

The DC link could, upon diagnosis of the problem and adequate reactive power support, ramp up to 150% of terminal rating very rapidly, thus causing area B to equal area A much sooner and diminishing the maximum angular separation of the systems.

12-22

Pow

er

0 90 180 Angle

B

A

B

Corrective action only in proportion to worsening of problem

Corrective action is virtually immediate

dc power

ac power

a b

Pow

er

0 90 180 Angle

B

A

B



dc power

ac power

Pow

er

0 90 180 Angle

B

A

B



dc power

ac power

a b

Figure 12-18 Early dynamic support capability of dc

The foregoing benefit may be forfeited when the ac fault is close to the dc inverter (receiving-end) bus. In that case commutation failure will delay restoration of the dc for following isolation of the fault.

12.7 Limits of AC System Optimization

In a classical ac system, i.e. one with fixed parameters, changes in effective susceptance of a particular transmission circuit will improve (n-1)-compliant transfer only to the extent that pu susceptance is either (a) made equal to that of the circuits it parallels or (b) bears the same ampacity/susceptance ratio as those circuits; which ever is less. Beyond that point, increases in susceptance will either reallocate flows among circuits or force a reduction in power transfer.

The availability of rapid post-contingency control of series reactance or phase angle can increase pre-contingency dispatch by creating a post-contingency load distribution within the rating of each remaining circuit thus maximizing use of remaining circuits.

While the foregoing relate to system recourses for increasing transfer, progress has also been made in ways to increase a line’s thermal limit, e.g. by ratings which reflect actual weather conditions, by monitoring of sag, and by sag mitigation devices. The latter may achieve increases in thermal ratings in the range of 10% to 30% of conventional maximum ratings and corresponding increases in operating level. In contrast, the effective thermal capability can be doubled or tripled through conversion to dc.

Whether further increases in transfer capability of an already heavily loaded ac system are achieved by system or transmission line recourses, those increases may come at a very high cost in incremental losses. That point is germane to the prospect of ac to dc conversion; an option which improves both system-constrained and circuit-constrained transfer limitations and does so with lower losses.

12-23

12.8 AC vs. DC Losses

Transmission owners’ primary economic incentive is to increase transmission capability; a move that may earn congestion-based revenues. That incentive notwithstanding, selection of alternative means for increasing capability should give careful weighting to electrical losses since their cost may be significant. Calculation of losses in MWhrs/yr is a relatively straightforward engineering matter. Assigning them a net present value (NPV) requires debatable economic assumptions; specifically (a) the discount rate, (b) energy value(s) and (c) write-off period. Values for the latter will differ for each stakeholder in the energy matrix. Economic responsibility for losses varies from one transmission line owner to another, one line to another, and from one regulatory entity to another. The beneficiary(s) of loss reduction will often be parties other than the one positioned give weight to losses in investment decisions. Thus the dollar credit given for reduced loss NPV may, depending on ownership, functional, and regulatory circumstances, be well below its true economic value. The significance of losses in diminishing transmission capability and therefore its value in gaining congestion rent is a value more closely coupled to transmission owners’ incentives, is discussed in paragraph 12.18f., below.

The simplest perspective on loss assessment is that of the economist, i.e. a focus on their real total cost to society – thus finessing the question of who bears their cost burden. That is the perspective taken in succeeding paragraphs.

a. Economic assumptions

Escalating energy prices make losses a more important factor in transmission investment economics than at any time in the history of the electric power industry. It is not uncommon for energy market prices to spike to the high hundreds of dollars per MWhr – often when system losses are at their maximum. Although the US Energy Information Administration (EIA) has predicted a relatively modest increase in energy prices up to the year 2025 [17] a number of authorities disagree and predict steady escalation. [18] In any case losses are sure to become an increasingly important factor both in future investment and operating decisions.

The likelihood that the long range percent escalation in energy costs will exceed prevailing interest rate percentages suggests use of a lower discount rate in electrical loss economics than is applied to investment decision in general. The NPV appropriate to losses of a given transmission line will also depend on its function. Where loading correlates with the system diurnal load cycle, energy values are likely to be very high when losses peak. For circuits which serve mainly as energy transporters (e.g. egress from base load power plants) average energy values are more appropriate.

Table 12-1 shows the net present value (NPV) of losses for a load duration curve having a maximum loss of 1 kW and a loss factor of 0.5 for various (constant) energy costs and discount rates - all for a 30 year time period.2 A discount rate of 2% and a constant energy value of $30 per MWhr yield a NPV of roughly $3,000 per kW of Loss. The same value would result from a

2 The assumption of constant energy cost is pessimistic in most cases since actual energy costs will usually be highest when loading and losses are highest.

12-24

variety of assumptions shown in Table 12-2. $3,000 per kW will be used for loss evaluations in later paragraphs.

Table 12-1 NPV of 1 kW of loss. 30 year write-off, Load Factor = 0.5

Table 12-2 Alternative financial assumptions corresponding to a $3,000 NPV for 1 kW

$30 $60 $90

0% $3,942 $7,884 $11,8261% $3,391 $6,782 $10,1732% $2,943 $5,886 $8,8293% $2,575 $5,151 $7,7264% $2,272 $4,544 $6,8175% $2,020 $4,040 $6,060

Value per MWHr

Dis

coun

t Rat

e PaybackYears $30 $60 $90 $120

30 1.9% 8.0% 12.8% 17.5%20 -1.3% 6.0% 11.8% 16.4%10 -13.0% -2.5% 5.3% 11.5%5 -36.0% -22.0% -12.0% -4.3%

Energy Value

b. Loss increase from extended ac operation

Fig. 12-19 shows loss per mile vs. real power for an example 230 kV transmission circuit using a 1,272 kcmil conductor. Operating rating is presumed limited to 220 MVA or, based on a power factor of 0.85, 187 MW. The solid curve extends up to that limit, the dotted curve up to the 90o C thermal limit, presumed to be 1,000 amperes or 400 MVA (340 MW). Extension into the dotted area might be achieved by capacitive compensation, a phase shifting device, a FACTS device to relieve 230 kV overload upon loss of another circuit, or simply by accommodation of higher loading by real time temperature measurement or sag mitigation measures.

Also shown in fig. 12-19 are the loss characteristics of the same circuit converted for dc operation at +/- 220 kV; about 17% above 230 kV line-to-ground crest voltage. DC losses include converter losses estimated at 0.85% of transmitted power per terminal. On that basis the bipole limit at 1,000 amperes is 440 MW and the tripole 1.37 times that or 603 MW; 2.4 and 3.2 times the prior ac MW loading respectively.

Suppose that the dc alternatives in fig. 12-19 were to supply the same maximum power as the ac circuit and have the same load duration curve. It is then a simple matter to compare the NPV of ac and dc losses for each power level up to the 400 MVA (340 MW) maximum rating of the conductor. Fig 12-20 shows that cost for 50,100 and 200 mile lengths, assuming a NPV of $3,000 per kW of loss (per table 12-1). Increasing power from 150 to 200 MW in the 100 mile case would increase loss NPV from $100 per kW transmitted ($15 Million for 150 MW) to $130 per kW transmitted ($26 million for 200 MW), an incremental loss cost of $11 Million or a loss-based cost of the extra transfer of $11x106/50x103 = $220/kW.

Fig. 12-21 shows the incremental loss relationship as a function of real power for 50, 100 and 200 miles lines where the reference point is the assumed 187 MW (220 MVA) operation. Incremental cost is double the absolute cost or, for the increment cited above, $440/kW.

12-25

0

25

50

75

100

125

150

175

200

225

250

0 50 100 150 200 250 300 350 400 450 500 550 600 650

Real Power Transfer - MW

Tran

smis

sion

Lin

e Lo

sses

- kW

/Mile

AC

AC - Extended

Operating Limit

Thermal Limit

Bipole Limit

Tripole Limit

Voltage 230 kVImpedance .08 +j.7 ohms/mileConductor 1,272 kcmilMax I (90oC) 1,000 amperesOp. Rating 220 MVA

Figure 12-19 Loss for various modes of transmission on an example 230 kV circuit

$0

$50

$100

$150

$200

$250

$300

$350

$400

$450

0 50 100 150 200 250 300 350

Maximum Real Power - MW

NPV

of L

osse

s - $

/kW

200 Miles

100 Miles

50 Miles

$0

$100

$200

$300

$400

$500

$600

$700

$800

$900

0 50 100 150 200 250 300 350


Incr

emen

tal L

oss

NPV

= $

/kW

200 Miles

100 Miles

50 Miles

Figure 12-20 NPV of future losses for various maximum ac real power transfers. Loss factor = 0.5

Figure 12-21 Incremental NPV of losses as maximum ac real power is increased

Suppose, for calibration purposes, that a 25% increase in flow (220MVA to 275MVA) can be achieved with 25% series capacitive compensation. At 275 MVA current would be 690 amperes and reactive losses, based on data shown in the insert at the top of fig. 12-19 would be 1,000 kVAR/mile. 25% compensation would equal 250 kVAR/mile. If the cost of that compensation was $25/kVAR, compensation would cost $25 x 250 = $6,250 per mile. Fig. 12-19 shows losses

12-26

at the original MVA loading to be 70 kW/mile. A 25% increase in current will increase losses to 70x (1.25)2 = 156 kW/mile, an increase of 39 kW per mile. That will translate to a NPV of $3,000 x 39 = $118,125/mile; almost 19 times the cost of the capacitor investment!

c. Losses following ac to dc conversion.

Figs. 12-22a, b and c re-plot the data in fig. 12-20 for 50, 100 and 200 mile lengths, respectively, but also show losses for the same real power transfer after the circuit is converted to +/-220kV dc. DC losses assume either (a) bipole operation with positive current on one conductor and negative current split between the two remaining conductors or (b) normal tripole operation. The dc curve includes converter losses equal to 0.85% of the transmitted power per terminal.

The difference between ac and dc curves in both figures, shown in red, is the loss-based credit or penalty applicable to conversion. Two adjustments to this credit should be made.

i. The dc case should include an additional cost for re-insulation. For example a cost of $30,000 per mile would add $3 Million to the 100 mile case, $6 Million to the 200 mile case. At 200 MW this would represent $15 per kW and $30 per kW respectively.

ii. The ac extended ac case should include the cost of apparatus necessary to achieve the extension, probably more than offsetting the cost cited above.

The curves ignore the benefit, illustrated in fig. 12-19, that conversion can achieve a much higher boost in capacity.

-$40.0

-$20.0

$0.0

$20.0

$40.0

$60.0

$80.0

$100.0

$120.0

0 50 100 150 200 250 300 350


NPV

of L

osse

s =

$/kW

DC Operation

AC Operation

Loss-based Conversion Credit per Terminal

-$50

$0

$50

$100

$150

$200

$250

0 50 100 150 200 250 300 350


NPV

of L

osse

s =

$/kW

AC Operation

DC Operation


-$50

$0

$50

$100

$150

$200

$250

$300

$350

$400

$450

0 50 100 150 200 250 300 350


NPV

of L

osse

s =

$/kW

DC Operation AC Operation


a. 50 Miles b. 100 Miles c. 200 Miles

Figure 12-22 NPV of losses, with the example 230 kV circuit converted to +/- 220 kV dc

It is apparent that, while raising maximum ac loading above its original 187 MW operating limit is very expensive in terms of losses, converter losses are likely to be much greater for short line lengths. For longer length and/or high load levels, loss economics alone can represent a significant credit toward the cost of conversion to dc, assuming the latter to be in the range of

12-27

$100 per kW per terminal. That credit approaches 50% for the 100 mile length and 100% for the 200 mile length since converter losses are independent of length.

If, in the above example, the circuit had substantially larger ac conductors than assumed, the loss-based incentive for conversion would be much less since the insertion of converter losses would tend even more to outweigh dc’s transmission line loss advantage.

d. Effect of system context on NPV of losses

The above analysis dealt with a 230 kV line outside of its system context. Fig. 12-23a shows the same circuit underlying three 500 kV circuits, all with a power factor of .85 and with (n-1)-compliant loading. It will be instructive to consider several means for increasing that loading.

The dispatch of fig. 12-23a is limited by loading on a 500 kV circuit following loss of another 500 kV circuit. Actual flows and flow limits in the figure reflect the fact that the 230/500 kV susceptance ratio is 0.20 but the ampacity ratio is 0.25. If the 230 kV circuit is compensated to the point where its outage becomes equally limiting with the 500 kV3, total path flow can be increased to 3,061 MW, a gain of 69 MW as shown in fig. 12-23b. The corresponding loss increase is 705-654 = 51 kW/Mile. Valued at $3,000 per kW the NPV of additional losses would be 51x $3,000 = $153,000/mile. Assuming that the 230 kV circuit had a reactance of .7 ohms per mile, 20% compensation at a 400 MVA maximum rating (~1,000 amperes) would represent 3x.20x10002x.7/1000 = 420 kVA per mile which, at $35/kVA would have a cost of $14,700/mile; less than one tenth the NPV of the new loss burden.

Figure 12-23 n-1-compliant dispatch of an example 230 kV circuit underlying three 500 kV circuits

Construction of new 230 kV circuits would obviously add more capacity at a lower loss penalty. Fig. 12-24 duplicates the system of fig. 12-23 except with a new double-circuit 230 kV circuit added. The system, again dispatched to comply with n-1 constraints, increases transfer by 3,537 – 2,992 = 545 MVA with a loss increase of 884 – 654 = 230 kW/mile or 23 MW for the 100 mile case. The NPV of that loss increase according to the above assumptions is $3,000 x 23,000/106 = $69 Million.

2. 100 x (1/.20-.1/25)x.20=20% compensation

500 kV 1,100 / 1,600 194

1,100 / 1,600 194 3,520 MVA

1,100 / 1,600 194 2,992 MW

230 kV 220 / 400 73

Total 654

Losses per Mile 500 kV 1,108 / 1,600 196

1,108 / 1,600 196 3,601 MVA

1,108 / 1,600 196 3,061 MW

230 kV 277 / 400 116

Total 705

Losses per Mile

a. As found system b. 230 kV line with 20% series compensation

12-28

500 kV 1,156 / 1,600 214

1,156 / 1,600 214

1,156 / 1,600 2144,162 MVA

230 kV 231 / 400 813,537 MW

231 / 400 81

231 / 400 81

Total 884

Losses per Mile

Figure 12-24 n-1-compliant dispatch of the system of Figure 12-23 with two new 230 kV circuits added

Suppose the 230 kV circuit of fig. 12-23 were converted to 220 kV dc. In that case the bipole rating, based on 1,000 amperes, could be as high as 2 x 220 = 440 MW which, with a 15% overload capacity, would yield an emergency rating of 506 MW. Fig. 12-25 shows an n-1-compliant dispatch for that case. The incremental power transfer is 3,227-2,992 = 235 MW; 43% of the increased achieved with the double circuit ac addition. DCE is 235/440 = .53; meaning that the increase in transfer capability would be only half the MW rating of the converters. On the other hand no new line construction would be necessary.

losses: kW per mile500 kV 1,093 / 1,600 191

1,093 / 1,600 191 3,719 MVA

1,093 / 1,600 191 3,227 MW

440 / 506 120

Total 693

220 kV DCBipole

losses: kW per mile500 kV 1128 / 1,600 204

1128 / 1,600 204 3,987 MVA

1128 / 1,600 204 3,479 MW

603 / 760 240

Total 851

220 kV DCTripole

Figure 12-25 n-1-compliant dispatch with the 230 kV converted to a +/- 200 kV dc bipole

Figure 12-26 n-1-compliant dispatch with the 230 kV converted to a +/- 220 kV tripole

Incremental losses are 693-654 = 39 kW/mile or 3.9 MW for the 100 mile case. Converter losses in this case would be 440 x .85 x 2 = 7.5 MW, bringing the total to 11.4 MW – translating to a NPV of $3 x 11.4 = $34.2 Million or $48.4 per kW of incremental power.

If the same circuit were converted to the tripole configuration, the maximum rating would be 440 x 1.37 = 603 MW and the emergency rating, 1.09x1.15=1.26 times that, or 760 MW (Fig. 12-26). Maximum transfer capability has now been increased by 3479-2992 = 487 MW and DCE to 487/603 = 0.81

The tripole’s maximum of 760 MW is critical to the allowable ac loading. However as previously suggested, the dc circuit would probably be operated at a level lower than 603 MW since doing so would reduce total losses without reducing the n-1-compliant dispatch. Fig. 12-27

12-29

shows such a re-dispatch which, though not optimized, reduces losses by 48 kW/mile.4 The increase in losses over the original ac case is then 803-654 = 149 kW/mile or 14.9 MW for the 100 mile case. Adding converter losses of 2 x 0.85 x 400 = 6.8 MW brings the total to 21.7 MW. At $3,000/kw this represents a loss NPV of $134/kW; still low compared to the loss cost of much smaller increases in capacity achieved by ac system augmentation.

losses: kW per mile500 kV 1227 / 1,600 241

1227 / 1,600 241 4,030 MVA

1227 / 1,600 241 3,478 MW

350 / 760 81

Total 803

220 kV DCTripole

Figure 12-27 Re-dispatch of fig. 12-26 to reduce total system losses

Table 12-3 summarizes the foregoing examples and shows the 220 kV tripole dc conversion case to be approximately equal to construction of two additional 230 kV lines in terms of both transfer gain and loss NPV. Those two options achieve from 7 to 8 times the transfer increase achieved with 20% series compensation.

Table 12-3 Summary of path flow enhancement options

NPV of ∆MW ∆Loss ∆loss/ Losses DCE

(MW) ∆MW ($ Mil.) Series Compensation of 230 kV 69 5.1 7.4% 15.3Build 2 new 230 kV Circuits 545 23.0 4.2% 69.0Convert 230 kV to Bipole HVDC 235 11.4 4.9% 34.2 0.53Convert 230 kV to Tripole HVDC 487 21.7 4.5% 65.1 0.81

System Configuration

e. Conversion Cost vs. the Cost of New Circuit Construction

The primary economic question posed by the examples of table 12-3 is whether 2 x 603 MW of converter capacity would be cheaper than a like rating of ac terminals plus 100 miles of double circuit ac line construction. Fig. 12-28 shows the circuit cost, in $Millions/Mile, corresponding to the “break-even” distance for that trade-off.

4 Total path losses for the bipole case cited above are very close to minimum with the bipole operating at full rating.

12-30

$0.00

$0.50

$1.00

$1.50

$2.00

$2.50

$3.00

$50

$60

$70

$80

$90

$100

$110

$120

$130

$140

$150

DC Terminal Premium - $/kW/Terminal

Tran

smis

sion

Lin

e C

ost -

$ M

il./M

ile

50

100

150 200

Line Length - Miles

Figure 12-28 Transmission line cost for 2-ckt 230 kV vs. difference between dc and ac terminal costs

f. Capacity Value of Reduced Losses

More directly related to the transmission owners’ business incentive is the impact of losses on MW transfer capability and on his revenue base. Even though the capacity penalty of losses is an order of magnitude lower than the energy loss penalty it may be significant in comparing alternatives that yield roughly the same transfer increase. Kelvin’s Law suggests that, for the optimum conductor choice, the NPV of future losses should be approximately equal to the conductor-dependent cost of the circuit’s construction. [19] That rule implies that if losses represent 5% of transmitted power at normal circuit loading, an option which decreases losses from 5% of transmitted power to 4% would have a value equal to 20% of the conductor-dependent portion of the circuit’s cost. If conductor-dependent costs are half total cost, the change would equal 10% of the total cost. The same reduction in losses will free up additional transmission capacity equal to 5%-4%=1% of the circuit’s rating; one tenth the loss advantage. The congestion rent value of that 1% capacity increase may be higher than the pro-rata circuit cost as cited above, but the NPV of energy loss reduction will still be much more significant from an economics standpoint than the increase that reduction brings to congestion rent.

g. Conclusions

• The cost of extending the load-carrying capability of ac circuits is largely the increased NPV of losses attendant with higher power operation; not the cost of enhancement measures.

• Increased energy loss also reduces the useful capacity of a circuit and therefore congestion-based rents. However the NPV of capacity reduction imposed by losses is an order of magnitude less than the NPV of the energy loss itself.

12-31

• The NPV of incremental losses resulting from increased loading of moderately long existing ac circuits may approach or exceed the cost of converting the same circuit to dc.

• In cases where a converted circuit parallels other ac circuits, it is the emergency rating of dc rather than its operating level, which limits (n-1)-constrained transfer.

• Dc links within an ac system can be used to optimize the division of losses between itself and parallel circuits.

• New AC line construction is apt to be less expensive than dc conversion in areas of low to moderate line construction cost. DC conversion may be competitive economically where circuits are long or the cost of new line construction is high.

12.9 Fire and Ice Issues

a. Ice prevention

An additional incentive for ac to dc conversion may exist where an ac circuit is subject to dangerous ice loading. During commissioning of conventional dc bipole lines, it is common to operate the circuit in “round power” operation, i.e. with the voltage polarity of both poles the same so that full power can pass down one pole and return on the other without supplying any load current. The tripole system is unique in that at any given time, three unique “blocks” of power characterize the power being transmitted. Figure 12-29 illustrates the case where a tripole is operated for maximum rating, i.e. where at any given time, one pole will operated at 1.37 pu current, another at .37 pu, and a third at 1.0 pu. The “Normal” column in Table 12-4 shows the percentage of total power which each block represents. Just as with bipole round power operation, it is possible to reverse any one of the three power blocks shown in Table 12-4 by reversing voltage during that block’s tenure. The columns under the “Reversed” heading show three possible reversing combinations and the effect each reversal has on total power.

12-32

Table 12-4 “Round Power” solutions using conventional modulation of a tripole system

Figure 12-29 Blocks of power associated with tripole operation at its maximum

Modifications in modulation pattern can produce power blocks other than those shown in table 12-4, making possible a relatively smooth reduction in net delivered power from 1.0 to near 0 while maintaining 1.0 pu current in line conductors. That regime is shown conceptually in fig.12-29. Whether or not 1.0 pu power will prevent ice formation is unclear. In any case a method has been proposed which will enable bundled conductors, normally operating in parallel, to be operated in series during hours of high icing exposure. That mode of operation could be limited to sections which are subject to ice build up. The system requires insulated bundle spacers and can be implemented with relatively simple conductor-mounted switches. [20]

Steps achieved by changes in modulation

Intermediate levels achieved by firing angle control

Modulation & Phase Angle Adjustments

Net D

eliv

ered

Pow

er

Current in each conductor remains = 1 pu

Figure 12-30 Adjustment of net delivered tripole power while maintaining 1.0 pu power on all poles

V1

Block 1 50%

Block 2 14%

Block 3 36%

Imin 14% -14% 14% 14%Imax 50% 50% -50% 50%

Imax - Imin 36% 36% 36% -36%100% 73% 0% 27%Total

ReversedBlock Normal

12-33

b. Brush Fire Response

Transmission circuits are subject to flashover during brush fires; common in many developing countries. Although flashover occurs on one ac phase first, all must be tripped with little chance of restoration until the fire has either burned out or passed the region under the line. Conversion to dc presents the possibility, illustrated for the tripole case in fig. 12-31, where a flashover on one pole leaves the remaining poles energized but triggers a reduction in voltage on all poles; here assumed to be 5%. The circuit then operates at 95% voltage until another pole flashes over, causing a reduction to 90%. In the illustration, a stable operating point is reached at 85% voltage. Although no field tests have been conducted to determine the practicality of such a scheme it could add to the conversion incentive in regions subject to fire exposure.

V= 100% V= 100% V= 95% V= 95%

V= 95% V= 90% V= 90%

V= 90% V= 85% V= 85%

V= 75%

. . . . . !

V= 100% V= 100% V= 95% V= 95%

V= 95% V= 90% V= 90%

V= 90% V= 85% V= 85%

V= 75%

. . . . . !

Figure 12-31 Example sequence of voltage reductions in response to brush fire trip-outs

12.10 Logistics of Non-Disruptive Conversion from ac to dc

The incentive for increasing transmission capacity is often greatest for circuits which have high impact on energy costs or reliability if removed from service. Conversion to dc can be accomplished by a relatively non-disruptive manner. The steps might be as follows:

a. Re-insulation

Replacement of insulators designed for dc operation is a necessary first step in conversion of an ac circuit for dc operation. The simplest change-out, i.e. a one-for-one replacement of porcelain or glass insulating units, will normally allow operation of the circuit at a dc voltage at least equal to the former ac line-to-ground crest voltage. Efforts to operate at a higher voltage, if allowed by clearance and gradient limits, may require increasing the number of dc insulators, going to

12-34

polymer units also available for dc operation, or changing the suspension configuration to restrain wind swing. In any case most such changes can be made “live,” i.e. while circuit remains in ac operation.

b. Converter Station Construction

Maximum benefit from conversion is realized when converter terminals are adjacent to or part of the sending and receiving ac substations. Very little of the space allocated to the pre-conversion ac line will be useful for the dc converter terminals, nor is it assured that the line’s ac breaker will be adequate for the higher rating achieved with dc. DC terminal construction need not disrupt operation of the ac circuit if busses are configured for rapid ac-to-dc changeover per fig. 12-32. Disconnect a is closed while b and c are open during terminal construction and ac operation of the circuit. Disconnect a is opened and disconnects b and c are closed during piece-wise commissioning tests - conducted during off-peak operation.

ac breakerac or dc high voltage line a

cb ac breaker

HVDC Converter Terminal

Figure 12-32 Example sequence of voltage reductions in response to brush fire trip-outs

c. Reconductoring

This chapter has compared certain ac flow enhancement options with conversion to both bipole and tripole dc of the same three-conductor system. AC reconductoring, perhaps combined with reactive compensation, is sometimes an attractive means of increasing ac capability. If reconductoring is undertaken as a part of an ac to dc conversion opportunity three important advantages may accrue:

i. The new conductor may allow a higher dc voltage if it is selected to reduce operating gradient, thus compounding its current and voltage benefits

ii. The ac benefits achieved through reconductoring are multiplied by the ac-to-dc power ratio, thus giving a substantially higher return on the reconductoring investment.

12-35

iii. During ac-to-dc conversion new conductor can be installed non-disruptively without live line work.

The third of these points is illustrated in fig. 12-33. In the case of a tripole configuration it is reasonable to assume that temporary bipole operation, as a first stage, will allow at least as much power transfer as under pre-conversion ac operation. Thus step 3 in fig. 12-33 provides an opportunity to reconductor the spare phase position. When that operation is completed, the reconductored pole can be restored to operation and another pole removed from service for reconductoring. When all three poles have been reconductored, the circuit can go into tripole operation with the compounded benefits of the new conductor.

////\\\\////\\\\////\\\\////\\\\////\\\\////\\\\////\\\\////\\\ ////\\\\////\\\\////\\\\////\\\\////\\\\////\\\\////\\\\////\\\

////\\\\////\\\\////\\\\////\\\\////\\\\////\\\\////\\\\////\\\ ////\\\\////\\\\////\\\\////\\\\////\\\\////\\\\////\\\\////\\\

* Once completed, rotate out of service position to other positions, maintaining continuous bipole operation

A B C A B C

Replace insulators with dc units - liveNormal ac operation

+Bipole HVDC operation

Out of Service for Reconductoring* Full Tripole Operation

+ +/

1 2

3 4

Figure 12-33 Non-disruptive change-over from ac to dc

12.11 Potential Conversion of Cable Circuits

Conversion of existing high voltage cable circuits to dc, either underground or undersea, is an attractive prospect inasmuch as the justification for undergrounding transmission is implicit evidence that the cost of additional transmission is very high. The issues are different with cable circuits in two important respects:

i. The “thermal averaging” principle associated with tripole operation of overhead circuits can be viewed as three separate thermal problems; one for each phase position. In most cable circuits the thermal limit is set by aggregate heating of three conductors, i.e. by the heat release capability of a pipe or closely spaced conductor group.

12-36

ii. Voltage limits for cables are more complex and involve the specific dielectric used including its vulnerability to voltage reversals.

a. Thermal limits for a system comprising an aggregate heat sink

The thermal limits of a three conductor system subject to an aggregate heat release limit can best be illustrated by the oil-filled paper cable case illustrated in fig. 12-34. The oil-paper insulation system used in pipe-type cable has long been considered adaptable to dc operation at a voltage at least equal to crest line-to-ground and possibly crest line-to-line voltage of the former ac operating voltage. Conversion to dc achieves a substantially higher dc/ac power ratio, in part because charging current, dielectric losses, and pipe losses, absent with dc operation, account for an important fraction (e.g. 1/3) of the ampacity of typical ac cables. However application of the thermal averaging characteristics of the tripole system to three cables enclosed in a common pipe, all sharing the same heat sink and aggregate thermal limit, requires careful analysis. Fig 12-34a shows a three-phase ac cable converted to bipolar dc operation. One of the former ac cables is idled. The current in the active cables is presumed equal to the former ac rms current. Per unit power and per unit losses shown in the table opposite fig. 12-34a are set at 1.0 for convenience in subsequent comparisons. L is the ratio of pu losses to pu power and Lmax is the pu loss on the most heavily loaded cable.

In fig. 12-34a two active conductors at 1.0 pu current will produce just 2/3 the aggregate losses which the pipe, duct, or trench was designed to accommodate, the latter being 1.5 pu for the rating convention used here. Assuming that total losses need only be kept within a fixed limit, Fig. 12-34b shows bipole current increased to the point where the aggregate loss limit of 1.5 is supplied by just two active cables. In that case losses in the two active cables can be increased to 3/2 and their current increased to •1.5 = 1.22. Total power will increase by the same multiple. The ratio of pu losses to pu power transmitted is simply 1.5/1.22 = 1.22.

Fig.12-31c duplicates the case shown in 12-31b but divides the return current between two conductors. Total power remains the same but losses drop by 1.222(1+.52+.52)/ (1.222+1.222) or 75% of the value in fig. 12-34b. One cable still needs to transfer 50% more heat from its inner conductor to the common heat sink than in the base case of fig. 12-34a. With total losses now again below the aggregate limit of the heat sink, current could be increased still further.

In Fig. 12-34d the highest dc current is increased to •2, the low currents to half that value, bringing aggregate losses back to 1.5 – the ac reference case and the maximum aggregate rating of the heat sink. However the heat transfer demand from the most heavily loaded conductor would now be twice the value seen in balanced ac operation.

In Fig. 12-34e, the tripole case, high level of current is alternated between two cables. In this case the power level stays the same but the rms current in the interchanging poles is (1.4142+.7072)/2= .5x•2.5 = .79, less on the average than the nominal ac heat release per conductor position. The total loss remains as in fig. 12-34d.

Because of the very long thermal time constant of cables, the period of voltage reversal could be in the order of 30 minutes, assuming converter station equipment was built to accommodate a similar thermal-averaging cycle.

12-37

According to the above cable rated for 345 kV ac, if operable at +/- 345 kV dc, would have a dc capability, according to fig. 12-34d of approximately:

3.237.3414.12

≈=− xxPP actripoledc (12-6)

where the factor 0.7 in the denominator is a conservative estimate of the gain from absence of charging current and magnetic losses. The high cost of urban converter terminal siting would weigh against such a conversion whereas alleviation of short circuit current levels would weigh in its favor. There remains too a question of whether tripole redundancy benefits would apply where (i) most faults, e.g. anchor damage, would take out all three conductors and (ii) even if that were not the case, the question of whether two poles could operate, even temporarily, with the third faulted.

Power 1.00

a Losstot 1.00

1.00Lmax 1.00

Power 1.22b Losstot 1.50

1.22Lmax 1.50

Power 1.22c Losstot 1.13

0.92Lmax 1.50

Power 1.41d Losstot 1.50

1.06Lmax 2.00

Power 1.41e Losstot 1.50

1.06Lmax 1.25

0

+1.00 -1.00

0

+1.22 -1.22

+1.22

-0.61 -0.61

+1.41

-0.71 -0.71

L

L

L

L

L

+1.41

-0.71 -0.71L

Figure 12-34 Various modes of dc current in a three-cable group

12-38

b. XLPE cable considerations

Conventional ac-rated cross-linked polyethylene (XLPE) cable has not been deemed suitable for dc operation because with long-term dc energization, space charge build-up (a) renders the voltage gradient within the cable incalculable, thus impossible to design for and (b) sudden voltage reversals with strong concentrations of space charge cause very high local stresses, ultimately leading to failure. [21, 22]

The time constant of charge build-up within the major portion of insulation is tens of hours while to the extent it is not prevented by semi-conducting wraps, charge build-up on the immediate surface of the conductor/semicon surface is measured in minutes. In either case ac voltage reversals characterizing power frequency, i.e. a period of ~ 16 milliseconds, are fast enough to prevent build up, thus finessing the problem. There is reason to suspect that reversal times characteristic of tripole modulation, while much longer than power frequency reversals, may also be frequent enough to allow standard xlpe cable to be used with this form of dc. In that case all three poles would have to be equipped either with back-to-back conventional valves or bi-directional valves so that polarity reversal could be achieved on each as shown in fig. 12-35. With symmetrical modulation of this sort, total power is 1.414 times the power that would be conducted on just two cables under a bipole scheme.

Pole 1

Pole 2

Pole 3

V1

V2

V3

Fig. 2 Asymmetrical tripole current modulation

Figure 12-35 Symmetrical modulation of a tripole dc system

However cable experts have divided opinions on the effectiveness of reversals in the 1 to 30 minute time frame. Recent work suggests that to achieve the benefit of polarity reversals as applied to XLPE cable, the tripole reversal period may have to be in the range of 1 to 2 minutes. [23] In any case a careful testing program would be needed before serious consideration could be given to use of standard XLPE cable with tripole dc.

12.12 References

1. “HVDC Conversion of HVAC Line to Provide Substantial Power Upgrading,” A. Clerici,

L Paris, P. Danfors, IEEE SM 300-4 PWRD, 1990

12-39

2. “Analysis of Possible Enhancement of Transmission Capacity While Converting 220 kV Alternating Current Overhead Lines into Direct Current Lines” A. Orzechowski. CIGRE B4-105, 2004”

3. “Conversion of AC Line into HVDC,” M.I.Khan, R.C.Agrawal, HVDC 2006 Congress, University of KwaZulu-Natal, July, 2006.

4. “EEI Survey of Transmission Investment” Edison Electric Institute, 701 Pennsylvania Ave. NW Washington DC May 2005

5. “Loss Economics as a Factor in Comparing FACTS vs. DC Conversion as a Means of Power Transfer Enhancement,” L. Barthold, H. Clark, 9th annual FACTS User’s Group Meeting, Montreal Quebec, September 6-8, 2006.

6. Reliability Standards for the Bulk Electric Systems of North America, May 2, 2006. North American Electric Reliability Council 116-390 Village Blvd. Princeton, N.J. 08540.

7. Discussion B4 Preferential Subject 1 by L. Barthold, CIGRE 2004 8. “Conversion of AC Transmission lines to HVDC using Current Modulation,” Lionel

Barthold, Hartmut Huang Inaugural IEEE PES 2005 Conference and Exposition in Africa. Durban, South Africa. 11-15 July, 2005

9. Reliability Standards for the Bulk Electric Systems of North America, May 2, 2006. North American Electric Reliability Council 116-390 Village Blvd. Princeton, N.J. 08540.

10. Bidirectional Valves: Specific Practical Applications of Interest. EPRI, Palo Alto, CA, and Bonneville Power Administration, Vancouver, WA: 2005. Product ID # 1014493

11. L.O.Barthold, D.A.Woodford, “Application of Voltage Sourced Converters to Tripole HVDC Transmission. 9th EPRI FACTS User’s Group Meeting, Montreal, Quebec, September 6-8, 2006

12. “National Electric Safety Code,” ANSI/IEEE Publication C2-2007 13. “Transmission Line Reference Book” 115-138 kV Compact Line Design EPRI project 260,

1978 14. P. Naidoo et al. “Investigations into Electrical and Corona Effects for the Upgrade of

HVAC Transmission Lines to HVDC. HVDC 2006 Congress, University of KwaZulu-Natal, July, 2006.

15. Jose A. Jardini. “CIGRE Working Group SC B2-05: JWG 17 Study Report for 800 kV HVDC.

16. Reference to EPRI transmission line design workstation 17. U.S. Energy Information Administration (IEA), “Annual Energy Outlook 2006 with

Projections to 2030,” http://www.eia.doe.gov/oiaf//aeo/forecast.html. 18. M.R. Simmons, “The Future Cost of Energy,” Offshore Technology Converence4 2004,

Houston, Texas, USA May 5, 2004 19. http://en.wikipedia.org/wiki/Overhead_powerline 20. Pierre Couture “Switching apparatus and method for a segment of an electric power line”

US Patent 6.727,604 B2 Apr. 27 2004 21. Behavior of Extruded HVDC Power Transmission Cables,” M. Pays. CIGRE Paper 2107,

1998 Plenary Session 22. “A Review of HVDC Extruded Cable Systems,” Ruter, F. Jicable 2001 23. “Characteristics of XLPE MV-size DC Cables by Means of Space Charge Measurements,”

R. Bodega, P.H.F. Morshuis, E.J. D. Straathof, U.H.Nilsson, G.Perego, 2006 Annual Report Conference on Electrical Insulation and Dielectric Phenomena. 1-4244-0547-5/06/$20©2006 IEEE

http://www.eia.doe.gov/oiaf//aeo/forecast.html

13-1

13 HARMONICS (OUTLINE)

13.1 Harmonic Sources

13.1.1 Industrial and Consumer Loads

13.1.2 Power Electronics Based Controllers, Compensators and AC-DC Converters

13.2 Possible Effects of Harmonics

13.2.1 Initiation of Oscillation at Network Resonant Frequencies

13.2.2 Overvoltages

13.2.3 Overcurrents

13.2.4 Increased Operating Temperature, Reduced Lifetime

13.2.5 Telephone and Radio Frequency Interference

13.3 Typical Utility Specified Harmonic Limits

13.3.1 Individual and Total RMS Harmonic Voltage

13.3.2 Individual and Total RMS Injected Harmonic Current

13.3.3 Telephone Interference TIF and IT Factors

13.4 Harmonic Reduction Techniques for Power Electronics Equipment

13.4.1 Circuit Topology

A. Multi–Pulse Structure

B. Multi-Level Structure

13.4.2 Various PWM Techniques

13.4.3 Filtering

13.4.4 Combinations of the Above

13-2

13.5 Passive Filter Design Consideration

13.5.1 Network Impedance Frequency Characteristic

A. Actual Measure

B. Simplified Worst Case Assumption

13.5.2 Avoidance of Resonances

A. Selecting Impedance “Zero” and “Pole” Frequencies

B. Damping

C. Filter Attenuation vs. Frequency Characteristic

13.6 Active/Hybrid Filtering and Damping Possibilities.


14-1

14 STUDIES FOR TRANSMISSION CONTROLLER SELECTION AND SPECIFICATION PREPARATION (OUTLINE)

14.1 Main Objectives of Studies

14.1.1 Identify System Needs and Candidate Solutions

14.1.2 Identify Possible Locations for the Installation

14.1.3 Compare Technical and Economic Benefits of Candidate Solutions

14.1.4 Prepare Specification for Equipment of Selected Solution(s)

14.2 Technical Information Needed for Controller Specification

14.2.1 Equipment Type and Operating Environment

14.2.1.1 Application Objectives

14.2.1.2 Controller Type(s) Wanted (or Acceptable)

14.2.1.3 Connection Point to the Transmission System

A. Bus Voltage

B. Max and Min Short Ckt. MVA

C. Max Unbalance (Negative Sequence)

D. Max Fault Clearance Time

E. Range and Rate of Frequency Change

F. Transient Overvoltage Levels and Durations

14.2.1.4 Continuous MVA/MW/MVar Ratings

14.2.1.5 Normal Operating Regions

14.2.1.6 Possible Abnormal and Emergency Operating Conditions

14.2.1.7 Required Short Term Ratings and Operating Limits

14.2.2 Operating Modes, Control Functions and Performance Requirements

Studies for Transmission Controller Selection and Specification Preparation

14-2

14.2.2.1 Identify Required Operating Modes and Control Functions

14.2.2.2 Define Steady-State Operating Characteristics

14.2.2.3 Define Dynamic and Transient Performance

A. During Small Disturbances

B. During and Following Large Disturbances

C. Coordination of Equipment Protection with Dynamic

Performance Requirements

14.2.2.4 Local and Remote Control Operation, Alarm Monitoring,

Diagnostic, and SCADA Requirements

14.2.3 Harmonic Performance

14.2.3.1 Existing Harmonic Environment at the Location

14.2.3.2 System Impedance Characteristic at the Bus (Measured or Worst Case Estimate)

14.2.3.3 Acceptable Harmonic Generation by Equipment

A. Voltage

B. Current

C. Telephone Interference

14.2.4 Criterion for Equipment Loss Evaluation

14.2.5 Acoustic Noise Restrictions

14.2.6 Availability and Reliability Requirements

14.2.6.1 Redundancy and Modularity in the Power Electronic Circuit

14.2.6.2 Full or Partial Control Redundancy

14.2.6.3 Spare Part Short- and Long-Term Availability

14.2.7 Acceptance and Commissioning Tests

14.2.7.1 Control System Operational and Performance Tests

14.2.7.2 Factory Type Tests

14.2.7.3 On Site Commissioning Tests

14.3 Main System Studies

14.3.1 Load Flow Studies for

14-3

A. Equipment Ratings (Steady-State and Short-term) and Location

B. Determine Steady-State Operating Point and Range

C. Determine Contingencies and Abnormal Operating Conditions

14.3.2 Small Disturbance (Steady-State Stability) Studies for

A. Determining Existing System Damping

B. Finding Effective Control Signal and Equipment Damping

Capability

14.3.3 Large Disturbance (Transient Stability) Studies for

A. Confirming Location and Adequacy of Rating

B. Determining Voltage and Power Excursion Levels at Required

Damping

C. Verifying Stability Improvement Controller Can Provide

D. Finalizing Controller Control Mode, Control Variables, Operating

Range and Limits

E. Determining if There is Any Interaction Between Controller and

Existing Power System Constituents

14.3.4 Voltage Stability Studies for Determining Voltage Levels and

Recovery After Disturbances and System Contingencies with Existing

Slow-Acting Voltage Regulators and Voltage Dependent Loads

14.3.5 Harmonic Studies for

A. Determining Acceptable Levels of Voltage and Current Distortion

Generated by the Controlled to Meet Utility Standards

B. Detecting Possible Harmonic Interactions

C. Detecting Possible Resonances with System and Controller Filters and Other Reactive Components.


15-1

15 ARCHIVE REFERENCE FOR SELECTED INSTALLATIONS (OUTLINE)

15.1 Introduction

Whereas the purpose of the previous chapters in the Reference Book are self-evident, this chapter’s purpose requires some explanation. The idea is to provide potential users with systematic, summary information on noteworthy installations in all categories of power electronics-based transmission controllers. The information would include application background and objectives, equipment description and site details with relevant technical data, operating modes and features, control functions, and achieved performance. The aim of these systematically organized installation descriptions is to provide a selection of applications precedents to new users to help their equipment selection and evaluation process.

15.2 Controller Groups

The following Controller Groups, to include the individual installation descriptions, are proposed:

15.2.1. Shunt Compensators

1A Static Var Compensator (SVC)

1B Static Synchronous Compensator (STATCOM) for Transmission Applications

1C Static Synchronous Compensator (STATCOM) for Arc Furnace Compensation

15.2.2. Series Compensators

2A Thyristor Controlled/Switched Series Capacitor (TCSC and TSSC)

2B Static Synchronous Series Compensator (SSSC)

15.2.3. Thyristor Controlled Phase Shifters (TCPS) and Phase Angle Regulators (TCPAR)

15.2.4. Thyristor Controlled Voltage Regulators (TCVR)

15.2.5. Multi Function Transmission Controllers

5A Unified Power Flow Controller (UPFC)

Archive Reference for Selected Installations

15-2

5B Interline Power Flow Controller (IPFC)

5C Convertible Static Compensator (CSC)

15.2.6. Voltage-Sourced Converter Based Back-to-Back (BtB) Ties

15.2.7. Voltage-Sourced Converter Based DC Transmission Systems

15.3 General Description Structure

The following general description structure is proposed for each installation within its relevant Controller group:

15.3.1. Title and Location of the Installation

15.3.2. Brief Background Description of the User Utility System

15.3.3. Reasons for, and Selection Process of the Installation

15.3.3.1 Planning and/or Contingency Studies

15.3.3.2 Reasons for, and Objectives of the Installation

15.3.3.3 Candidate Equipments for the Installation

15.3.3.4 Selection Criteria

15.3.3.5 Main Specification Requirements

15.3.3.6 Reason for the Selection Result

15.3.4. Description of the Equipment Installed

15.3.4.1 Singe Line Diagram of the Total Installation

15.3.4.2 Simplified Layout of the Installation

15.3.4.3 Rating and Structure of the Power Electronic Circuit

• Nominal Rating

• Short-Term Rating

• Transmission Voltage

• Power Converter/Circuit Output/Operating Voltage

• Coupling Transformer Winding Arrangement and Voltages

• Type of Converter/Circuit Topology

15-3

• Pulse Number

• Number of Converters/Circuits Used to Provide the Output

• Method Used to Combine the Outputs of Individual Converters/Circuits

• Type of Valves Used

• Type of Semiconductor Used in the Valve

• Number of Semiconductors used in Series and/or Parallel in the Valve

• Voltage and Current ratings of the Individual Semiconductors

• DC Capacitor (if used)

− Nominal DC Voltage

− Rating in kJ

• Cooling System

15.4 Functional Structure and Operating Features of the Control System

15.5 Summary of Commissioning and other Performance Tests

15.6 Results and Benefits of the Installation


1

A CONTRIBUTORS

Dr. Abdel-Aty (Aty) Edris, Editor Dr. Edris was born in Cairo, Egypt. He received his BS with honors from Cairo University, the MS from Ein-Shams University, Egypt, and the Ph.D. from Chalmers University of Technology, Sweden. Dr. Edris spent 12 years with ABB Company in Sweden and the United States, in the development and application of Reactive Power Compensators and High Voltage DC Transmission systems. In 1992, Dr. Edris joined EPRI as Manager of Flexible AC Transmission System (FACTS) Technology.

Dr. Edris is Technology Manager of EPRI Power Delivery and Markets, and a member of several IEEE and CIGRE Working Groups.

Dr. Edris is the recipient of the IEEE 2006 Award for industry leadership and scientific contribution to FACTS Technology, pioneering transformation of electric transmission system into flexible, controllable, yet secure system operated at thermal capacity.

Peter Lips – Chapter 2: Power Semiconductors and Valves Mr. Lips received the German equivalent to the MSEE from Technische Hochschule Darmstadt, Germany in 1965. Following a 23-year career in HVDC at BBC Brown Boveri in Germany and Switzerland, he joined Siemens AG in Germany in 1988 as manager for design and manufacturing of HVDC and SVC thyristor valves. He retired in 2004 after 39 years in HVDC.

Peter was active in the IEEE-PES in a number of committees and working groups and served as PES Secretary from 1995-1998. He also served as

chairman of IEC Technical Committee 22 "Power Electronics" from 1989-2002 and as German Representative to Cigré Study Committee B4 from 1996-2004.

Lionel O. Barthold – Chapter 12: AC to DC Conversion Mr. Barthold is president of iMod, Inc. Following a career in power transmission research with General Electric, Barthold founded and served as CEO and Chairman of Power Technologies, Inc. (PTI) from 1969 to 1998. He has served on and chaired numerous committees in CIGRE, IEEE, and ANSI and is a past president of the IEEE Power Engineering Society. Author of over 75 technical papers and five US patents, Barthold is a Fellow of IEEE, holder of its Power Life Award, and a member of the

National Academy of Engineering.

2

Dr. Bjarne R. Andersen – Chapter 11: VSC-Based DC Transmission Dr Bjarne R. Andersen has his own independent consultancy company, Andersen Power Electronic Solutions Ltd, which focuses on HVDC, SVC and FACTS systems and provides assistance to existing and prospective owners of such systems. Prior to starting his own company, Dr. Andersen spent 26 years working for ALSTOM T&D Power Electronic Systems Ltd. Dr. Andersen is active within CIGRE, IEEE and the IET (formerly the IEE). He was the UK Regular Member for Study Committee B4 (HVDC and Power Electronics) from 2000 to 2006. He was the convenor of

CIGRE WG B4-37, VSC Transmission, and is currently the convenor of WG B4-39, Integration of Large Scale Wind Power using HVDC and Power Electronics. He received the CIGRE Technical Achievement Award in 2004. He is a Fellow of the IET, and a senior member of the IEEE.

Dr. Laszlo Gyugyi – Chapter 1: AC Transmission Systems and Chapter 8: Generalized AC Transmission Controllers (UPFC) and (IPFC)

Dr. Gyugyi received his education at the Budapest University of Technology, the University of London, the University of Pittsburgh, and the University of Salford, England. He joined the Westinghouse Science & Technology Center in 1963, where he was appointed Manager of the Power Electronics Department in 1979 and Technical Director in 1995, and was responsible for the Corporation’s Power Electronics technology development for 23 years. Dr. Gyugyi has 78 US patents, authored/coauthored several books and numerous publications. He received the Westinghouse Order of Merit (the Corporation's highest honor) in 1992, the IEEE William E. Newell Power Electronics Award in 1994, and the first FACTS Award in 1999. Since 2002 he provides consultation on Power Electronics and its utility applications.

1

B AUTHOR’S GUIDE

Introduction

This Author’s Guide is designed to provide guidelines for authors to follow in the writing of chapters for the Gold Book. These guidelines will help to ensure that the completed book has a consistency and uniformity of presentation and style.

Sources The guidelines are based on several sources including the EPRI Editorial Style Guide, The Authoritative Dictionary of IEEE Standards Terms (7th edition), and the EPRI AC Transmission Line Reference Book, 200 kV and Above (Red Book).

Editorial Committee The Editorial Committee for the Gold Book will review each chapter to ensure the accuracy and consistency of the book. The Editorial Committee includes: (*to be determined*).

Author’s Guide Contents

Section Page Number

Production Process 2

How to Set Up Files 6

Format: Headings 7

Format: Tables and Figures 7

Format: Equations 8

SI Units and English Measures 8

Writing Principles 10

Language Guidelines 10

References 12

Index 13

Copyright Issues 13

Terms 14

2

Figures Page Number

Figure 1. Production Process 3

Figure 2. Sample Copyedit Page 4

Figure 3. Sample Layout Page 5

Production Process

The production process is designed to be easy for authors to use, and to require no additional purchase of software for writing or review. The production will take place in five steps (see Figure 1): 1. First Drafts First drafts are written by authors or author-teams. First drafts should not be submitted for production until the entire chapter is complete and the draft has been reviewed and approved by all members of the chapter team. 2. Technical Review Once the first draft is complete, it should be submitted for Technical Review. The reviewer(s) may draft a detailed review memo. Alternately, reviewers may also choose to comment on and suggest revisions using the “Track Changes” mode in Microsoft Word (Tools, Track Changes). Authors will review these suggestions and accept or reject these changes using Track Changes to produce a clean manuscript. 3. Copyedit After changes from the Technical Review have been resolved, the chapter will be copyedited. Copyediting will revise the text for spelling, punctuation, grammar, and adherence to the style guidelines. The copyedit will be made in the “Track Changes” mode of Microsoft Word (see Figure 2). The copyedited chapter will sent back for review and approval by the authors before proceeding. 4. Layout Once changes from copyediting have been resolved, the chapter will be laid out. The production team will lay out chapters in a two-column page format (see Figure 3). Pages will be modeled on the format and look of EPRI’s AC Transmission Line Reference Book, 200 kV and Above (Red Book). Important: Substantive revisions to chapters should be made BEFORE layout—as much as possible. Once the book is in layout, changes are more time consuming. (Compression ratio: The laid-out pages reduce the chapter page counts by about half.) 5. Publication Once the document has been laid out and approved by authors, the book will be published.

3

Figure 1 Production process.

2. Technical Review

3. Copyedit

4. Layout

5. Publication

Author Review

Author Review

Author Review

1. First Draft

4

Figure 2 Sample copyedit page.

5

Figure 3 Sample page layout.

6

How to Set Up Files

When writing first drafts, please use the following guidelines for creating your files:

Text

• Write files in Microsoft Word. • Drafts should be in 12 point type, Times New Roman, single-spaced, in one column, with 1-

inch margins. • For new paragraphs, do not indent first line. Just do a double return between paragraphs. • Include all text, tables, equations, and figures in the file.

Equations

• Write equations in Microsoft Equation 3.0.

Figures—Graphs and Photographs

• Important: If graphs or photos are created in any other program than Word, PowerPoint or Excel, please go to File, Save As. Then save files as .tif, .jpg, or .bmp format and insert that figure in the document. (.tif is preferable.)

• If you include figures in Excel, supply the original spreadsheet. • Annotate graphs. Label parts of the graph. Do not use legends. • If necessary, hand-drawn figures may be submitted to the editor to be drawn. • Use the font Helvetica for wording in figures, if possible. Do not use the font Arial; it may

lose or change characters during production. • For photos, use a version as close to the original as possible.

Index

• Make an alphabetical list of key words and their section location for an Index. See instructions on page 13.

Completed Files • Post completed files on the FTP site in the folder for your chapter. • If your completed chapter is large (more than a few MB), please post it in separate section

files. This will create files that are more manageable for others to handle.

7

Format

Headings

For consistency, all chapters should have the same style of main headings and subheadings. There are four levels of headings, as shown below.

Main Heading

One-Decimal Number, All Capital Letters, Boldface Example: 1.2 FUNDAMENTAL RELATIONSHIPS OF AC POWER TRANSMISSION

First Subheading Two-Decimal Number, Capital and Lowercase Letters, Boldface Example: 1.2.1 Analytical Characterization

Second Subheading Capital and Lowercase Letters, Boldface and Italic Example: Ideal (Lossless) Line

Third Subheading Capital and Lowercase Letters, Italic Example: Conclusions

Tables and Figures

Tables and figures should be numbered with the number of the subsection in which they appear. That is, the first table in Subsection 2.2 would be Table 2.2-1. The first figure in Subsection 2.2 would be Figure 2.2-1.

Tables

Table titles should be initial capped. Example: Table 2.5-1 Steady-State Operating Characteristics Table titles appear above tables. Table footnotes are superscript numbers.

Figures

Figure captions should be first-word-only capped and with a period at the end. Example: Figure 3.2-23 Simplified diagram of VSC transmission scheme. Figure captions appear below figures.

8

Photographs are considered figures.

Equations

Equations should be written using Microsoft Equation 3.0, which is available free with Microsoft Word. (To access it, go to Insert, Object, and select Equation 3.0.) Equations should be numbered to the right of the equation, using the same system as for tables and figures. Example, the third equation in Section 3.4 would be 3.4-3. If equations are long, they may need to be broken in the layout. Please indicate where they can broken. When identifying symbols in an equation, put them in a vertical list, not run-in on one line. Put the “Where:” at the top of the list. If equations are jpg files, rewrite them in Equation 3.0. Or, submit them to the editor to be generated in Equation 3.0.

SI Units and English Measures

Wherever reasonable, use the International System of Units (SI), with the English Units following in parentheses. Please note: If existing tables or figures use English units, leave as is. If conversion to SI units is difficult, leave as is. Imperial units may be used, but SI units are preferred, as the Gold Book is migrating to them for the future.

SI Base Units

Quantity Name Symbol length meter m mass kilogram kg time second s electric current ampere A thermodynamic temperature kelvin K

9

SI Derived Units

Derived Quantity Name Symbol

area square meter m2 volume cubic meter m3 speed, velocity meter per second m/s acceleration meter per second squared m/s2 wave number reciprocal meter m-1 density, mass density kilogram per cubic meter kg/m3 specific volume cubic meter per kilogram m3/kg current density ampere per square meter A/m2 magnetic field strength ampere per meter A/m

SI Derived Units with Special Names and Symbols

Derived Quantity Name Symbol frequency hertz Hz force newton N pressure, stress pascal Pa energy, work, quantity of heat

joule J

power watt W electric charge coulomb C electric potential difference volt V capacitance farad F electric resistance ohm Ω electric conductance siemens S magnetic flux weber Wb magnetic flux density tesla T inductance henry H Celsius temperature Degree Celsius °C

Exceptions

Use gauss, milligauss. Put tesla in parentheses. Usage Periods are not ordinarily used with abbreviations for units of measure (e.g., m, s, kg). One exception is the abbreviation for inch (in.). When it stands alone (e.g., 6 in.), it should have a period so that it is not confused with the preposition “in”. When used with an exponent (in2) or as part of a compound unit of measure (in/s), the period is omitted.

10

For more information, see the 2001 edition of The International Systems of Units (SI), NIST Special Publication 330, which is available on the web at http://physics.nist.gov/Pubs/SP330/sp330.pdf.

Writing Principles

Authors should keep in mind a number of general principles, so that all chapters are consistent in their presentation and style.

Introduction

Each chapter should begin with a brief (5-10 paragraphs) introduction. The introduction should describe the scope of the chapter and its context within the book. It may describe the topic’s significance and changing attitudes/approaches toward it. In addition, it should provide a quick “roadmap” to the main sections included within the chapter.

Flow

When writing your chapter, be sure the information is developed logically from section to section and within each section. Be aware of the number of sections: Is it within the average of other chapters? Avoid overly long or overly short sections. If you list three topics at the start of a section, discuss them in that order.

Depth

Present information in an appropriate level of detail. Avoid drilling down too deeply in one area and covering another too shallowly. Be aware of total page count, which will be discussed prior to the start of writing: Are you within the expected range?

Tone

Emulate the tone of other chapters. Use clear, declarative sentences.

Appendices

Present data, figures, and other detailed information, which would clutter the main text, in appendices.

Language Guidelines

Guidelines for language should follow the EPRI Editorial Style Guide, which is at: http://www.epri.com/corporate/discover_epri/epri_facts/reportspecs/styleguide.html. Specific guidelines are as follows:

http://physics.nist.gov/Pubs/SP330/sp330.pdf

http://www.epri.com/corporate/discover_epri/epri_facts/reportspecs/styleguide.html

11

Numbers

Use numerals: • For all numbers 10 and above. • In tables and figures. • With units of measure (5 Hz, 12°C) • With percentages (67%) Use words to express: • All numbers zero through nine in text (e.g., “A determination of conductor tension depends

on whether we select three, four, or five beats as being indicative of the traveling wave return time.”)

A zero is ordinarily used before a decimal point: 0.01%

Percentages

Use %, not percent.

Hyphens

Hyphenate compound adjectives containing units of measure and time. a 7.3-m (24-ft) room a 2.54-cm (1-in.) diameter pipe a 3-m pole

In almost all cases, hyphenate high- and low- adjectival compounds. high-voltage transmission

Parallelism

Be aware of parallelism. E.g., if you create a bulleted list in which most lines begin with a verb, start all lines with a verb.

Length

Avoid overly long sentences and paragraphs. Clarity is usually improved by shortening.

Periods/Spaces

Use a single space—not a double space—after the period at the end of a sentence.

Units Spacing

Follow IEEE style. Put one space between number and unit. E.g., 21.1 Hz.

12

Abbreviations

Follow IEEE spelling.

Page Breaks

In first drafts, do not pay attention to page breaks. Do not worry if tables break over two pages. These breaks will be fixed in page layout.

References

References should be cited using the “author-date” method. The advantage of this system is that it avoids the need to track and revise footnote numbers. In text, to cite a source, write the author’s last name and the date of publication in parentheses. For example: (Gyugyi 1988). Two or more authors would be written as: (Hingorani and Gyugyi 2000). If there are more than two authors, use et al. (Edris et al. 1950). If the author has more than one work in one year, use letters to distinguish (Grunbaum 1950a). If the author is an organization, include enough of the name to enable a reader to locate it in the end-of-chapter list. For example: (IEEE Working Group 1978). If there is no author, include the first few words of the title. For example: (EHV Transmission Line 1968). If you wish to specify a page number, put that after the date. For example: (Dalziel 1950, p. 1163). Citations should be inserted within the sentence as close as possible to the point being referenced. At the end of the chapter, provide a list of references alphabetized by the author’s last name. If there is no author, alphabetize by the citation’s title. See examples below for the correct format. Only reference publicly available documents.

Bibliography

If desired, include a Bibliography separate from the References. It should include all major seminal works.

Website References

Website references may be used. In text, include a recognizable short form of the URL. At the end of the chapter, include the full URL.

Reference List Style

Scientific Journal, One Author Gyugyi, L. 1994. “Dynamic Compensation of AC Transmission Lines by Solid-State Synchronous Voltage Sources.” IEEE Transactions on Power Delivery. Vol. 9. No. 2. April.

13

Scientific Journal, Multiple Authors

Enjeti, P. N., P. D. Ziogas, and J. F. Lindsay. 1990. “Programmed PWM Techniques to Eliminate Harmonics: A Critical Evaluation.” IEEE Transactions on Industry Applications. Vol. 26. No. 2. pp. 302-316. March/April.

Conference Paper

Schettler F., H. Huang, and N. Christl. 2000. “HVDC Transmission Systems Using Voltage Sourced Converters—Design and Applications.” IEEE PES Summer Meeting, Seattle, Washington.

Index

When submitting your chapter, create an alphabetical list of key words and section numbers. This will be merged with lists from other chapters.

Keywords in the index will be spelled in adherence with the spelling of the IEEE Dictionary of Standards Terms.

Copyright Issues

If you plan to reproduce figures, photos, or tables from another publication, EPRI will need to seek copyright permissions from the publisher and/or author of the original publication. The Legal Department of EPRI has a specific form and process for obtaining copyright permissions for reproduction of material. Following completion of the first draft, the production team will work with EPRI Legal to obtain the required permissions. As an author, your responsibility is to maintain a clear list of references, showing the source of all reproduced material. Required information for books includes author’s name, title, edition, publisher, year, and page number. Required information for journal articles includes author’s name, article title, journal name, volume and issue numbers, and page numbers. Please note that page numbers are required in both forms, and are required in most publisher’s copyright documentation. Note: • Government Sources. Much material published by government agencies is in the public

domain, and does not require copyright permissions. However, please document all reproduced material, so that the production team can verify each case.

• Author’s Publications. If you are reproducing material from your own earlier publication, we may still need to obtain permission to publish, because, in some cases, the publisher, not the author, holds the copyright.

• Denied Permissions. In some cases, the publisher may deny permission or require an onerous fee or pre-condition. In this event, we may need to create a new figure or table, or paraphrase the information.

14

• Revised Figure. If a figure is substantially changed, you do not need permission from the original source.

• Product Photo. A photo from a utility of a product on a line requires permission from the utility, not the product manufacturer.

Terms

Bibliography. List of documents (books, articles, and websites) related to the subject. (See “References.”) Copyediting. Revision of text to correct spelling and grammar, and to ensure consistency of style, format, and usage. English Units. Units of measure commonly used in North America (e.g., feet, pounds, miles). Page Layout. Placement of the text, tables, and figures on pages in a consistent design, with specifications for type size, line spacing, margins, etc. References. List of documents (books, articles, and websites) cited in the text. (See “Bibliography.”) SI Units. (Le Système International d’Unités, or International System of Units). Metric measurement system, dominant measurement system used in international commerce. Technical Review. Review of draft documents by experts in the content field.

ELECTRIC POWER RESEARCH INSTITUTE 3420 Hillview Avenue, Palo Alto, California 94304-1338 • PO Box 10412, Palo Alto, California 94303-0813 • USA

800.313.3774 • 650.855.2121 • [email protected] • www.epri.com

Export Control Restrictions

Access to and use of EPRI Intellectual Property is granted with the specific understanding and requirement that responsibility for ensuring full compliance with all applicable U.S. and foreign export laws and regulations is being undertaken by you and your company. This includes an obligation to ensure that any individual receiving access hereunder who is not a U.S. citizen or permanent U.S. resident is permitted access under applicable U.S. and foreign export laws and regulations. In the event you are uncertain whether you or your company may lawfully obtain access to this EPRI Intellectual Property, you acknowledge that it is your obligation to consult with your company’s legal counsel to determine whether this access is lawful. Although EPRI may make available on a case-by-case basis an informal assessment of the applicable U.S. export classification for specific EPRI Intellectual Property, you and your company acknowledge that this assessment is solely for informational purposes and not for reliance purposes. You and your company acknowledge that it is still the obligation of you and your company to make your own assessment of the applicable U.S. export classification and ensure compliance accordingly. You and your company understand and acknowledge your obligations to make a prompt report to EPRI and the appropriate authorities regarding any access to or use of EPRI Intellectual Property hereunder that may be in violation of applicable U.S. or foreign export laws or regulations.

The Electric Power Research Institute (EPRI)

The Electric Power Research Institute (EPRI), with major locations in Palo Alto, California, and Charlotte, North Carolina, was established in 1973 as an independent, nonprofit center for public interest energy and environmental research. EPRI brings together members, participants, the Institute’s scientists and engineers, and other leading experts to work collaboratively on solutions to the challenges of electric power. These solutions span nearly every area of electricity generation, delivery, and use, including health, safety, and environment. EPRI’s members represent over 90% of the electricity generated in the United States. International participation represents nearly 15% of EPRI’s total research, development, and demonstration program.

Together…Shaping the Future of Electricity

© 2006 Electric Power Research Institute (EPRI), Inc. All rights reserved. Electric Power Research Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc.

Printed on recycled paper in the United States of America 1012414

epri power electronics-based transmission controllers reference book (the gold book)

Documents