critical pq phenomena and sources of pq disturbances in pe ...€¦ · name company author/s: j....

REPORT

This project has received funding from the European Union’s Horizon 2020

research and innovation programme under grant agreement No 691800.

........................................................................................................................................................... Acronym: MIGRATE – Massive InteGRATion of power Electronic devices Grant Agreement Number: 691800 Horizon 2020 – LCE-6: Transmission Grid and Wholesale Market Funding Scheme: Collaborative Project ...........................................................................................................................................................

Critical PQ phenomena and sources of PQ disturbances in PE rich power

systems

Deliverable 5.1

Date: 19.12.2016 Contact: [email protected]

REPORT

of 158

Disclaimer The information, documentation and figures in this deliverable are written by the MIGRATE project consortium under EC grant agreement No 691800 and do not necessarily reflect the views of the European Commission. The European Commission is not liable for any use that may be made of the information contained herein. Dissemination Level: Public X

Restricted to other programme participants (including the Commission Services) Restricted to bodies determined by the MIGRATE project Confidential to MIGRATE project and Commission Services

REPORT

of 158

Document info sheet Document name: Critical PQ phenomena and sources of PQ disturbances in PE rich

power systems

Responsible partner: Tallinn University of Technology (TUT)

WP: 5

Task: 5.1

Deliverable number: D5.1

Revision: 1.0

Revision date: 19.12.2016

Name Company Name Company

Author/s:

J. Kilter Elering S. Abdelrahman UoM1

M. Löper Elering J. Milanovič UoM

T. Kangro TUT S. Schilling TUB2

I. Palu TUT K. Strunz TUB

M. Val Escudero EirGrid A. Božiček UL3

C. Buchhagen TenneT B. Blažič UL

J. Kosmač ELES D. Matvoz EIMV

J. Žvab ELES A. Souvent EIMV

Task leader: T. Kangro TUT

WP leader: J. Kosmač ELES

Revision history log Revision Date of release Author

Summary of

changes

0.1 - Draft 03.08.2016 TUT Initial Draft

0.2 - Draft 28.09.2016 TUT Comments from WP5

partners.

1.0 19.12.2016 TUT

Comments from

Executive Board and

WP5 partners

1 The University of Manchester, UK 2 Technische Universität Berlin, Germany 3 University of Ljubljana, Slovenia

REPORT

of 158

CONTENT

EXECUTIVE SUMMARY ............................................................................... 12

Index of Terms and Abbreviations .............................................................. 15

1 Introduction ....................................................................................... 17

1.1 Background ................................................................................. 17

1.2 PQ Phenomenon and Principles ...................................................... 17

1.3 PQ Standards and Technical Recommendations ............................... 20

1.4 Overview of the Report ................................................................. 22

2 PQ Disturbance Categorization .............................................................. 23

2.1 Introduction ................................................................................ 23

2.2 Frequency Deviation .................................................................... 25

2.2.1 Background ...................................................................... 25

2.2.2 Evaluation Indices ............................................................. 25

2.2.3 Consequences .................................................................. 26

2.2.4 Mitigation Methods ............................................................ 28

2.2.5 Practical Studies ............................................................... 29

2.3 Steady-State Voltage Variation ...................................................... 31

2.3.1 Background ...................................................................... 31


2.3.3 Consequences .................................................................. 33



2.4 Voltage Dip and Temporary Power Frequency Over-voltage ............... 36

2.4.1 Background ...................................................................... 36


2.4.3 Consequences .................................................................. 41


REPORT

of 158


2.5 Voltage Fluctuation and Flicker ...................................................... 46

2.5.1 Background ...................................................................... 46


2.5.3 Consequences .................................................................. 49



2.6 Voltage Unbalance ....................................................................... 52

2.6.1 Background ...................................................................... 52


2.6.3 Consequences .................................................................. 56



2.7 Harmonic Distortion ..................................................................... 61

2.7.1 Background ...................................................................... 61


2.7.3 Consequences .................................................................. 66



2.8 Harmonic Stability ....................................................................... 72

2.8.1 Background ...................................................................... 72


2.8.3 Consequences .................................................................. 76



3 PQ in Transmission Systems ................................................................. 77

3.1 TenneT ....................................................................................... 77

3.1.1 Introduction ..................................................................... 77

3.1.2 Generation ....................................................................... 78

REPORT

of 158

3.1.3 Consumption .................................................................... 79

3.1.4 PQ Overview .................................................................... 79

3.1.5 PQ Monitoring Systems ...................................................... 81

3.2 ELES .......................................................................................... 81

3.2.1 Introduction ..................................................................... 81

3.2.2 Generation ....................................................................... 83

3.2.3 Consumption .................................................................... 83

3.2.4 PQ Overview .................................................................... 83


3.3 Elering ....................................................................................... 91

3.3.1 Introduction ..................................................................... 91

3.3.2 Generation ....................................................................... 93

3.3.3 Consumption .................................................................... 94

3.3.4 PQ Overview .................................................................... 95


3.4 EirGrid Group .............................................................................. 99

3.4.1 Introduction ..................................................................... 99

3.4.2 Generation ..................................................................... 101

3.4.3 Consumption .................................................................. 102

3.4.4 PQ Overview .................................................................. 103

3.4.5 PQ Monitoring System ..................................................... 105

3.5 Questionnaire ............................................................................ 106

3.5.1 Introduction ................................................................... 106

3.5.2 The Survey .................................................................... 106

3.5.3 The Results of the Survey ................................................ 107

3.5.4 Conclusions .................................................................... 115

4 PQ Monitoring and Assessment ........................................................... 117

4.1 PQ Monitoring ............................................................................. 117

4.1.1 Introduction ................................................................... 117

REPORT

of 158

4.1.2 Objectives for PQ Monitoring ............................................ 117

4.1.3 Selection of Monitoring Locations ...................................... 118

4.1.4 Duration of Monitoring ..................................................... 119

4.1.5 Selection of Monitoring Parameters ................................... 119

4.1.6 Selection of Monitoring Equipment .................................... 120

4.2 PQ Assessment .......................................................................... 123

4.2.1 Introduction ................................................................... 123

4.2.2 PQ Reporting .................................................................. 124

4.2.3 Compliance Verification .................................................... 125

4.2.4 Performance Analysis/Benchmarking ................................. 126

4.2.5 Site Characterization ....................................................... 127

4.2.6 Troubleshooting .............................................................. 128

4.2.7 Advanced Application and Studies ..................................... 129

4.2.8 Active PQ Management .................................................... 129

4.3 PMU Measurements .................................................................... 130

4.3.1 Introduction ................................................................... 130

4.3.2 Standards for PMU Measurements ..................................... 132

4.3.3 Testing of PMUs .............................................................. 133

4.3.4 Results .......................................................................... 136

5 Conclusion and Future Work ............................................................... 141

5.1 Conclusion ................................................................................ 141

5.2 Future Work .............................................................................. 142

References ............................................................................................ 145

REPORT

of 158

List of Figures Figure 1-1: Causes of PQ problems on customer and utility side [1] ......................................... 18

Figure 1-2: Overview of several PQ definitions [1] .................................................................. 19

Figure 1-3: General PQ evaluation and mitigation process [1] .................................................. 19

Figure 2-1: Average frequency values in Continental Europe, June 2003 and June 2010 [28] ....... 30

Figure 2-2: Frequency quality behaviour in Continental Europe during years 2001-2011 [28] ...... 30

Figure 2-3: Voltage dip and temporary power frequency over-voltage phenomenon .................... 37

Figure 2-4: FRT criteria by EU regulation [65] ........................................................................ 39

Figure 2-5: Voltage dips at 110 kV substations as reported in [105] ......................................... 44

Figure 2-6: Waveform and RMS voltage performance of the three phases due to fault at 130 kV

[32] .................................................................................................................................. 46

Figure 2-7: Measurement data for flicker Pst95 at MV, HV and EHV [62] ..................................... 51

Figure 2-8: Measurement data for flicker Plt95 at MV, HV and EHV [62] ...................................... 51

Figure 2-9: Measurement data for flicker Pst99 at MV, HV and EHV [62] ..................................... 52

Figure 2-10: Low order harmonic voltages from 7 surveys at HV for Vhsh95 [62].......................... 72

Figure 2-11: Frequency analysis of the current at the AC side [216] ......................................... 75

Figure 3-1: TenneT grid in Netherlands and Germany in 2015 .................................................. 78

Figure 3-2: Steady-state harmonic in offshore grid ................................................................. 80

Figure 3-3: Voltage (155 kV, L1) during harmonic instability in offshore grid ............................. 81

Figure 3-4: Voltage (155 kV, L1) during harmonic instability in offshore grid in 0.2-0.35-s time

frame ............................................................................................................................... 81

Figure 3-5: Slovenian 400 kV and 220 kV transmission network ............................................... 82

Figure 3-6: Trend lines of VRMS at 400 kV voltage level ............................................................ 84

Figure 3-7: Trend lines of VRMS at 110 kV voltage level ............................................................ 84

Figure 3-8: Trend lines of Pst at 400 kV voltage level .............................................................. 85

Figure 3-9: Trend lines of Pst at 110 kV voltage level .............................................................. 86

Figure 3-10: Weekly average profile of Pst at 400 kV voltage level ............................................ 86

Figure 3-11: Weekly average profile of Pst at 110 kV voltage level ............................................ 87

Figure 3-12: Daily average profile of Pst at 400 kV voltage level ............................................... 87

Figure 3-13: Daily average profile of Pst at 110 kV voltage level ............................................... 88

Figure 3-14: Trend lines of THD at 400 kV voltage level .......................................................... 89

Figure 3-15: Trend lines of THD at 110 kV voltage level .......................................................... 89

Figure 3-16: ELES PQ monitoring system .............................................................................. 91

Figure 3-17: Estonian transmission network .......................................................................... 92

Figure 3-18: Power plants in Estonia ..................................................................................... 93

Figure 3-19: Energy consumption in Estonia .......................................................................... 94

Figure 3-20: Comparison between energy consumption and generation in Estonia ...................... 94

Figure 3-21: Seasonal variations of 99% voltage measurement values at 110 kV level ............... 95

Figure 3-22: Seasonal variations of 99% asymmetry ka values at 110 kV level .......................... 96

Figure 3-23: Seasonal variation of 95% Plt values at 110 kV level ............................................ 97

Figure 3-24: Seasonal variation of 99% THD values at 110 kV level ......................................... 98

REPORT

of 158

Figure 3-25: 95% values of 2nd-13th voltage harmonics (as a fraction of normal voltage; blue line

expresses allowed levels) during years 2010-2013 at 110 kV level ........................................... 98

Figure 3-26: Transmission system in the isle of Ireland ......................................................... 100

Figure 3-27: Expected growth in wind capacity in Ireland ...................................................... 102

Figure 3-28: Observed trend in harmonic voltage distortion in the Irish transmission system (2014-

2015) ............................................................................................................................. 103

Figure 3-29: Harmonic voltage distortion measurements (10-min average values) before and after

underground cable connection ........................................................................................... 104

Figure 3-30: Harmonic voltage distortion measurements (95 percentile values) before and after

underground cable connection ........................................................................................... 104

Figure 3-31: PQ monitoring devices in Ireland ..................................................................... 105

Figure 3-32: Response to Q0 – What are your system characteristics with respect to generation

mix and type of circuits? ................................................................................................... 108

Figure 3-33: Response to Q1 – Are you monitoring power quality in your network? .................. 108

Figure 3-34: Responses to Q3 – Why do you to the monitoring? ............................................. 109

Figure 3-35: Responses to Q2 – How many PQ monitors do you have? ................................... 109

Figure 3-36: Responses to Q6 – What types of transducers are used for PQ monitoring? ........... 110

Figure 3-37: Responses to Q9 – Do you use of plan to use PMUs for PQ monitoring? ................ 110

Figure 3-38: Responses to Q9 follow-up question – What do you monitor or plan to monitor? .... 111

Figure 3-39: Responses to Q10 – How many PMUs in total do you have installed in your network at

different voltage levels? .................................................................................................... 111

Figure 3-40: Responses to Q4 – Which aspects of PQ do you monitor? .................................... 112

Figure 3-41: Responses to Q5 – What electrical parameters do you monitor? ........................... 112

Figure 3-42: Responses to Q7 – What are the main issues of PQ in your network? ................... 113

Figure 3-43: Responses to Q8 – What are the main sources that cause PQ issues in your network?

...................................................................................................................................... 113

Figure 3-44: Responses to Q11 – Do you use power quality mitigation in your network? ........... 114

Figure 3-45: Responses to Q11 – Do you use power quality mitigation in your network? What

devices are used for power quality mitigation? ..................................................................... 114

Figure 3-46: Comparison of responses to Q12 – How many of each of the following PQ mitigation

devices do you have in your network? To which voltage level these devices are installed and what

is their typical size? .......................................................................................................... 115

Figure 4-1: Voltage transformer technologies frequency range [221] ...................................... 122

Figure 4-2: Current transformer technologies frequency range [221] ...................................... 122

Figure 4-3: Trend graph including percentile values [3]......................................................... 124

Figure 4-4: Example of a histogram for continuous disturbances in single site [3] .................... 126

Figure 4-5: The bar chart of disturbance levels at sites [3] .................................................... 127

Figure 4-6: Distribution of site disturbance levels [3] ............................................................ 127

Figure 4-7: Time varying disturbance trend plot [3] .............................................................. 128

Figure 4-8: Sinusoidal signal phasor representation; (a) sinusoidal signal, (b) phasor

representation [230] ........................................................................................................ 131

Figure 4-9: General scheme of PMU [231] ........................................................................... 131

REPORT

of 158

Figure 4-10: PMU phasor signal processing model [232] ....................................................... 132

Figure 4-11: PMUs testing scheme...................................................................................... 134

Figure 4-12: Harmonic content in Test B Case 1; CP95 – harmonic value 95% of measurement

period, CP99 – harmonic value 99% of measurement period, CPMax – harmonic maximum value

...................................................................................................................................... 135

Figure 4-13: Average voltage TVE ...................................................................................... 136

Figure 4-14: Average current TVE ...................................................................................... 137

Figure 4-15: Current (left) and voltage (right) harmonics TVE in Case 1.................................. 138

Figure 4-16: Current (left) and voltage (right) harmonics TVE in Case 2.................................. 138

Figure 4-17: 75 Hz current (left) and voltage (right) harmonics TVE with first set of filters ........ 139

Figure 4-18: 75 Hz current (left) and voltage (right) harmonics TVE with second set of filters .... 139

REPORT

of 158

List of Tables Table 2-1: PQ disturbance categorization .............................................................................. 24

Table 2-2: Categories and typical characteristics of steady-state voltage RMS [35] .................... 32

Table 2-3: Voltage ranges for reference voltages defined by TSOs at 110-300 kV [30] ................ 32

Table 2-4: Voltage ranges for reference voltages defined by TSOs at 300-400 kV [30] ................ 32

Table 2-5: Expected number of voltage dips for some substations [106] ................................... 45

Table 2-6: Indicative values of planning levels for flicker in MV, HV and EHV [72] ...................... 48

Table 2-7: Indicative planning levels for harmonic voltages (in percent of the fundamental

component) in HV and EHV [139] ......................................................................................... 65

Table 2-8: Magnitudes of measured harmonics [165].............................................................. 72

Table 3-1: Key figures of TenneT in Netherlands and Germany in 2015 ..................................... 77

Table 3-2: Energy generation in Germany year 2015 .............................................................. 79

Table 3-3: Voltage harmonic trends for the Slovenian 400 kV network ...................................... 90

Table 3-4: Voltage harmonic trends for the Slovenian 110 kV network ...................................... 90

Table 3-5: Energy consumption and generation in Estonia ....................................................... 94

Table 3-6: Total length of transmission circuits .................................................................... 101

Table 3-7: Total transformer MVA capacity .......................................................................... 101

Table 3-8: Reactive power compensation devices ................................................................. 101

Table 3-9: All-island demand ............................................................................................. 103

Table 3-10: Survey questions ............................................................................................ 107

Table 4-1: Appropriate timescales for monitoring different PQ phenomena [3] ......................... 119

Table 4-2: PQ disturbances and measurement intervals [221] ............................................... 121

Table 4-3: Transformer parameters influencing PQ measurement [221] .................................. 121

Table 4-4: PQ monitoring data presentation approaches ........................................................ 123

Table 4-5: Presenting tabulated statistical data [3] ............................................................... 128

Table 4-6: Harmonic content in Test B Case 1 and Case 2 ..................................................... 135

Table 4-7: TVE change in percentiles with different interharmonics ......................................... 140

REPORT

of 158

EXECUTIVE SUMMARY There has been a significant increase of interest in power quality (PQ) in modern transmission networks. This topic has been historically more related to the distribution level but constant increase of power electronic (PE) based technology over the last few decades has given the push to look into this topic also on transmission level. Due to higher penetration levels of PEs several issues have been encountered by many European Transmission Systems Operators (TSOs). These are related to the modified dynamic behaviour of the power system, interactions between the controllers of PE, weaknesses of existing protection schemes and power quality. Increasing penetration of PE-interfaced wind generators (and to some extent PV) is detrimental to PQ, when, at the same time, the use of storage technologies or HVDC lines may impact the PQ performance of such networks. PE devices are indeed one of the major sources of network disturbances (harmonics, flicker, voltage and frequency variations), while also being very sensitive themselves to PQ disturbances. In a response to this increased interest in PQ, a separate work package (WP), which deals only with PQ related questions, was formed within the MIGRATE project. Activities in MIGRATE WP5 should be looked at together with scenarios developed in MIGRATE WP1. WP5 addresses the following objectives:

• to develop and validate a focused set of simulation models of PE devices for PQ studies in line with the scenarios studied in WP1, which allows evaluating their influence on PQ in transmission networks;

• to identify critical PQ disturbances in order to numerically study their propagation through transmission networks in line with the retained WP1 scenarios (different levels of PE penetration and different network operating conditions)

• to examine how far the use of PMUs can assist in the mapping of PQ issues, including the implementation of real-time data retrieval;

• to analyse the impacts of PQ disturbances on the performance of PE devices; • to combine the above findings into a methodology and numerical models which can be used

by TSOs to assess PQ issues in future power systems experiencing WP1-like scenarios and to propose appropriate, cost-effective mitigating solutions to reach a required level of PQ.

In general, the objective is to evaluate the PQ levels and propagation in future PE-rich power networks and to develop methodologies for cost-effective mitigation options by TSOs, in order to keep PQ levels within affordable values. In order to solve these challenging objectives with respect to future power systems, the work in WP5 has been divided into five parts. These are as follows:

• identification of critical PQ phenomena and sources of PQ disturbances (task 5.1); • development of PE numerical simulation models for PQ studies (task 5.2); • propagation of PQ disturbances through power networks (task 5.3); • assessment of the influence of PQ disturbances on operation of PE rich power networks (task

5.4); • mitigation of PQ disturbances and provisions of differentiated PQ (task 5.5).

The first of these objectives has been the task for WP5 team for 2016. The remaining parts are covered in the work from 2017 to 2019.

REPORT

of 158

This report covers the first part of WP5 work and deals with identifying critical PQ phenomena and sources of PQ disturbances. Critical PQ issues in power systems (harmonics, flicker, voltage and frequency variations) are analysed including determining the contribution of different network components (PE-interfaced wind and PV generation, thyristor-controlled devices, VSC, FACTS, etc.) to the PQ disturbances. The report includes critical assessment of PQ issues in transmission power networks considering real-world issues as well as studies published by research laboratories. PQ disturbances are categorized by cause, type (such as harmonics, frequency variations, voltage-level violations), and reported impacts while considering the temporal variation of the level of disturbances due to renewable generation and load. Influence of harmonics on PMU accuracy (in particular, if the PMUs are installed close to sources of harmonics and inter-harmonics, e.g. generation buses with PE-connected generation), has been evaluated. Moreover, assessment and recommendations for network measurements (e.g. harmonics phase-angles) so as to capture on different PQ phenomena are provided. The work has been led by Tallinn University of Technology with the support from other academic partners (The University of Manchester, Technical University of Berlin, and University of Ljubljana) and from TSOs (ELES, Elering, TenneT and EirGrid). This report is divided into five parts. The first chapter provides a generalised overview on the background of PQ in modern power systems, e.g. what is the definition of PQ, what phenomena can be considered as PQ, and basic evaluation principles together with mitigation possibilities. A discussion from the TSO perspective is given in respect of defining the planning levels for PQ for transmissions systems and specific limits for customer. This, in most cases, is not straightforward and depends on TSOs understandings and principles. Currently there is no commonly agreed method to define the PQ limits for transmission network and for its customers. Nevertheless, there are documents, e.g. EN 50160, IEC 61000-3-6, IEC 61000-3-7, IEC 61000-4-30, CIGRE Technical Brochures 261, 467, 596, and 5th CEER Benchmarking Report on the Quality of Electricity Supply, available, which cover this item in quite sufficient manner and provide good basis for the TSOs to select and define PQ parameters. Following this, it summarises the previous work available in the literature and including real-life experience from TSOs with respect to various PQ phenomena. The discussion includes known PQ phenomena including frequency deviation, steady-state voltage variation, voltage dips and temporary power frequency over-voltage (in the literature could also be referenced as voltage swells), voltage fluctuations and flicker, voltage unbalance, and harmonic distortion. Various aspects, including background, evaluation indices, consequences, mitigation methods and practical studies, are covered. Emphasis has been on transmission network related aspects and studies considering the increasing level of PE. In addition to the known PQ phenomena, the report includes discussion on relatively new topic for the TSOs – harmonic resonance. This issue has become relevant to the TSOs lately after the level of PE based generation has increased significantly. Here, the harmonic impedance and emissions from multiple PE devices can interact with the harmonic impedance of the supply system leading to excitation of harmonic resonances and instability of the wideband controllers. As a conclusion, results of an analysis with respect to contribution to PQ disturbances by PE devices and the effect of these PQ disturbances on the performance of the PE devices is presented. Based on the analysis it can be stated that PQ in future power systems is relevant and there are various items that must be carefully considered when moving towards higher penetration levels of PE.

REPORT

of 158

The report continues with real life examples from TSOs. Partner TSOs in this MIGRATE WP have composed an overview and a discussion about their systems and various PQ challenges in their networks. This discussion is elaborated further using the results from a survey on PQ issues in current transmission networks. As one of the first task of WP5, a questionnaire on a range of PQ issues in current transmission networks was developed and distributed to more than 30 European TSOs. The key findings of the questionnaire, based on 23 responses, are used to affirm current TSO experiences with PQ phenomena. It has been established that at present most of the TSOs monitor PQ in their network to ensure compliance against standards. Furthermore they record events to compile statistical information and to assess the individual events. The results also indicate that the PMUs are used as one additional option for PQ monitoring and assessment mainly concentrating on frequency and voltage issues. A discussion related to PQ monitoring and assessment is given next for the purpose of providing background about how exactly PQ should be monitored and assessed in transmission networks and what are the parameters that influence the results. For successful PQ monitoring it is essential to define PQ monitoring objectives, select suitable monitoring locations, define the duration of monitoring campaign, select sufficient number of monitoring parameters, and select appropriate monitoring equipment (this is especially important in transmission networks). The next step is to assess the information and define appropriately the following steps for mitigating existing PQ issues. When assessing PQ in transmission networks different aspects have to be considered. It is important to understand and define the purpose and objectives of the analysis. Most common ones are compliance verification, performance analysis and benchmarking, site characterization, troubleshooting, advanced applications and studies, and active PQ management. A separate section has been devoted to PMUs and their applicability for PQ monitoring with respect to harmonics. Based on conducted studies and analysis it has been evaluated that harmonics do not have significant influence on PMU accuracy. Nevertheless, further assessment on this item is needed in order to consider also dynamic criteria defined in relevant standards. Finally, the report clarifies areas and directions for further research and development in the scope of WP5 of the MIGRATE project. These include activities related to modelling, propagation, assessment and mitigation of PQ in future transmission networks. Each of these parts will be led by different academic partner and supported strongly by all involved TSOs. All specific tasks declared for MIGRATE deliverable 5.1 have been completed. Additionally, an international survey on TSO practice with PQ issues, involving 23 European TSOs, has been carried out and its findings summarised in the report. The survey`s results complement the work and serve as a consolidated summary of the present PQ issues in the industry. Moreover, the findings delivered in this report are planned to be used for publications in international journals and conferences.

REPORT

of 158

Index of Terms and Abbreviations

All terms and abbreviations provided in this section are for the purpose of this Deliverable 5.1.

AC Alternating Current

ACM Authority for Consumers & Markets (Netherland)

A/D Analogue/Digital

ANSI American National Standards Institute

ASD Adjustable Speed Drive

CBEMA Computer and Business Equipment Manufacturers' Association

CCGT Combined-Cycle Gas Turbine

CDS Closed Distribution Systems

CEER Council of European Energy Regulators

CER Commission for Energy Regulation

CHP Combined Heat and Power

CIGRE International Council on Large Electric System (Conseil International des Grands

Réseaux Electriques)

CIRED International Conference on Electricity Distribution (Congrés International des

Réseaux Electriques de Distribution)

CT Current Transformer

DC Direct Current

DCMLI Diode Clamped Capacitor Multilevel Inverter

DFIG Doubly Fed Induction Generator

DSO Distribution System Operator

DVR Dynamic Voltage Restorer

EHV Extra High Voltage

EMC Electromagnetic compatibility

EMT Electromagnetic Transient

ENTSO-E European Network of Transmission System Operators for Electricity

EWIC East West Interconnector (HVDC link between Ireland and Great Britain)

FACTS Flexible AC Transmission System

FRT Fault-ride-through

GPS Global Positioning System

HPP Hydro Power Plant

HV High Voltage

HVDC High Voltage Direct Current

IDVR Interline Dynamic Voltage Restorer

IEC International Electrotechnical Commission

IEEE Institute of Electrical and Electronics Engineering

LCC Line-commutated Converter

MMC Modular Multilevel Converter

REPORT

of 158

NEC National Electricity Code (Australia)

NEMA National Electrical Manufacturers Association

NGO Non-Governmental Organisation

NRS National Standard (South-Africa)

NEW North West European

PDC Phasor Data Concentrator

PE Power Electronic

PMU Phasor Measurement Units

PQ Power Quality

PQMS PQ Monitoring System

PTP Precision Time Protocol

PV Photovoltaics

PWM Pulse Width Modulation

ROCOF Rate of Change of Frequency

RC-divider Resistive-Capacitive Divider

RES Renewable Energy Sources

RTE Reseau de Transport d’Electricite (French TSO)

RVC Rapid Voltage Changes

SARFI System Average RMS Variation Frequency Index

SCADA Supervisory Control and Data Acquisition

SHE Selective Harmonic Elimination

SSC Static Series Compensator

STATCOM Static Synchronous Compensator

SVC Static var Compensator

TCP/IP Transmission Control Protocol/Internet Protocol

TDD Total Demand Distortion

THD Total Harmonic Distortion

TSO Transmission System Operator

TVE Total Vector Error

UPS Uninterrupted Power Supply

UTC Coordinated Universal Time

VSC Voltage Source Converter

VT Voltage Transformer

VUF Voltage Unbalance Factor

WPP Wind Power Plant

WTG Wind Turbine Generator

REPORT

of 158

1 Introduction

1.1 Background The aim of the MIGRATE (Massive InteGRATion of power Electronic devices) project is to help the pan-European Transmission System Operators to anticipate and adjust progressively to the impacts of increasing penetration of power electronics (PE) onto the power system operation, with an emphasis on power system stability, the relevance of existing protection schemes and the expected degradation of power quality (PQ). The project is organised in a number of Work Packages (WP) dealing with each individual technical issue. MIGRATE WP5 is dedicated to investigate power quality in transmission networks with high PE penetration. The increasing penetration of PE-interface renewables has already resulted in PQ challenges as evidenced by harmonic distortion, voltage sags and other disturbances. PE devices are one of the major sources of PQ disturbances in power systems but they are also very sensitive to PQ disturbances themselves. The main objective for WP5 Task 5.1 is to identify critical PQ phenomena and sources of PQ disturbances in order to numerically study their propagation through transmission networks. Task 5.1 involves the assessment of PQ issues in transmission networks considering real-world issues as well as studies published by research laboratories, categorization of the types of PQ disturbances (including causes, evaluation indices, consequences and mitigation methods) and evaluation of the influence of harmonics on Phasor Measurement Unit (PMU) accuracy. The focus of this work is on impact of PE devices on transmission systems. However, in some cases the aspects on distribution level are also pointed out for comparative purpose and to evaluate PQ disturbance transfer to higher voltage levels.

1.2 PQ Phenomenon and Principles PQ covers a wide range of disturbances and deviations in voltage magnitude or waveform shape from the nominal values. The criticality of PQ is directly related to the sensitivity/resilience of customer and transmission system equipment to those deviations, as well as the consequences of failures. In general, PQ can be defined as the characteristics of the power supply required to ensure that all equipment works properly and efficiently [1]. The Council of European Energy Regulators (CEER) considers the assessment of the impact on customers as the main reason for monitoring and regulating PQ. As it is stated in [2]: “The ultimate aim of voltage quality regulation is to ensure that the functioning of equipment is not impacted by voltage disturbances coming from the network”. Therefore, when analysing PQ issues the main focus should be on compatibility between the customers’ equipment and different PQ phenomena, e.g. voltage dips, unbalance, harmonics, flicker, transients, over-voltages etc.

REPORT

of 158

Determination of the exact cause of PQ problems is not always straightforward. Similarly, identification of stakeholders’ responsibilities in the area of PQ can be challenging and lead to different interpretations. As an example, Figure 1-1 shows the results of a survey performed by the Georgia Power Company between customers and utilities aimed to identify and understand the different perceptions related to PQ problems. The results suggest similar perception from both customers and utilities on the dominant cause of PQ problems (i.e. natural causes). However, the survey results highlighted opposing perceptions when it came to identifying responsibility for non-natural causes of PQ problems [1]. Common trend nowadays is to monitor and measure different electromagnetic phenomena under the term of PQ. Some regulators are becoming to apply the PQ standards. The CEER report [2] and CIGRE brochure [3] showed the increasing adoption of monitoring different PQ phenomena in the European countries.

Figure 1-1: Causes of PQ problems on customer and utility side [1]

PQ phenomena can be divided into two major categories: disturbances and steady-state variation [1]. The disturbances include phenomena like the transients, voltage dips and temporary power frequency over-voltages or interruption of supply. These phenomena cannot be measured in a traditional continuous manner as they occur only momentary during long periods of time. The assessment of such PQ phenomena is usually based on the count of occurrences during a relatively long time-period. On the other hand, the steady-state variation category includes phenomena like voltage variation, voltage unbalance, harmonics distortion and voltage flicker. These types of phenomena persist for a long time. Therefore, continuous monitoring is needed. The performance of PQ in such phenomena could be measured and benchmarked for very short time-periods (seconds). In [1] the PQ phenomena is divided into more categories:

• Transients category, which includes the impulsive and the oscillatory phenomena; • Short-duration variations category, which includes the interruptions, dips and temporary

power frequency over-voltages; this category can be further divided based on duration i.e. instantaneous (cycles), momentary (seconds) or temporary (up to 1-min);

• Long-duration variations category, which includes sustained interruptions, under-voltages and over-voltages;

• Steady-state category, which includes phenomena like voltage unbalance and harmonics. Sometimes additional characteristics are needed to classify the measurements and fully describe electromagnetic phenomena that lead to PQ problems [1]. Characteristics like amplitude, frequency, spectrum, and source impedance are required to describe steady-state phenomena, and characteristics like rate of rise, duration, rate of occurrence are needed to describe non-steady-state

REPORT

of 158

phenomena. Figure 1-2 shows an overview of some of the PQ phenomena definitions, illustrated in a voltage root-mean-square value (RMS) plot.

Figure 1-2: Overview of several PQ definitions [1]

In spite of the variety of PQ phenomena, general steps can be applied in any PQ evaluation process. Based on the available information and measurements, identifying the causes of disturbances and available solutions are the steps of the general process. Interaction between customers and utilities is critical in evaluation process. The general PQ evaluation and mitigation process is summarised in Figure 1-3 [1].

Figure 1-3: General PQ evaluation and mitigation process [1]

REPORT

of 158

The PQ solutions can range from equipment based, to plant based, to utility based solutions. It is obvious that the mitigation cost increases with the increase of the mitigation solution size and coverage. For example, the utility PQ mitigation solutions could benefit a number of neighbouring buses, however improving equipment specifications and immunity or utilising a local solution could be more attractive and optimal in some applications. The sensitivity of loads and expected levels and types of PQ disturbances must be considered when optimising PQ mitigation solutions. The increasing amount of grid-integrated renewable energy sources and their associated power electronic devices have a major impact on the PQ of the electric network. Furthermore, the usage of storage units as well as high voltage direct current (HVDC) connections within the AC transmission networks affects the PQ performance of the network. Besides the PQ disturbances caused by the operational behaviour of power electronic devices, the devices themselves are sensitive against PQ disturbances. This means, that the operational behaviour of power electronic devices can be negatively influenced by other PQ disturbance sources. Due to increasing concerns about PQ issues associated with PE devices and the commitments made to integrate high levels of renewable generation, a common trend nowadays is to monitor and measure different electromagnetic phenomena under the term of PQ. The CEER report [2] and CIGRE brochure [3] showed the increasing adoption of monitoring different PQ phenomena in the European countries.

1.3 PQ Standards and Technical Recommendations Currently there is no harmonised approach to PQ management amongst the European TSOs. Due to the limitation of the European Standard EN50160, which does not adequately cover transmission voltage levels, different PQ indices are applied in each country. Furthermore, the practice of allocating emission limits to customer connections differs significantly from country to country. While some TSOs have fixed emission limits defined in their Grid Codes (for example REE and RTE), other TSOs allocate emission limits on a case-by-case basis following detailed technical studies (for example EirGrid, National Grid and EnerGinet). Some countries (Ireland and Denmark) follow the IEC 61000 recommendations for “equal rights” apportioning of harmonic capacity while, in contrast, UK has adopted the “first-come-first-serve” allocation principle described in G5/4. This lack of consistency and gap in regulation has already been identified by the Council of European Energy Regulators (CEER), who published a PQ benchmark report in 2011 [2] with the following recommendation - “Further improve EN 50160 as a harmonised instrument for voltage quality regulation”:

• An effective extension to the high voltage networks (with effective limits and requirements) and the consideration of extra high voltage networks;

• The adoption of new limits for supply voltage variations in distribution networks (especially in low voltage networks);

• The introduction of limits for voltage events, taking into account the different characteristics of the European networks; one or more responsibility-sharing curves should be defined for voltage dips and voltage swells;

• A general framework for sharing the voltage quality responsibilities between network companies, equipment manufacturers and users.

A list of the most relevant standards and technical recommendations dealing with PQ is provided below for reference:

REPORT

of 158

EN 50160 – Voltage characteristics of electricity supplied by public distribution systems (2010) IEC 61000-2-4 – Electromagnetic compatibility (EMC) - Part 2-4: Environment - Compatibility levels in industrial plants for low-frequency conducted disturbances (2002) IEC 61000-2-12 – Electromagnetic compatibility (EMC) - Part 2-12: Environment - Compatibility levels for low-frequency conducted disturbances and signalling in public medium-voltage power supply systems (2003) IEC 61000-3-3 – Electromagnetic compatibility (EMC) - Part 3-3: Limits - Limitation of voltage changes, voltage fluctuations and flicker in public low-voltage supply systems, for equipment with rated current ≤16 A per phase and not subject to conditional connection (2013) IEC 61000-4-30 – Electromagnetic compatibility (EMC) - Part 4-30: Testing and measurement techniques - Power quality measurement methods (2015) IEC TR 61000-2-1 – Electromagnetic compatibility (EMC) - Part 2: Environment - Section 1: Description of the environment - Electromagnetic environment for low-frequency conducted disturbances and signalling in public power supply systems. IEC TR 61000-2-8 – Electromagnetic compatibility (EMC) - Part 2-8: Environment - Voltage dips and short interruptions on public electric power supply systems with statistical measurement results (2002) IEC TR 61000-2-14 – Electromagnetic compatibility (EMC) - Part 2-14: Environment - Overvoltages on public electricity distribution networks (2006) IEC TR 61000-3-6 – Electromagnetic compatibility (EMC) - Part 3-6: Limits - Assessment of emission limits for the connection of distorting installations to MV, HV and EHV power system (2008) IEC TR 61000-3-7 – Electromagnetic compatibility (EMC) – Part 3-7: Limits - Assessment of emission limits for the connection of fluctuating installations to MV, HV and EHV power systems (2008) IEC TR 61000-3-13 – Electromagnetic compatibility (EMC) – Part 3-13: Limits - Assessment of emission limits for the connection of unbalanced installations to MV, HV and EHV power systems (2008) IEC TR 61869-103 – Instrument transformers - The use of instrument transformers for power quality measurement (2012) IEEE 1159 – IEEE Recommended Practice for Monitoring Electric Power Quality (2009) IEEE 1564 – IEEE Guide for Voltage Sag Indices (2014) IEEE 493 – IEEE Recommended Practice for the Design of Reliable Industrial and Commercial Power Systems (2007) IEEE 241 – IEEE Recommended Practice for Electric Power Systems in Commercial Buildings (1990) IEEE 519 – IEEE Recommended Practice and Requirements for Harmonic Control in Electric Power Systems (2014) ANSI C84.1 – Electric Power Systems and Equipment - Voltage Ratings (60 Hertz) (2006) NRS 048-2 – Electric Supply – Quality of supply – Part 2: Voltage characteristics, compatibility levels, limits and assessment methods (2003) CIGRE TB 261 – Power Quality Indices and Objectives (JWG C4.07, 2004) CIGRE TB 372 – Voltage Dip Evaluation and Prediction Tools (TF C4.102, 2009)

REPORT

of 158

CIGRE/CIRED TB 467 – Economic Framework for Power Quality (JWG C4.107, 2011) CIGRE/CIRED TB 596 – Guidelines for Power Quality - Measurement Locations, Processing and Presentation of Data (C4.112, 2014) CEER – 5th CEER Benchmarking Report on The Quality of Electricity Supply (2011) ER G5/4 Planning levels for harmonic voltage distortion and the connection of non-linear equipment to transmission systems and distribution networks in the United Kingdom (2011) ENTSO-E/Eurelectric Report – Deterministic frequency deviations – root causes and proposals for potential solutions (2011)

1.4 Overview of the Report This report is structured into five chapters. In addition to this introductory chapter, the outline of the other chapters is described below. Chapter 2 provides a comprehensive assessment of PQ phenomena in power systems as well as a review of studies published by research institutions. Emphasis has been on transmission network related aspects and studies considering the increasing level of PE. Chapter 3 provides an overview of the PQ phenomena experienced by four European TSOs as well as a description of their PQ monitoring practices and associated systems. In addition, a questionnaire on range of issues of PQ in current transmission networks was developed and distributed to more than 30 European TSOs. The key findings of the questionnaire, based on 23 responses, are included in this chapter. Chapter 4 deals with PQ monitoring and assessment by introducing the proposed monitoring objectives. An overview of the selection of PQ monitoring locations and the duration of monitoring period is given, as well as the selection of monitoring parameters and equipment. For the PQ assessment, a unified approach is discussed. The assessment is firstly dependent on the PQ monitoring objectives and also for the requirements derived from standards, regulators, as well as the TSOs and clients requests. Finally, a section has been devoted to PMUs and their applicability for PQ monitoring with respect to harmonics. Chapter 5 brings together conclusions of the report and provides overview for the further research and development in the PQ in transmission networks with high PE penetration.

REPORT

of 158

2 PQ Disturbance Categorization

2.1 Introduction Following considers the categorization of PQ disturbances to identify critical PQ phenomena and study their propagation through transmission networks. For the identification of PQ phenomena and sources of PQ disturbances, the focus is on the increasing penetration of PE-interfaced devices in transmission networks. PQ issues in transmission networks are analysed by considering real-world experiences and examples, as well as studies published by researchers. Different PQ disturbance types and their main causes are categorized as the traditional well-known disturbance, such as frequency variations, voltage-level violations, flicker, voltage unbalance and harmonics. In addition, non-traditional harmonic stability issue, which derives from interaction between PE devices and their control systems and power networks, is covered. Analysis of PQ issues involves disturbance evaluation indices considering available standards, technical recommendations and system operators’ practices. The possible consequences of PQ disturbances are pointed out, as well as the potential mitigation methods. PE devices are one considerable sources of network disturbances (namely harmonics, flicker, voltage and frequency variation), while also being sensitive themselves to PQ disturbances. This bilateral relation is presented in comprehensive manner in Table 2-1. Objective of the table is to present PQ phenomena and provide the main criteria for understanding the impact of PE on PQ phenomena and PQ phenomena effect on PE. Based on the assessment it can be concluded that PE devices have an impact on PQ phenomena. The main phenomena under observation are frequency deviation, voltage fluctuations and flicker, harmonic distortion and harmonic stability. In case of frequency deviation, the issue is on the increased penetration of PE, which may cause higher frequency variations due to reduction of system inertia. It is stated that some PE devices have significant influence on voltage fluctuations and flicker and that PE devices are dominant source of harmonics in power system. In addition, PE devices have an effect on harmonic stability issue, as being source of harmonics they interact with the supply system harmonic impedance and as a consequence change resonant points. There are also some PQ phenomena, e.g. voltage unbalance, temporary power frequency over-voltages, voltage dip and steady-state voltage variation, which seem not to have direct relation regarding the impact of PE on PQ. In case of impact of PQ phenomena on PE it is acknowledged that many devices are susceptible to voltage dips, temporary over-voltages, and harmonic distortion in the supply network and some to voltage unbalance. High impact can be considered in case of harmonic stability phenomena. However, additional assessment is needed to verify these impacts in case of high penetration of PE in future transmission networks. For phenomena as frequency deviation, steady-state voltage variations, voltage fluctuation and flicker further research is required as currently the impact of these is not known or it is limited.

REPORT

of 158

Table 2-1: PQ disturbance categorization

PQ Phenomena Description Impact of PE on PQ

phenomena Impact of PQ

phenomena on PE Comments

Frequency

deviation

Small variations with respect to the nominal value of 50 Hz

Increased penetration of PE contributes to higher frequency variations due to a reduction in system inertia

Unknown; further research needed

Inertia emulation can be implemented in PE to assist with primary frequency control

Steady-state

voltage variation

Steady-state increase or decrease of RMS voltage, normally due to load variations

None apparent (assuming that PE devices comply with the reactive power requirements dictated in grid codes)


Large penetration levels of PE embedded in distribution networks present a challenge for voltage control at transmission level; there is a need for effective co-ordination and shared responsibility for voltage regulation at TSO/DSO interface See Note 1

Voltage dip

Temporary reduction of the RMS voltage below a specified threshold

None apparent from PE devices; negative effects are more related to the volatility of renewable energy sources (i.e. large cloud, solar eclipse)

Many PE devices are susceptible to voltage dips

Low Voltage Ride Through (LVRT) requirements must be defined in grid codes for all PE devices See Note 1

Temporary

power

frequency

over-voltage

(Voltage swell)

Temporary increase of the RMS voltage above a specified threshold

None apparent from PE devices; negative effects are more related to the volatility of renewable energy sources (i.e. large cloud, solar eclipse)

Many PE devices are susceptible to temporary over-voltages

High Voltage Ride Through (HVRT) requirements must be defined in grid codes for all PE devices See Note 1

Voltage

fluctuation,

Flicker

Impression of unsteadiness of visual sensation induced by a light stimulus, the luminance or spectral distribution of which fluctuates with time

Some PE devices are high contributors to flicker in power systems


Flicker is typically a symptom of voltage disturbance caused by the use of large fluctuating loads, i.e. those that have rapidly fluctuating active and reactive power demand See Note 1

Voltage

unbalance

The condition in a poly-phase system in which the RMS values of the line-to-line voltages (fundamental component), or the phase angles between consecutive line voltages, are not all equal

None apparent (assuming balanced three-phase PE devices)

Some PE devices are susceptible to voltage unbalance

The main sources of voltage unbalance are single-phase loads and generation connected at distribution or transmission voltage levels (especially traction systems); the inherent asymmetry of the power system is also a contributor

Harmonic

distortion

Harmonics are quasi-stationary distortions of voltage or current waveforms

PE devices are dominant sources of harmonic currents and voltages in power systems

PE devices are susceptible to harmonic voltage distortion in the supply network

Active filtering and harmonic cancellation can be implemented in the control algorithms of PE devices

Harmonic stability

Harmonic impedance and emissions from multiple PE devices can interact with the harmonic impedance of the supply system leading to excitation of harmonic resonances and instability of the wideband controllers

PE devices are dominant sources of harmonic currents and voltages in power systems; they also interact with the harmonic impedance of the supply system, changing resonant points

High New PQ issue introduced by the proliferation of PE devices in power systems

Note 1: FACTS devices can be utilised to mitigate this PQ issue.

REPORT

of 158

2.2 Frequency Deviation

2.2.1 Background The grid frequency is an indicator for the balance of power generation and demand within a power system [4]-[5]. Considering the PQ of an electrical power system, frequency control plays an important role. Thus, the quality of the frequency is of special concern not only to TSOs, who are responsible for the reliable operation of the electricity system, but also to generating companies whose generators have to react to frequency deviations. Inherent system inertia helps the power grid to resist frequency deviation and damps the frequency even before the primary control of the power grid sets in [6]. In case the loss of a large generation unit leads to a sudden unbalance of active power generation and demand, kinetic energy stored in the rotating mass in synchronous machines is released. As a result, the grid frequency drops because of the direct dependency of the rotation of synchronous machines and grid frequency. It can be stated that the higher the inertia, which is integrated in the transmission network, the smaller the resulting frequency drop [6]. All rotating (synchronous) machines, which are part of conventional power plants, add to the inherent inertia of the power system. In these days, power generation by renewable energy sources begins to replace conventional generation, such that the ability of the power system to resist frequency deviations decreases [6]. Consequently, replacing rotating mass of conventional generation, i.e. synchronous generators, by renewable energy sources significantly reduces the naturally available inertia of the power system and increases possible frequency deviations [6]. All the mentioned effects have a significant impact on the PQ of the power system. With a high penetration of renewable energy sources, the amount of associated grid connected converters also increases. Considering wind power conversion systems, variable frequency wind turbines do not operate with the same frequency as the power grid [7]. In order to counteract the effects of reduced system inertia, the necessity of implementing adequate control algorithms for renewable energies arises such that those generation technologies can also contribute to frequency control [7]. Frequency stability questions are more thoroughly discussed and analysed in MIGRATE project WP1 and WP3. WP5 covers this topic only in a scope of PQ disturbance.

2.2.2 Evaluation Indices According to [8], the mean value over 10-s of the fundamental frequency of the European electricity transmission network should be within 50 Hz ± 500 mHz range for 99.5% of week interval. Additionally, this mean value should not deviate -6% to +4% from the nominal value for 100% of week [8]. According to IEEE/CIGRE Joint Task Force on Stability Terms and Definitions, the measure of frequency stability refers to the ability of a power system to maintain steady frequency following a severe system upset resulting in a significant unbalance between generation and load [9]. The authors state that the measure of frequency stability depends on the ability of a power grid to preserve a balance between power generation and demand, by simultaneously minimizing the actions of unintentional loss of generation units or load. According to [5], the quality of frequency during normal system operation can be assessed using the standard deviation of the frequency error by Equation 2-1.

REPORT

of 158

� = �� ∑ �� − �� (2-1)

Where n is the number of measurements, which are taken every 15-min in a month. Main frequency is represented as f and fref is the frequency rated value (base value). The percentage share of frequency deviations higher than 50 mHz is calculated together with the times of their appearance [5]. Considering frequency deviation, it is to be distinguished between small-signal stability and transient stability of a power system [4]. Small-signal stability describes the ability of a power system to maintain synchronism when subjected to small disturbances [4]. In this context, a disturbance is considered small if the associated equations, which describe the resulting response, can be linearized for the purpose of analysis [4]. Transient stability is associated with the ability of a power system to maintain stable in case of large disturbances, i.e. an outage of a large generator unit. A loss of synchronism due to transient instability will be measurable within 2-3-s after the initial disturbance [4]. Considering transient stability, the gradient of the power grid frequency is a direct measure for the ability of a power system to remain stable with respect to sudden system events, for instance after an outage of a large generation unit or a disconnection of a large load [10]. Generally, if the power grid is not able to stabilize the grid frequency within a predefined range, the frequency deviation can lead to system instability [4]-[5]. Considering large interconnected power systems, e.g. the European transmission network, this type of situation is associated with conditions following splitting the systems into islands [9]. In this context, the evaluation of frequency stability is a question if the different islands are able to preserve the grid frequency with a minimum of unwanted disconnection of generation or load. It is determined by the overall response of the island as evidenced by its mean frequency, rather than relative motion of machines. It can be stated that frequency stability problems in transmission network are getting worse with inadequacies in equipment responses, poor co-ordination of control and protection equipment, or insufficient generation reserve [9]. Due to the increasing amount of integrated wind power generation units, this implies that control algorithms have to be adapted in order to adequately react to frequency deviations.

2.2.3 Consequences As an immediate result of an unbalance between generation and load, the system frequency starts deviating from its steady-state value. For instance, in the synchronous transmission network of Continental Europe depending on the system load frequency gradients of 5-10 mHz/s have been observed after power plant outages of 1 GW [10]. In case of grid frequency deviation, stored kinetic energy is used for primary frequency control to recover the system steady-state frequency. The inertia constants of directly grid connected synchronous generators of a large power plant are in the order of 2–9-s [11], and this inertia is inherently available, as the frequency of the power grid tends to decrease following a disturbance or loss of generation units and/or transmission lines [12]. With a high penetration of renewable energy sources, available kinetic energy, which is stored in the rotating mass of synchronous generators, is reduced. With the reduction of integrated system inertia, i.e. rotating mass of synchronous generators, the power system faces increasing problems regarding frequency recovery [10]. To allow converter-interfaced renewable energies to contribute to frequency

REPORT

of 158

stability of the grid, associated control algorithms have to be adapted. In case the primary control reserves of the power grid are exhausted, the system frequency as well as its gradient reaches exceptional values and the system state is called alert [10]. When the global security of the system is endangered or/and under frequency, load shedding is activated and the system is in the emergency state [10]. In extreme cases of frequency deviation, i.e. if the grid frequency exceeds the range of 47.5 Hz or 51.5 Hz, a system blackout can hardly be avoided [10]. Besides the consequences of reduced system inertia on frequency stability of a power grid, stochastic power fluctuations due to wind power generation units also negatively affect the frequency stability. This implies that the PQ of the transmission network suffers if (stochastic) power fluctuation is present [13]. In extreme cases of power fluctuation the resulting frequency deviation can lead to a partial or total collapse of the power system [13]-[16]. Generally, power system devices are protected by the circuit breakers that connect them to the network. Each breaker is attended by a protection relay and operates upon receiving an open or close signal from relay. Associated protection relay monitors device variables defined for its safe operation. If the variable violates its limit and the violation lasts longer than a pre-set period, the device breaker would be signalled to open [15]. As a result, excessive grid frequency deviation can cause power system blackout if key circuit breakers are tripped by frequency relays [15]. Considering the increasing demand of system flexibility, the evaluation of frequency stability becomes more and more important. The authors in reference [17] conducted an evaluation of power system flexibility considering a high penetration of variable power generation, i.e. renewable energy sources. It has been shown that a power grid always has to maintain a certain flexibility in order to adequately react to changes in the grid conditions. In addition to the mentioned phenomena, frequency deviations in transmission networks have also negative effects on grid connected electrical devices such as induction machines [18]. This means that the performance of induction motors will vary from the rating, if frequency deviation is present. A frequency, which is higher than the rated frequency, decreases load rotor torque and increases the speed and friction and winding losses [18]. At a frequency lower than the rated frequency, the locked-rotor torque is increased and power factor is decreased [18]. Furthermore, the speed decreases during under frequency conditions [18]. Electrical equipment demands a steady frequency in order to maintain an effective and efficient operational behaviour. If the grid frequency deviates from the nominal grid frequency, the operational behaviour of electrical devices will be negatively affected such that in extreme condition equipment can be damaged [19]. Malfunction of generation equipment can also increase frequency deviation and lead to additional cascading malfunctions [6], [19]. In particular, this implies that frequency deviation has a negative impact on the PQ of the power system such that in extreme cases power generation units or loads have to be disconnected from the power grid for security reasons. Furthermore, the power supply reliability suffers such that frequency deviations can lead to additional economic costs [20]. Nevertheless, the choice of appropriate counteractions in order to face frequency deviations in transmission networks is generally a matter of making a trade-off between PQ, power system security and reliability as well as resulting economic costs. With the purpose of finding a suitable balance of those aspects, the impact of frequency deviation on consumers’ equipment, on generation equipment as well as on the resulting costs of possible ancillary services has to be considered.

REPORT

of 158

2.2.4 Mitigation Methods Short-duration under frequency deviations require power generation assets to increase real power output, reducing frequency dips. Similar to the usage of kinetic energy stored in the rotating mass of conventional generators, i.e. synchronous generators, an adequate mitigation technique for providing synthetic system inertia is the use of kinetic energy stored in the rotor blades and generators of wind power plants (WPP) [12], [21]-[22]. In [12] a control strategy to mitigate the impact of reduced system inertia is proposed. The advantage of providing inertia this way is that the availability can be adjusted according to the demand. To increase the affected contribution by wind power generation units, a supplementary control for Doubly Fed Induction Generators (DFIG) power converters is designed. The authors also propose the idea of adjusting pitch compensation and maximum active power order to the converter, in order to improve inertial response during the transient with response to drop in grid frequency. It has been shown that the proposed control algorithm is not only valid against sudden wind speed changes, but also improves the frequency response in case of a failure of a large power generator. Furthermore, the authors state that a beneficial impact in terms of damping power system oscillations could be observed, which has been validated by eigenvalue analysis. The authors in [21] proved that kinetic energy, stored in the spinning wind turbine generators (WTG), is suitable for frequency control. The authors propose a frequency threshold detector such that the power electronic control is sufficiently fast to discharge the kinetic energy in storage. Studies showed that even with low wind velocity the wind turbines can still serve as kinetic energy storage and contribute to mitigate frequency deviations. The authors in [22] also propose a supplementary control loop for a controller of a DFIG, in order to reintroduce inertia response. To analyse the dynamics of the presented control loop, the inertia response is compared with the response of a fixed-speed induction generator. It has been shown that the greater variations in rotor speed, available from the DFIG wind turbine, lead to considerably more kinetic energy being released, in order to support the frequency control of the power system. In these days, some manufacturers already offer adequate technology to face the problem of reduced power system inertia. For instance, there has developed a solution that allows a short-term power increase of 5-10% in case of a frequency drop [23]. The proposed technology utilises proprietary control algorithms to transform the mechanical inertia of the rotor into a temporary increase in electrical power output over a short period of time. By applying the proposed control algorithm, other options for increasing the real power output of wind turbine generators can be avoided such that negative impact on annual energy production can be reduced [23]. It is stated that the proposed control system is able to detect under frequency events to utilise active power controls, in order to shape the power response of the turbine. The resulting impact on the grid frequency is similar to that of conventional synchronous generator in case of under frequency events. There is also developed other similar solution [24]. The maximum available power of the wind power conversion system is provided within 0.5 to 1-s. In a pilot project realised in 2012, this technology was enabled in 60 wind turbines with 2.3 MW rated power in the wind park Le Plateu in Quebec. It showed that it is possible to increase power injection by up to 10% of the rated power for a duration of 10-s.

REPORT

of 158

In extreme cases, when no spinning reserve is available, generation units are not able to increase their power generation such that the frequency drop is high and frequency collapse might occur [5]. Many power systems can be protected from frequency collapse by importing power from neighbouring systems to make up for the lost generation [5]. If this is not possible, i.e. if an islanded power system is considered, the only way to prevent a frequency collapse following from a large system disturbance is to apply automatic load shedding. Load shedding is a control measure, which is implemented in stages with each stage triggered at a different frequency level [5]. This allows the least important loads to be shed first. In [25] an adaptive optimal estimation technique for calculating the frequency deviation and its mean rate of decay for load shedding applications is presented. The proposed method uses an extended Kalman filter algorithm, in order to estimate the instantaneous frequency deviation from highly distorted voltage measurements. The authors in reference [26] propose a two-stage load control scheme in case of higher load than generation. During the first stage of the load control an event based under frequency load shedding is performed, which guarantees fast response to the detected high-consequence disturbances. In the second stage model, predictive control is triggered for step-by-step online closed-loop control of interruptible loads. The proposed control scheme has been tested and validated on a 9-bus system and on the New England 39-bus system.

2.2.5 Practical Studies In past years, several companies have conducted practical studies that focus on the possible effects of a reduced system inertia on frequency stability. A study performed by the German Energy Agency analyses the effects of the rising share of renewable generation in Germany on the frequency stability of the European power system [27]. As already explained above, renewable generation without additional control systems does not support the inertia of the system. The authors state that because of the increasing share of renewable energy sources in Germany, the contribution of this country will become very low in 2030 for times of high renewable generation. Therefore, countermeasures have to be taken, in order to keep the German inertia share on the same level. For the year 2011, the results obtained from modelling show that for a power step of 3 GW, Germany provides frequency response reserve of 372 MW and kinetic energy of 0.95 MWh [27]. Without retaliatory action, the authors state that the German participation could be reduced to a third of today’s share for certain points of time. However, given that the remaining European power system does not change, for transient deviations the minimum frequency would still be set at 0.25 Hz above the critical value of 49.2 Hz [27]. In order to keep the electric system stable, for the same step of 3 GW, additional 254 MW of frequency response reserve and 0.68 MWh of kinetic energy have to be supplied [27]. In 2011, the European Network of Transmission System Operators for Electricity (ENTSO-E) published a report, in which the frequency of the synchronous grid of Continental Europe is analysed [28]. The authors state that in the last years all synchronous areas of ENTSO-E have been experiencing increasing frequency deviations in amplitude and duration. Considering this problem, permanent grid frequency monitoring has been established in order to analyse those effects [28]. Frequency deviations in the transmission network with peak-to-peak values up to 150 mHz could be observed primarily within a time interval of 10-min [28]. In Figure 2-1, the average frequency of the transmission network in Continental Europe for June 2003 and June 2010 [28] is shown. It can be seen that the shape of the grid frequency reacts to events, i.e. the load evolution in the grid, and, therefore, fluctuates around the rated 50 Hz. In Figure 2-2, the result of permanent monitoring of

REPORT

of 158

the system frequency of the Continental Europe transmission network for the years 2001-2011 considering frequency quality [28] is shown. The analysis is based on a monthly resolution considering the criterion of exceeding ±75 mHz around the nominal 50 Hz grid frequency.

Figure 2-1: Average frequency values in Continental Europe, June 2003 and June 2010 [28]

Figure 2-2: Frequency quality behaviour in Continental Europe during years 2001-2011 [28]

REPORT

of 158

The Canadian TSO Hydro-Quebec has integrated nearly 3.5 GW of generation from WPPs by 2015 [29]. Because of the fact that Hydro-Quebec’s system is asynchronous from its neighbouring systems, and because of its relatively low inertia (approximately 42 GW of installed capacity), it is of great importance to maintain system stability despite installing more and more wind turbines. The presented concept in [29] was applied to specific inertia emulation parameters, using a negative load model injecting or absorbing active power at the wind farm collector buses. The inertia emulation scheme was developed for a step function and a proportional function. Currently, Hydro-Quebec is in the phase of validating the manufacturer’s models integrating inertia emulation features. It has to be awaited if the settings prove to be sufficient, especially for the case of more WPPs being built in the future.

2.3 Steady-State Voltage Variation

2.3.1 Background In recent years, for reasons including delayed network expansion, many power systems worldwide have been operating close to their stability limits [4]. In order to achieve an adequate level of PQ and supply reliability, the magnitude of the three phase voltages of an electric power grid is required to be within a certain predefined range at any time [4]. Considering the European transmission network, the voltage limits are defined by the ENTSO-E network code on Operational Security [30]. According to [30], the voltage limits depend on the voltage level of the transmission network, as well as on the considered synchronous area. These voltage constraints should be satisfied in normal grid conditions, as well as in N-1 outage conditions. However, it is worth mentioning that a violation of the predefined voltage range can occur during different time ranges. This means that one has to distinguish between short-duration voltage violations events, such as voltage fluctuations or voltage drops, and long-duration voltage violations. Short-duration voltage violations are described in a later section of this report, whereas this section focuses on long-duration voltage variations named as also as steady-state voltage violations. The voltage stability issues and operational behaviour of power grids with high penetration of PE devices will be analysed in MIGRATE project work packages 1 and 3. However, since PQ includes the continuity of power supply then the grid stability also affects PQ. According to [4], power system enters to the state of voltage instability when the disturbance, increase in load demand, or the change in system conditions causes a progressive and uncontrollable decline in voltage. The magnitude of the three phase voltages in a transmission network can be affected by grid faults, load, transmission paths, energy management, supply distances, power flow dispersion and reactive power compensation [9]. Further reasons for voltage variations in transmission networks are heavy load variations and system switching operations [31]. Incorrect tap settings on transformers can also result in system over- or under-voltages. Considering a linearity of the relation between power and voltage in the transmission network, the voltage drop due to an increasing load will be the same as the voltage rise due to a decreasing load [32]. The phenomenon of voltage variation in power systems can initiate severe stability problems, such that the power system can collapse [4]-[5]. This effect becomes even more critical if an inadequate supply of reactive power is provided. In case of voltage instability, the collapse of the power system can be either partial or total (blackout) [4]-[5]. In order to prevent a possible voltage collapse of the power system, it is desirable to measure the

REPORT

of 158

bus voltages. Therefore, for any bus of the power system, the Mvar-distance to voltage collapse has to be above a predefined value [33]-[34].

2.3.2 Evaluation Indices According to [9], voltage stability refers to the ability of a power system to maintain steady voltages at all buses in the system after being subjected to a disturbance from a given initial operating condition. The authors state that voltage stability depends on the ability of the power system to maintain or restore equilibrium between load supply and demand. Generally, voltage variation is defined as an increase or decrease of the magnitude of the three-phase voltage system. According to European Standard EN 50160, the 10-min mean value of the supply voltage should be within a 10% range around the nominal value for at last 95% of week interval [8]. For maximum 5% of the week, the mean value can be up to 15% below the nominal value [8]. In comparison, IEEE standard 1159 [35] defines steady-state voltage RMS variation as a variation of the RMS value of the voltage from the nominal for a time greater than 1-min [35]. Furthermore, a sustained interruption is defined as a type of steady-state voltage RMS variation, where the complete loss of voltage (<0.1 p.u.) on one or more phases exists for a time-period between 3-s and 1-min [35]. According to [8], the time- period is defined as longer than 3-min. However, voltage interruptions, which occur longer than 1-min, have the permanent nature in most of the cases. In particular, this means that the usage of an auto-recloser cannot mitigate the source of the voltage interruption such that a manual intervention for restoration is required [35]. However, the duration is further described using a modifier, indicating the magnitude of a voltage variation (e.g. under-voltage, over-voltage, voltage interruption) [36]. The steady-state RMS voltage variations are classified according to Table 2-2. Table 2-2: Categories and typical characteristics of steady-state voltage RMS [35]

Categories Steady-state RMS variations

Duration Voltage

magnitude

Interruption, sustained >1-min 0.0 p.u Under-voltages >1-min 0.8-0.9 p.u Over-voltages >1-min 1.1-1.2 p.u

Table 2-3: Voltage ranges for reference voltages defined by TSOs at 110-300 kV [30]

Synchronous area Voltage range Duration

Continental Europe 0.9-1.118 p.u unlimited Nordic 0.9-1.05 p.u unlimited Great Britain 0.9-1.1 p.u unlimited Ireland 0.9-1.118 p.u unlimited Ireland offshore 0.9-1.1 p.u unlimited Baltic 0.9-1.12 p.u unlimited

Table 2-4: Voltage ranges for reference voltages defined by TSOs at 300-400 kV [30]

Synchronous area Voltage range Duration

Continental Europe 0.9-1.05 p.u unlimited Nordic 0.9-1.05 p.u unlimited Great Britain 0.9-1.05 p.u unlimited Ireland 0.9-1.05 p.u unlimited Ireland offshore 0.9-1.05 p.u unlimited Baltic 0.9-1.05 p.u unlimited

REPORT

of 158

The critical level of voltage variations in transmission network is defined by the associated TSOs. Therefore, the requirements and standards of the TSO should be considered to evaluate voltage variations. According to ENTSO-E network code on Operational Security [30], voltage ranges for transmission networks between 110 kV and 300 kV are defined in Table 2-3. Considering voltage levels between 300 kV and 400 kV, voltage range for transmission networks are defined in Table 2-4. It can be seen that the critical level of voltage variations, i.e. emergency high/low warning depends on the considered voltage level, as well as on the local area of the associated TSOs. In these days, the permanent PQ monitoring plays already an important role in reliability-centred maintenance [32]. This includes that TSOs monitor system voltages in real time. Considering voltage variations, the TSOs are responsible for the voltage control in the system as a part of a secure supply. Every TSO must ensure voltage stability in its controlled area. This includes the power grid itself, generation units, power stations and boundary areas. Therefore, voltage variation recordings and statistics has been an established method to evaluate the impact of voltage variation on the transmission network and to test the performance of power grid protection devices [37]. If a generation unit reaches to the reactive power generation limit, the TSO should be notified. Consequently, if bus voltages reach a critical level, additional supplementary generation units have to get in charge to supply reactive power [38]. When considering the providers of voltage control services, the generation of reactive power can be divided into basic and enhanced reactive power service. The basic service respects limitations, which have to be fulfilled if a generation unit is about to be connected to the grid. The non-obligatory enhanced reactive power is provided as supplement to the basic requirements [39].

2.3.3 Consequences According to IEEE standard 1159 [35], steady-state RMS variations in supply voltage, i.e. lasting longer than 1-min, can cause load equipment problems [35]. Furthermore, it is stated that over- and under-voltage problems are less likely to occur on utility feeders, as most electric power providers strive to maintain ±5% voltage regulation. However, voltage variation can occur due to overloaded feeders, incorrect tap settings on transformers, malfunctioning regulators or controls, blown fuses on capacitor banks, and capacitor banks in service during light load conditions [35]. Under-voltages in excess of 1-min can also influence motors, so that controllers can drop out during such grid condition [35]. The dropout voltage of motor controllers is between 70-80% of the nominal voltage [35]. Another effect of a steady-state under-voltage is the increase of heat in motors caused by a higher current drawn [32]. Speed changes are possible for induction machinery during under-voltage conditions due to loss of torque [32]. The reduced starting torque increases the time needed to accelerate the motor. It can be stated that voltage variation leads to a reduced efficiency of the induction machine including negatively affected lifetime and higher power consumption with additional losses [32]. Furthermore, in [32] it is stated that in extreme cases the motor can stop the acceleration at all and stalls, such that the stalled motor draws a high current without any rotation. As a result, the motor becomes a short-circuited transformer, so that if protection devices do not disconnect it, it will overheat quickly [32]. This implicates that induction motors must be disconnected under such grid conditions. For capacitor banks, the under-voltage conditions result in a reduction of reactive power output, since the reactive power output is proportional to the square of the applied voltage. According to

REPORT

of 158

IEEE standard 1159, under-voltage conditions on transformers, cable, bus, switchgear, metering devices and transducers do not directly cause problems for the equipment [35]. Nevertheless, it is stated that during under-voltage conditions the visible light output from lighting devices can be reduced [35]. Considering steady-state over-voltage, electronic devices can experience immediate failure during such grid conditions [35]. Sustained over-voltages can overexcite transformers, generate harmonic distortions and cause transformer overheating [35], [40]. Over-voltage on transmission cables, busses, switchgears and rotating machineries can result in severe loss of equipment lifetime due to insulation degradation [35], [40]. Furthermore, during over-voltage grid conditions, the reactive output of capacitor banks increases with the square of the voltage [35]. Over-voltage conditions can also increase the visible light output from lighting devices [35].

2.3.4 Mitigation Methods One mitigation technique for voltage variations is the usage of reactive power controlling devices. In [41], the dynamic regulation models for Voltage Source Converter (VSC) based Flexible AC Transmission System (FACTS) controllers are described in order to enhance power transfer capability during voltage variations. An optimal allocation method for FACTS devices considering congestion relief and voltage stability is presented in [42]. Beside the technical issues, there is also considered an economic impact of FACTS devices. It is concluded that the presented method in [42] minimizes annual device investment cost and maximizes annual benefit. Shunt capacitors can also be used to supply reactive power and increase local bus voltages in order to mitigate voltage variations [4]. The integration of shunt capacitor units contributes to the efficiency of power transmission [4]. Nevertheless, the principle disadvantage of shunt capacitors is that their reactive power output is proportional to the square of the voltage. This implies that the reactive power output during low voltage conditions, i.e. when it is needed most, is reduced [4]. The authors in [43] also propagate the operation of capacitor banks, in order to mitigate voltage variations in transmission networks. A method for the optimal allocation of shunt capacitors with respect to voltage correction of a power system in steady-state condition is presented. The proposed method also considers load uncertainties. However, it has to be mentioned that the increasing integration of capacitor banks can also lead to transient voltage stability problems during switching operations [4]-[5], [9]. Considering the increasing integration of renewable energy sources, the impacts caused by wind power generation on voltage stability have been investigated in [44]-[45]. The authors used several tools for static analysis such as classical PV-curves. Nevertheless, power converters of renewable energy sources can also contribute to the mitigation of voltage variations [46]. By applying adequate control strategies, power converters are capable of injecting positive, as well as negative, reactive power into the grid [46]. Hence, the network bus, to which the power converter of a renewable energy source is connected, can operate as a PV-bus (generator bus) with a controlled voltage magnitude [46]. However, is has to be considered that those bus voltages can also be a source of transient voltage variations, in case of power fluctuations of the renewable energy source [44]-[45]. In order to avoid a collapse of the power system due to voltage variation, it is desirable to determine available reactive power reserves. In [47], a two-step approach is proposed to evaluate reactive

REPORT

of 158

power reserves with respect to operating constraints and voltage stability. During the first step, the minimum overall needed reactive power reserve is determined. This problem is formulated as a security constrained optimal power flow. In the second step, using dynamic simulation the available reactive power reserves has to be specified. The [48] proposes a reactive power reserve management program based on the optimal power flow. Considering the optimisation, the authors in [48] assign different weights for each generator to maximize reactive power reserves within the system areas, which are most vulnerable against voltage stability problems. A decomposition technique is proposed to formulate sub-problems, which can be solved with various stress levels afterwards. In order to maintain near-optimal voltages in the transmission network, the authors in [49] propose a hierarchical two-level voltage controller. The presented method co-ordinates discrete reactive power control devices in the transmission network. The local controller use local PMU measurements to carry out control decisions on switching var devices. The central controller determines optimal voltage schedule and co-ordinates operation of the local controllers to achieve benefits for the whole power system. As a last option to avoid system instability due to voltage collapse, load shedding can be performed [4]-[5]. In reference [50], several concepts of load shedding during under-voltage grid conditions are discussed. It proposes a load shedding scheme to counteract the imminent danger of voltage collapse of the power system. The presented method uses a set of distributed controllers, where every controller monitors the three phase voltages in a zone of the transmission network. Furthermore, the controllers are regulating a group of related loads. Depending on the evolution of the monitored bus voltage, the controller acts on a set of loads located at distribution level and having influence on the voltage. In order to stabilise the bus voltage in the transmission network, the controller successively sheds parts of those loads.

2.3.5 Practical Studies In 2013, the French TSO Reseau de Transport d’Electricite (RTE) published a report of the main events and reliability in the European Electricity System [52]. Amongst other topics, the report analyses the voltage profile in the transmission network in France regarding voltage stability issues. In this context, the report considers the emerging phenomenon of over-voltages not only in the RTE network, but also in those of other European TSOs (notably in Spain, Belgium and Germany). Reactive energy is produced during periods of low power consumption, mainly in summer. Over-voltages in the transmission network occur in case of lightly loaded overhead lines. The authors state that this phenomenon gets worse by the presence of high voltage cables, which have notable capacitive behaviour. Furthermore, over-voltages in the transmission network are more likely because of the development of generation sites spread widely across the distribution grid, meaning these sites consume less and less or are even injecting power into the transmission network. The authors state that during the years 2011-2013 the amount of hours, where high voltage cables had to be disconnected has steadily increased. However, according to [52], RTE proposed several approaches in order to control voltage variations in the transmission network:

• Installing new capacitor banks with a combined nominal power of 2500 Mvar; • Installing Static var Compensators (SVC) in the 225 kV substations in Domloup (Brittany)

and La Merlatiere (Vendee);

REPORT

of 158

• Carrying out additional analyses to define the operating principles for the future voltage protection relay designed to handle potential voltage collapses over the Northern part of France;

• Improving the dynamic simulation study process to further develop anticipation of delicate situations and to ensure that any decisions to trigger safeguard orders are as accurate as possible.

The dynamic voltage behaviour studies allow the analysis of “slow-dynamic” voltage plan changes and the characterization of needs for further implementation of control resources. The aim is to accurately locate zones in the transmission network, which have a higher risk of voltage collapse. This analysis considers several parameters like consumption level, available compensation resources, network and generation availabilities and limitations of the reactive capacities.

2.4 Voltage Dip and Temporary Power Frequency Over-voltage

2.4.1 Background The voltage amplitude throughout the power system is normally maintained at a value between predefined tolerances. Occasional rapid increases of current through conductors may cause large voltage drops on the impedance of these conductors. These voltage drops may reduce the voltage at nodes to such level that it affects the operation of loads connected to the nodes. This type of phenomenon is called voltage dip in standards (e.g. IEC 61000 2-1 [53]; in some papers there is used term voltage sag, as it corresponds to U.S English) [54]. The voltage dip is defined as a decrease in the voltage RMS value between 0.9 and 0.1 p.u in RMS voltage lasting typically from half a cycle (10-ms in 50 Hz system) up to 60-s [35]. Beside the magnitude and duration, as only typical characteristics described in international standards, the other voltage dip characteristics include, phase angle shift, point on the wave of dip initiation, voltage recovery and symmetry (i.e. single, two or three-phase dips). Most common reason for the rapid current increase is the occurrence of a fault somewhere in the power system. Due to the meshed characteristics of transmission networks, voltage dips caused by faults can spread over long distances within the network and may affect even urban and rural areas [55]. A fault can occur due to a lightning strike, manmade accident, bird nesting or many other reasons. Less commonly, the rapid current increase can stem from a very large load increase, e.g. starting of a large motor or transformer energising. Voltage dips caused by a large load increase will generally be less severe than those caused by faults. At the node of short-circuit, the voltage will be near zero for the duration of the short-circuit time, while other nodes in the system will have their voltages reduced. Generally, the nodes closer to the short-circuit will have their voltage fall deeper. Voltage dip pertains to the phenomenon when the voltage of at least one node falls below predetermined tolerances, but not to zero. A voltage that falls to zero implies that a certain node has lost connection to sources of electrical power, which is regarded as a supply interruption. As the voltage dip is normally the consequence of a fault, the clearing of the same fault may cause some parts of the system to experience a supply interruption afterwards.

REPORT

of 158

With high amount of integrated renewable energies, especially wind and PV, voltage dips in one or more phases of the electric network are also likely, because of the volatility of power injection. In this case, voltage dips can be the result of large clouds or solar eclipse [56], which lead to an abrupt loss of power injection by the PV power plants. Studies have shown that this effect can lead to critical transient behaviour of the transmission network and short-term voltage stability [57]-[58]. On the other hand, it is worth mentioning that, for instance, the operational behaviour of PV power plants is as well sensitive against voltage dip disturbances, since a disconnection of the power plants due to stability reasons is possible. Beside the voltage dips, there is also a violation in voltage amplitude predefined tolerances as a voltage increase. A temporary power frequency over-voltage (or temporary over-voltage in some cases) is the term used to describe the increase of the RMS of the voltage value above a certain threshold level for a short amount of time (short-duration). This is in its principle the opposite of a voltage dip. The causes for a temporary power frequency over-voltage can be any of the following: a large load decrease or transfer of loads from one power source to another, the Ferranti effect, self-excitation, resonances and ferroresonances [59]. The temporary power frequency over-voltage will last for the duration of the fault [35], [60]. The IEEE Standard 1159 [35] defines temporary power frequency over-voltages as phenomena, which last from half a cycle up to 1-min. During the temporary power frequency over-voltage, depending on the network neutral-point treatment, the magnitude of the phase voltage in the un-faulted phases can reach to √3 in case of grounded systems, as the reference is shifted and the phase-to-ground voltages obtain the magnitude, which was previously phase-to-phase. A general overview of magnitude-duration relationship of the voltage dip and temporary power frequency over-voltage phenomena is provided in the Figure 2-3 below.

Figure 2-3: Voltage dip and temporary power frequency over-voltage phenomenon

REPORT

of 158

2.4.2 Evaluation Indices According to the IEC 61000 2-4 [61], voltage dip is a sudden voltage drop that falls below a predetermined threshold level, followed by a recovery after a brief interval. The threshold level is typically set to be the same as the tolerance for the voltage regulation, e.g. if the tolerance is ±5% the threshold level will be a drop of 5% or 95% residual of declared supply voltage. According to the standard, if the voltage drops below a certain pre-set level (10% of voltage for example), the occurrence is no more considered as voltage dip, but as a supply interruption. In a three-phase system the above mentioned is to be applied to each phase separately, with the depth of the dip, being defined by the phase, which had its voltage dropped the most. According to EN 50160 [8], a voltage dip ranges from 90% to 1% residual voltage, if the voltage falls below 1% the occurrence is a supply interruption. A voltage dip single-event indices are primarily described with two numbers, these are the depth of the voltage dip and its duration. The depth is the difference between the minimal value of remaining voltage and the reference voltage and it can be estimated from a fundamental frequency complex network theory model [62]. The duration is the time-period for which the voltage remained below its threshold levels. The duration starts from the fault-clearing period [62]. In IEEE 1564 [63], two single-event indices for voltage dip have been introduced [64]. The voltage dip energy and voltage dip severity. The voltage dip energy E is defined with Equation 2-2. � = � 1 − ��(�)�� ∙ � !" (2-2)

Where T is the duration of dip, V(t) is the voltage at the terminals and Vnom is the nominal voltage value. The voltage dip severity S (Equation 2-3) is defined from the retained voltage V in per-unit and the duration d by comparing these values with so-called SEMI-curve. # = �$��$�%&'()(*) (2-3)

Where V is the amplitude of the dip, d is the duration, Vcurve(d) is the magnitude value of the reference curve for the same duration. As the voltage dip phenomenon is regularly occurring in the power system, the system operators have recommended different fault-ride-through (FRT) criteria for generation units. According to new EU regulation [65], the FRT curve is defined in the manner shown below on the Figure 2-4. The FRT criteria in ENTSO-E is established for different types of generation units by different definitions of the time-points in Figure 2-4.

REPORT

of 158

Figure 2-4: FRT criteria by EU regulation [65] There have been defined a set of voltage dip indices in IEEE 1564 and the methods of reporting dip results for a site as well as for a network are suggested in [62]. The most common index for reporting the results is the System Average RMS variation Frequency Index (SARFI). This includes voltage dips, temporary over-voltages and long interruptions [62]. Site performance, also the system performance, is often described in a voltage dip table form. The most common in use is the density-table [66]. A method for reporting site information from event magnitude and duration is described in IEEE 493 as the voltage dip co-ordination chart method [62]. Some utilities have developed approaches to define specific areas on a magnitude/duration plane that attempt to provide generalised guidelines on areas where dips are likely to reduce the number of indices that need to be reported and managed, based on the most “appropriate” grouping of dip events. Different dip indices and reporting methods are more broadly discussed and compared in [62]. There are no suitable voltage dip objectives defined by international standards [62]. The main reason is the lack of information concerning voltage dips and the widely different network topologies and operational environments [62], [67]. Some immunity standards define minimum levels to guide users of some equipment connected at low voltages, but as such these cannot be considered directly in comparison relating to objectives applicable to supply network [62]. One approach to try to limit the dips at the transmission level might be to assess the impact of dips at HV and EHV onto the LV level and limit the HV level dips accordingly. For example, The Netherlands Authority for Consumers & Markets (ACM) has announced a new standard dealing with voltage dips in the HV networks [68]. The authors in [69] also mention that due to lack of regulation for dips in EN 50160 Norway has defined its own standard. The European standard EN 50160 gives the voltage dip objectives in general terms for LV, MV and HV (≤150 kV). It states that under normal operating conditions the expected number of voltage dips in a year may be from up to few tens to one thousand. The majority of voltage dips have a duration less than 1-s and a retained voltage above 40% [62]. Standard EN

REPORT

of 158

50160 gives so-called voltage characteristics, which implies that these are values within which any costumer can expect voltage characteristics to remain most of the time [62]. In the UK, the limitations of the Rapid Voltage Changes (RVC) have recently been modified [70]. The new grid code GC0076 [70] ranges the RVC into three categories with different limitations of allowed voltage excursions. No limits, by means of frequency and number of events, are applied to category 1 where the maximum voltage change is within 1%. The category 2 is defined for events with maximum voltage variations between 1 to 3% with limited occurrences per hour (as a function of voltage variation amplitude). The category 3 defines allowed number of lower voltage decreases and increases as follows: maximum 12% decrease for duration of 80-ms with reduction dynamics to 10% and 3% after 80-ms and 2-s respectively; and maximum 5% is permissible if it is reduced to 3% after 0.5-s. Similarly, a new modification to the Irish grid code has been recently approved by the Irish Commission for Energy Regulation (CER) [71]. The purpose of this grid code modification is to clarify the roles, responsibilities and requirements expected from users and the TSO in relation to PQ. New limitations of the RVC are in line with IEC TR 61000-3-7 [72] as:

• Temporary voltage depression, in range of 5% of the nominal voltage with 3-s recovery time to nominal voltage;

• Voltage step change, in range of 3%. The temporary power frequency over-voltage is defined in several standards with small distinctions. According to IEC 61000-4-30 [73], a temporary power frequency over-voltage is any rise of voltage above a pre-set threshold level. It begins when the VRMS(1/2) voltage of one or more channel rises above the temporary power frequency over-voltage threshold and ends when the VRMS(1/2) voltage on all measured channels is equal to or below the temporary power frequency over-voltage threshold minus the hysteresis voltage. A typical value for the hysteresis is 2%. In the description above, VRMS(1/2) denotes the value of the RMS voltage measured over 1 cycle, commencing at a fundamental zero crossing, and refreshed each half-cycle. A channel in the above definition means any individual measurement path through an instrument. This encapsulates measurements of line-to-line, as well as phase-to-ground voltages. According to IEC 61000-4-30, a temporary power frequency over-voltage is characterised in a two-dimensional manner, by the maximum over-voltage and by the duration of the over-voltage. The maximum over-voltage magnitude is the largest VRMS(1/2) value measured on any channel during the temporary power frequency over-voltage. The duration of an over-voltage is the time difference between the beginning and the end of the temporary power frequency over-voltage. The standard IEEE 1564 [63] handles the temporary power frequency over-voltage in the same manner as IEC 61000-4-30. According to IEEE 1564, the single-event criteria mentioned previously for voltage dip energy (Equation 2-2) and voltage dip severity (Equation 2-3) can also be used for a temporary power frequency over-voltage. The only difference is in using the equation for the energy index. For the over-voltage, the Equation 2-4 is used. � = � 1 − ��(�)�� − 1 ∙ � !" (2-4)

REPORT

of 158

The standard IEEE 1159 defines temporary power frequency over-voltage as being an increase in RMS voltage above 1.1 p.u for duration 0.5 cycle to 1-min. The typical magnitudes of temporary power frequency over-voltage are between 1.1-1.2 p.u, where the magnitude of the over-voltage is to be described by its remaining voltage, hence will always be greater than 1.0 p.u.

2.4.3 Consequences Voltage dips are the most disruptive and costly PQ phenomenon, as they often lead to industrial and commercial process interruptions and subsequent financial losses to end-use customers. This is mostly due to the large number of occurrences throughout a typical transmission and distribution network, the increasing sensitivity of customer equipment to voltage dips, and the high costs of lost productivity and downtime [74]. The number of voltage dips that can occur at service entrance to a facility depends on where the facility is located in the network, the characteristics of the utility's supply system (e.g. underground or overhead lines), lengths of the distribution feeder circuits, the number of feeders, the lightning activity in the area, the number of trees adjacent to the power lines and etc. Some studies have found that nearly all disruptive voltage dips are caused by current flowing to short-circuit, either within the plant or on utility lines in the electrical neighbourhood [75]. Although the faults occur relatively rarely, a large portion of the power system experience voltage dips whenever a fault occurs, meaning that voltage dips are much more common than actual interruptions of supply. The reduction in voltage experienced during a voltage dip abates the energy transfer capability of the system, limits the fault clearing time in transmission networks or could results in tripping of embedded generation from the grid [76]. As during dip events (i.e. reduced voltage), the rotor angle increases to transfer the same amount of power, which causes the synchronous machine to operate closer to stability limits. On the customer side, the main consequence of voltage dips is the misoperation of sensitive equipment. Many modern devices like computers, process controllers, and Adjustable Speed Drives (ASD) experience operational problems when the voltage drops below 85% for more than 40-ms [76]. Equipment sensitivity to voltage dip varies according to load type, control settings, and applications. Equipment immunity to voltage dips is represented by the dip characteristics that influence the behaviour of the equipment. The sensitivity of individual equipment mainly determines how severely an industrial process will be affected by the dips. The growth of demand side technologies (most of these technologies utilise power electronics) is driven by the need for efficient consumption of electricity. This implies that equipment has become less tolerant to voltage disturbances. Some of the most sensitive devices, namely, AC ASD, AC coil contactors and personal computer have been thoroughly investigated and detailed sensitivity curves have been developed and reported in [77]-[79]. Voltage dips at the transmission level cause damages indirectly, as voltage dips from the transmission level spread to lower voltage levels through transformers. The impact that a certain voltage dip at the HV level has on the LV level, is largely dependent on the way transformers transferring the disturbance from the primary to the secondary side. In [80], the effect of disturbance propagation related to different transformer types have been studied. Assessing the connection between voltage dips at different voltage levels, in [80] authors concentrate on some different types of transformers and produced an explanation as how the disturbance propagates from one side of

REPORT

of 158

the transformer to another. The authors conclude that voltage dips at the transmission level influence those at the distribution level, while voltage dips at the distribution level seem not to influence the transmission level notably. In literature [75], [81] it has been shown that PQ disturbances due to voltage dips can lead to costly interruptions in industry facilities. The authors in [83] established a method to estimate the cost of downtime of industrial processes caused by voltage dip disturbances. However, PQ disturbances due to voltage dips have a severe impact on economic costs, the authors in [84] emphasise that economic losses show up in many aspects such as loss of revenue, lost opportunities, product damage, wasted energy, decreased equipment lifetime, field service warranty work, manufacturing interruptions and loss of productivity. The impact of process interruptions may include lost or wasted work in progress, costs caused by idle staff who are no longer able to work, process slow down, equipment damage, costs associated with restarting the process, potential penalties due to late delivery of finished goods and decreased competiveness [85]. The reported financial losses resulting from process interruption due to voltage dips are often expressed in millions (or even billions) of pounds/dollars/euros per year [86]-[88]. The reported losses vary widely depending on the type of industry. It is estimated that PQ problems cost industry and commerce in the EU about 10 billion € per annum [88]. The effects of temporary power frequency over-voltage on the equipment may be even more destructive than those caused by voltage dips. The temporary over-voltage may overload insulation or the equipment itself. Over-voltage can cause break-down of the power supplies components. In addition, the effect of temporary power frequency over-voltage can accumulate in the system [89] and generate even greater issues. The broad usage of electronics in the industry puts great emphasis on the importance of over-voltages as a PQ factor, as the capability of electronic components to withstand over-voltages is often lower than that of conventional equipment. When it comes to shunt capacitor banks used in system, they may give rise to interactions, such as resonance that increase or amplify these over-voltages [59]. The effect of the over-voltage on any particular item of equipment is dependent upon the magnitude and duration of the over-voltage, as well as the resilience of the equipment. The effect can range from slight degradation of performance to total failure. The resilience level of equipment should correspond to the risk of experiencing an over-voltage, in the sense that the insulation of the equipment should be co-ordinated with the expected level of over-voltage. The IEC 61000-2-14 [59] standard brings out the consequences of temporary power frequency over-voltage as lifetime reduction in filament lamps. However, this is not an issue for the EU market, as the product has been prohibited for several years. Additionally, there is also addressed a negative effect of over-voltages on IT equipment, where the CBEMA (Computer and Business Equipment

Manufacturers' Association) curve is mentioned in this context. The temporary power frequency over-voltage can also cause data errors, flickering of lights and degradation of electrical contacts [80]. However, these are mostly the issues on a distribution level. The temporary power frequency over-voltage on the transmission level could cause equipment malfunction either on the transmission level, or by propagating through a transformer onto the distribution level [90]-[92].

2.4.4 Mitigation Methods The mitigation of voltage dips can be either network based or device based [93]-[94]. The network based solutions are mainly concerned with improving the network performance regarding short-

REPORT

of 158

circuit faults, which are the main cause of voltage dips. This can be performed by taking measures to reduce the number of faults in the network, which will certainly reduce the number of voltage dips. Undergrounding of lines, insulating overhead lines, trimming trees, isolation washing, installing animal guards and additional re-closer placements and improving the information and signs systems of the underground cable locations to prevent dig-in faults, will reduce the probability of fault occurrence. Furthermore, limiting the faults severity and area of impact can improve the voltage drop performance of the network. This can be performed by having improved protection systems so that faults can be cleared faster, installing fault current limiters to reduce short-circuit currents and installing surge arrester to limit the impact of lightening surges. However, these methods are costly, as voltage dips also happen due to faults that are hundreds of kilometres away. The device based mitigation solutions of voltage dips can be performed in two ways: by either installing new devices that are capable of negating the effect of voltage dip disturbance or by improving the immunity of sensitive equipment to dip disturbances. Increasing the immunity of devices can be performed in the manufacturing stage by improving the ride-through voltage dip capabilities of equipment, which is also a trade-off between cost and performance. In [95], it is indicated that the installation of mitigation devices at the system equipment interfaces may be the most attractive short-term solution for customers. The mitigation device solutions introduced in [95] are motor-generator sets, transformer-based solutions and inverter-based solutions. Devices like on-line Uninterrupted Power Supply (UPS) and Flexible AC Transmission System (FACTS) devices, like Static var Compensator (SVC) and Dynamic Voltage Restorer (DVR) have proved very efficient in mitigating voltage dips by injecting the required reactive power to support the voltage during disturbances. However, high cost and sophisticated control of these devices hinder the wide deployment [94]. DVR and STATCOM are the most commonly used devices to mitigate voltage dips as for example discussed in [96]-[98]. The simulation results have showed that these devices can also contribute to system power factor correction and harmonics cancellation even though they were originally designed for voltage dip mitigation. As compared with voltage dips, the temporary power frequency over-voltage can also be mitigated by limiting the faults occurrence in the power system. However, the equipment and network insulation levels should be co-ordinated to prevent damaging the equipment from transient over-voltages. For this purpose, surge arresters are commonly installed into the system. However, extra care should be taken, since temporary over-voltage, which last longer than transient over-voltages may cause arrester failure [99] as they have limited temporary over-voltage withstand capability. As the same manner as mitigating the voltage dips, the researches on power systems seems to focus on various FACTS devices as the mitigation method for the temporary over-voltage. On the distribution system level, researches are dealing with the DVR and Interline Dynamic Voltage Restorer (IDVR) device, which is connected serially into the system via voltage injection transformer. DVR operates by injecting the desired voltage into the line. Several researches [100]-[103] show that both, DVR and IDVR can be useful for mitigating the temporary over-voltage phenomenon, as well as voltage dips. The Static Series Compensator (SSC) device, as the mitigation method for temporary over-voltage is examined in [104], which is more commonly used in distribution levels, but can also be extended to transmission level [104]. STATCOM is a shunt device, which more generally can be used for PQ improvement, since it can provide flicker compensation, voltage stability and power factor control.

REPORT

of 158

2.4.5 Practical Studies Regarding to voltage dips, there are some publications available on HV systems. However, they mostly deal with simulated networks and do not provide measurements. The authors normally gather some statistical data for faults, from which they try to infer the statistical data for voltage dip. The reference [105] stands out in the sense that it provides real measurements. In all the publications, those dealing with measurements or with simulation, the number of voltage dips in the transmission network seems to be consistent at a first glance in the sense that a node will experience somewhere in the order of tens, but less than a hundred voltage dips annually, most of which will be minor. Typically, less than 10 voltage dips per year have the depth of over 50% (or the retained voltage under 50%). In [105] authors present the results of long term monitoring of voltage dips at the HV, MV and LV level in the Czech Republic. They present the results for measurements conducted at seven distinct substations at 110 kV level for years 2006 to 2010. The number of voltage dips at 110 kV level is shown below in Figure 2-5.

Figure 2-5: Voltage dips at 110 kV substations as reported in [105]

The results obtained in [105] show that 80% of dips at the transmission level tend to last less than 100-ms. Over a third of all voltage dips (36.5%) are relatively mild as they have residual voltage higher than 85%. No measured voltage dip lasted longer than 1-s. It can be concluded, based on the results in [105] that the annual number of voltage dips experienced by any single node in the 110 kV system should probably range from fewer than ten to a few tens. In [106] the authors approach the problem of estimating the state of voltage dips occurrence in 97-bus National Grid of England and Wales transmission network. The results have brought out in Table 2-5 containing the number of occurrences that a certain node is expected to experience a voltage below certain level within one year.

REPORT

of 158

Table 2-5: Expected number of voltage dips for some substations [106] Voltage Desside Walhalm Bramley Palhalm Killingholme Cellarhead

90% 27.95 42.09 36.94 33.21 10.24 31.00 85% 21.75 29.41 21.42 22.63 6.66 15.11 80% 18.14 22.69 16.72 14.04 5.3 9.47 75% 16.07 16.58 8.95 9.35 2.44 7.69 70% 14.36 16.42 6.65 8.27 1.76 6.88 65% 12.73 15.46 5.55 5.44 1.76 5.20 60% 12.73 11.02 3.64 5.36 1.76 3.55 55% 9.09 7.98 3.56 3.36 1.76 2.30

The authors in [107] examined the impacts of phase-angle jump, voltage-recovery progress and various voltage dip parameters. It has been shown that at the beginning and end of the symmetrical voltage dip disturbances the response of the stator flux generates large transient overshoots in the rotor current as well as in the DC-link voltage. Furthermore, the authors observed transient overshoots and steady-state oscillations in the rotor current and the DC-link voltage when applying asymmetrical voltage dip disturbances to the system. An increase of negative-sequence components was detected if voltage dip disturbances were associated with phase-angle jumps. This has a major effect on offshore wind turbines rotor current and DC-link voltage overshoots increase, by 45-60% and 20-40% respectively. Moreover, a linear relationship between rotor current or DC-link voltage overshoot and the depth of the voltage dip disturbance was detected [107]. In [108] the authors provide the frequency of expected annual voltage dips for the Vietnam transmission network in several figures. The frequency of the dip of 0.9 p.u voltage remaining may happen over 30 times per annum and dips with voltage around 0.5 p.u may occur more than 5 times, but in general not more than 10 time at a certain node.

In [109] the authors try to gain some insight about the impact of wind power generation on voltage dips in transmission power network. They simulated the IEEE 30-bus benchmark model with various setups of wind generation and different voltage dips scenarios. The authors concluded that wind generation might improve conditions during the voltage dips. They also conclude that voltage stability may improve if wind penetration increases. Studies dealing with temporary power frequency over-voltage can be found more on distribution system level than transmission system level. Temporary over-voltages can occur in the phase-to-ground voltage of non-faulted phases during an earth fault. The over-voltage can reach up to 2 p.u, although values higher than 1.7 p.u are rare. These high voltages can be recorded for impedance earthed systems. For solidly and effectively earthed systems, voltage values between 1.2-1.4 p.u can be recorded in the non-faulted phases [110]. The following examples of events show the voltage performance during unsymmetrical ground faults. Figure 2-6 presents the recorded voltage at 130 kV bus due to a fault at the same level. The figure shows a very short duration over-voltage event (compared to the dip duration at the faulty phase) on one of the healthy phases [32].

REPORT

of 158

Figure 2-6: Waveform and RMS voltage performance of the three phases due to fault at

130 kV [32] In high impedance grounded systems or resistance earthed systems (mainly in distribution levels), the voltage rise in the non-faulted phases could reach up to √3 times the nominal values. For a single phase-to-ground fault in low resistance earthed systems, the voltage drop over the earthing resistance causes a zero-sequence voltage at different angles to the positive- and negative-sequence voltages at different locations. Therefore, the same fault may cause an over-voltage in one or two phases depending on the location of the fault. For example in [111], there have analysed temporary over-voltage mitigation in EHV network by using FACTS devices. The over-voltage in a real network study is generated by disconnecting a heavy load in a certain system bus. The study analyses temporary over-voltage levels at different system buses. The results showed that over-voltage could even reach up to 2.68 p.u in case of disconnecting transmission line from the system [111]. An example at distribution level [112], reports that in measurement period of one year an average customer disturbances per costumer was 0.5 with the amplitude overreaching 110%, but not surpassing 120% of the rated voltage. It can be stated that for some cases the temporary over-voltages do occur in the power system, however compared to voltage dips they appear to be rare [113]-[114].

2.5 Voltage Fluctuation and Flicker

2.5.1 Background Voltage fluctuation is defined as the cyclic variation of voltage amplitude that usually does not exceed 10% [115]. It is the effect of the interharmonics and subsynchronous harmonics (subharmonics) produced mostly by non-linear, fluctuating loads in the distribution and industrial networks. The effect on the network voltage varies with the ratio of the short-circuit and fluctuating power. In the local distribution network the voltage fluctuation can be caused by different small scale loads, while in the transmission network, to cause any noticeable flicker the variable power of the load needs to be much higher and the transmission path of the power at higher impedance.

REPORT

of 158

The main sources of interharmonics at the transmission level are the industrial devices as high-power cycloconverters, used for rolling mill drives or low frequency railways, ASD, electric arc furnaces, induction furnaces, HVDC links, traction drives, integral cycle control, low-frequency power line carrier etc. [116]. Flicker is the quantity that has been defined based on the human perception of the light-emitting sources (incandescent light) with the low frequency pulsating light intensity. Flicker is the PQ phenomena caused due to the fast changing, fluctuating load. The most crucial from the costumer’s point of view is the voltage level at the distribution level, medium and low voltage, where the devices or lights causing flicker are connected. In this case, the voltage fluctuation does not significantly propagate to higher voltage levels. In industrial and transmission networks, the arching devices, welders or high-power electric arc furnaces produce interharmonics and subharmonics, which give rise to voltage fluctuation in network, wherefrom it can propagate downstream to the low voltage networks and produce flicker. The flicker effect is produced at low variations of the voltage (few percent) and it becomes higher within the most sensitive frequency regions. Although the human perception of the flickering lights can vary, the flicker measurement has been standardised and it is based on the measurement of the variation of the voltage amplitude. The new technology of light sources that are using electronic ballasts can offer flicker-free operation with resilience of the light intensity to the low frequency voltage fluctuations. On the other hand, the low-quality products (ballasts, dimmers etc.) can amplify the flicker. One of the new contributors to the voltage fluctuations at the transmission network are the wind turbines, especially synchronously connected, fixed-speed types which suffers from periodical torque variation due to influence of the turbine tower to the wind [117]-[118]. On the other hand, the wind turbines based on full or partial connection through a converter can help reducing flicker in the network [119]-[120]. One important issues and the goals of many studies is the determination of the flicker level before connection of the new installation. Such analyses demand accurate and comprehensive knowledge of the future installation operation. This defines the expected power fluctuation and consequently flicker generation. Such studies are mostly carried out for the industrial client or the TSOs in the phase of evaluation of impact of new installations on the network. For realistic model of the fluctuating loads e.g. electric-arc furnace, complex models needs to be utilised in order to combine the broad harmonic spectrum of the injected distortion with complex propagation phenomena and mitigation principles [121]-[122].

2.5.2 Evaluation Indices

In the IEC 61000 standard series, the flicker is defined as an impression of unsteadiness of visual sensation induced by a light stimulus whose luminance or spectral distribution fluctuates with time. Since the flicker occurs as a result of voltage fluctuation, the standards define limitations for both, rate of relative changes of the voltage amplitude and flicker itself. Evaluation of the flicker is based on the human perception and irritation level of the illuminance variation, which is a non-linear function of the frequency and amplitude of the voltage variation.

REPORT

of 158

Therefore, the international flickermeter has been standardised in IEC 61000-4-15 [123] to define the following indices:

• Short-term flicker severity (Pst) is the flicker severity evaluated over a short period (usually in 10-min), where value 1 is the conventional threshold of irritability;

• Long-term flicker severity (Plt) is the flicker severity evaluated over a long period (2-h) using Pst values as in Equation 2-5

+,� = � �� ∑ +-�./��.��0 (2-5)

The flicker emission levels are assessed at the point of evaluation of a fluctuating device and the possible flicker attenuation to the lower network levels needs to be considered within the flicker assessment. The perceptible limits of the flicker are defined with the standardised test conditions, where at the supply terminals of the equipment under test connected low voltage network the following limits apply [124]:

• The value of the Pst shall not be greater than 1; • The value of the Plt shall not be greater than 0.65.

For the HV and EHV networks, the flicker limit values are defined by means of planning levels, while the compatibility levels, which are used for co-ordinating the emission and immunity of equipment in the systems to ensure electromagnetic compatibility (EMC), are defined for LV only. Planning levels, defined in IEC 61000-3-7 [72], are the indicative values that can be used for the purpose of determining emission limits, taking into consideration all fluctuating installations. Planning levels are specified by the system operator or owner. They are defined for each voltage level and can be considered as internal quality objective of the system operator. Indicative values proposed by the IEC 61000-3-7 are given in Table 2-6 below. In some countries, the planning levels are defined in national standards or guidelines. Table 2-6: Indicative values of planning levels for flicker in MV, HV and EHV [72]

Planning levels

MV HV-EHV

Pst 0.9 0.8 Plt 0.7 0.6

For the purpose of setting the emission limits, it is recommended to weight the given planning levels at HV level by taking into account the flicker transfer coefficient from the source of emissions to all other voltage levels. Planning levels must allow co-ordination between different voltage levels. Therefore, the operator has to evaluate the flicker transfer coefficients for various operating conditions of the system. To compare the measured flicker with the planning levels, the minimum measurement period is one week with normal business activity with some part of the period of expected maximum flicker levels. One or more indices may be used to compare the actual flicker levels with the planning levels. The indices are the following:

• The 95% probability weekly of Pst; • The 99% probability weekly of Pst; • The 95% probability weekly of Plt.

REPORT

of 158

By the standard IEC 61000-3-7, the 95% probability value should not exceed the planning level, while the 99% probability may exceed the planning levels by a factor in range defined by the operator depending on the system and load characteristics (for example 1.0-1.5). The European standard EN 50160 [8] sets the Plt≤1 as the permissible flicker level, which should not be exceeded in 95% of the weekly measurement time. This level is defined for public distribution systems. For the purpose of setting flicker emission limits to different voltage levels (also HV-EHV) and allow co-ordination between voltage levels, the flicker transfer coefficient can be used. The permissible flicker level can be recalculated by evaluating the flicker transfer coefficient for various operating conditions of the system. In practice, the transfer coefficients for different voltage levels are less than 1. For example, a typical value between HV and LV is 0.8. Further discussion of the assessment of flicker transfer coefficient can be find in IEC 61000-3-7 [72].

2.5.3 Consequences The sensitivity of the network equipment to the voltage variation is mostly relatively low and the operating problems are present only in rare cases. The main disturbing effect of voltage fluctuations is producing variations of the illumination intensity of light sources, the flicker [125]. The flicker is an unpleasant sensation experienced by the human visual system. People exposed to the flicker may experience nausea, distraction, problems with concentration, headaches. Flicker may even trigger the epileptic fits. Voltage fluctuations can also cause interference with communication equipment, tripping of electronic equipment or relays and in severe cases even prevent other loads to be started at the reduced voltage. Other sophisticated devices could malfunction and have reduced efficiency, which are costly in downtime and rejects [126]. Heating elements exposed to severe voltage fluctuations will have lower efficiency, which can increase the time of the process (e.g. metal melting). Synchronous motors and generators that are exposed to the fluctuated voltage at the terminals deal with increased losses, premature wear of rotors, changes in torque and power. Induction motors, on the other hand, cope with variation of torque and slip causing vibrations, reduction of mechanical strength and eventually shortening of service life [126]. Especially sensitive to the voltage variations is the electrolyser equipment, where the high current supply can become degraded, which increase maintenance and repair costs. Reduction of the power factor and generation of non-characteristic harmonics and interharmonics can occur to the static phase-controlled rectifiers with DC-side parameter control. In case of drive breaking in an inverter mode it can result in commutation failure and damage system components. Other effects of voltage fluctuation to the network equipment are the inconsistency of motor speed, triggering of the UPS systems to switch, the battery malfunctioning of the automated production line, etc.

2.5.4 Mitigation Methods The voltage fluctuations can be mitigated by using different approaches, such as decreasing the variation of load power (especially reactive) or by increasing the system short-circuit level. With help of modern FACTS devices, reactive power can be dynamically generated close to the fluctuated loads in order to reduce the overall power and voltage variations. The state of art technology offers different solutions that can be used as a compensation device. The most frequently used ones are the SVC and VSC based devices.

REPORT

of 158

The amount of the reactive power that is generated by the compensation device to reduce the voltage fluctuation is defined by the device’s control algorithm. Theoretically, if the load variates only by means of reactive power, then fast compensation could fully compensate the voltage variations in the network. In most cases, the variable loads consist of variable active and reactive components. Therefore, the active component remains uncompensated by causing voltage variations in the network. With the advanced control solutions, it is also possible to reduce the voltage amplitude variation that is caused by the active power fluctuation. The overcompensation is carried out by additional reactive power. Implementing the short-circuit increase as a voltage fluctuation reduction, is oriented on the system short-circuit power increase. This means reducing the impedance in point of connection between the network and generators. In most cases, this can be done by introducing additional parallel lines or additional parallel transformers, or connecting the desired network to higher voltage level. Decreasing the networks lines series impedance, decreases the voltage drop, as well as voltage fluctuations during power transmission. The solutions for the system short-circuit power increase are often very expensive and not always applicable. In order to reduce the flicker at the consumer sites, the following measures can be performed [115]:

• Separation of the supply that feeds fluctuating load from the rest of the network; • Time-scheduled operation of the fluctuating load in order to minimize the operation during

the experienced flicker periods (e.g. during the evenings or nights); • In some cases the change of fluctuating load operation practices, operation at lower power; • Supplying the fluctuated source from isolated source.

2.5.5 Practical Studies The literature that describes the flicker in the transmission network mostly consider the electric arc furnaces that are used for melting in steel factories, rolling mills and wind turbines as the source flicker. Short-circuit levels in the point of connection of the steel factories cause different levels of flicker at the point of connection, while the network impedance ratio from the factory (furnace) to the source of the energy defines the propagation of the voltage variation, or in other words the cause of flicker. Different authors tune their simulation models to suit the real network topology and measurements taken at different locations in the network. For example in [127], the flicker is measured in the north-west of Spain, where in local steel factory is 80 Mvar arc furnaces. The flicker that penetrates into 135 kV network at the factory location causes Pst95% up to 8.22 while on other locations it is around 2 or less. In [128] a detailed propagation analysis has investigated in Slovenian network where three larger steel factories are placed. At the 110 kV connection points of individual furnace the Pst maximum values reach 7, 4, and 2.5 respectively, while their contribution to the flicker at nodes that is in the middle of all arc furnaces locations reaches values up to 1, 0.5 and 0.3 for individual flicker source, while the total contribution to the Pst maximum value is around 1.1. Another detailed study in Slovenian transmission network [129] has shown that at three voltage levels (400 kV, 220 kV and 110 kV) the long-term flicker Plt exceeds value 1 only in 110 kV level at 29% of the busses. On higher voltage levels, the Plt exceeds value 0.6 at 33% of 220 kV busses and at 29% of 400 kV busses. In

REPORT

of 158

[130], the influence of the two steel plants is analysed and the correlation between the flicker at different levels is evaluated. The flicker level reach around 1.8 at the 220 kV bus in distances of 40 km and 20 km away from both observed plants respectively. Additional 50 km of the line to the next 220 kV bus reduces the Pst flicker to the level that is below 1. CIGRE working group C4.07 [62] has gathered available flicker measurements for different sites on MV, HV and EHV level. The results are shown in Figure 2-7 to Figure 2-9.

Figure 2-7: Measurement data for flicker Pst95 at MV, HV and EHV [62]

Figure 2-8: Measurement data for flicker Plt95 at MV, HV and EHV [62]

REPORT

of 158

Figure 2-9: Measurement data for flicker Pst99 at MV, HV and EHV [62] The availability of statistical parameters is low, particularly for EHV. Additionally, the gathered data is not always directly comparable with the 99% indices for planning levels. However, the results show that the actual flicker disturbance is sometimes more than double of the planning levels and jet no problems are observed in network. It follows, as many researchers report a concern that Pst threshold of 1.0 is too strict, as it is based on laboratory studies and since the flicker influences mostly the light emitting sources, not necessarily other equipment [131]. Reasons for this include the fact that a daily assessment does not weight daylight hours less severely, the use of lighting technologies other than incandescent, and the influence of other sources of lighting [62]. In addition, measurements are often conducted on the worst network locations, in the vicinity of important flicker source [62], which will not present the network conditions objectively.

2.6 Voltage Unbalance

2.6.1 Background The three-phase system is in unbalanced condition when the phase voltages are not equal and/or they do not have phase shift of 120° with respect to each other. Voltage unbalance is a steady-state quantity, defined as the maximum deviation from the average of the three-phase voltages or currents, divided by the average of the three-phase voltage or currents, expressed in percent (ANSI) [132]. Unbalance in voltage and current in a three-phase system, can be also analysed and quantified with the method of symmetrical components. Three complex voltages or currents are decomposed, a balanced component (positive-sequence), an unbalanced component (negative-sequence), and a common-mode component (zero-sequence). Transfer of energy from generation to load takes place in the positive-sequence only, thus if the three-phase system is balanced then the system voltages contain only a positive-sequence component [32]. The more common (IEC) approach defines voltage unbalance as a ratio of negative-sequence component to the positive-sequence component [132]. It is also known as the negative-sequence unbalance [3].

REPORT

of 158

The positive-sequence voltage represents the amount of voltage contributing to the power flow from generators to motor, whilst negative-sequence voltage indicates to the amount of unbalance in the system. The presence of zero-sequence voltage indicates a connection to the earth, measuring the amount of current not returning through the phase conductors [32]. The power system is designed as a voltage source characterised primarily by the positive-sequence voltage, and where the negative-sequence (and zero-sequence) voltages are small. In case of severe disturbances in the system, the situation may change significantly. In addition, since the negative-sequence (and zero-sequence) power flows from unbalanced load back to the system, the negative-sequence power is injected into the system [32]. The appearance of voltage unbalance indicates an inefficiency in the use of the three-phase system [32]. Most common and apparent reason for voltage unbalance is when the load currents are in unbalanced situation. The load unbalance is partly due to the natural variations between the single-phase loads in the three-phase system [32]. The uneven distribution of single-phase loads can be continuously changing across three-phase system [133] and even in cases where the loads are equally distributed over the three phases, the variation over time of the individual loads never result in the perfect balance between the load currents [32]. For an example, the problem areas can be rural electric power systems with long distribution lines. Industrial and commercial facilities may have well balanced incoming supply voltages, but unbalance can develop within the building from its own single-phase power requirement if the loads are not uniformly spread among three phases [133]. Since in MV and above networks, the loads are in most cases three-phase loads and the power supplied from transmission level is regulated to be balanced, it may seem that voltage unbalance is therefore more likely a concern in lower voltage levels and in distribution networks. However, there are some common exceptions in heavy single-phase demands that are imposed by large commercial facilities and are connected into the transmission network. Such load may be lighting loads, railway traction supplies and arc furnaces [32], [133]. In heavy industrial single-phase loads, the balancing problem becomes difficult to compensate, since the unbalance is continually varying [133]. Voltage unbalance also results from balanced current flowing through unbalanced impedances, since transformers and transmission lines are not fully identical in the three phases. In case of transformers, the centre leg of the three-phase transformer has different magnetizing current than the outer legs [32]. Depending on the transformer winding connection type, zero-sequence power flow becomes possible if neutral current exists. Although, it does not cause large problem for the network, it may circulate within the delta winding and contribute in winding temperature increment, leading to additional losses and damage. In addition, considering the open wye and open delta transformer banks, the transformer winding impedances are asymmetrical [133]. Although this type of transformer banks are more known in distribution level. Incomplete transposition of transmission lines and capacitor banks anomalies, such as a blown fuse on one phase of a three-phase bank may cause asymmetrical transmission impedances [132]-[133]. Three-phase overhead transmission line have also different inductances and capacitances, which results in coupling between positive- and negative-sequence voltages and currents [32]. Unbalance may also occur due to differences between the phases of three-phase equipment. Equipment that is supposed to be balanced, such as induction motors, can have due to design limitation or erroneous design unbalance current [32]. Within user facility, unbalance voltages can

REPORT

of 158

also be caused by unbalanced and overloaded equipment and high impedance connections, such as bad or loose contacts [133]. As voltage unbalance yields extra costs on operation, maintenance or replacement for both, network operators and customers, it is timely topic for the future design of smart grids where the presence of single-phase generations and loads will be noticeably increased. The large-scale integration of single-phase loads amplify the unbalance level in networks. Therefore, the initial network construction plan may not provide balanced working condition for customers. The distributed generation, such as single-phase renewable energy sources (e.g. PV), may either mitigate or aggravate unbalance. Furthermore, due to the fact of decentralized power injection by renewable energy plants and their associated power electronic devices, there is also a spatial dependency of voltage unbalance [134]-[135]. In consideration of transmission level, the highest influence on voltage unbalance originating from traction supply loads, which are heavy single-phase loads that are directly connected to the transmission network and which are in further details discussed in following Subsection 2.6.5.

2.6.2 Evaluation Indices PQ indices are broadly covered in a report by CIGRE working group C4.07 [62], including the comparison of voltage unbalance indices and objectives between different standards and reference documents. The standard parameter to evaluate the voltage unbalance is the ratio between the negative- and positive-sequence voltages, known also as Voltage Unbalance Factor (VUF) [32]. According to IEC 61000-4-30 [136], the standard measurement window is 10 (50 Hz) or 12 (60 Hz) cycles. From the fundamental component or RMS value of the three phase-to-phase voltages, it is possible to calculate the (negative-sequence) unbalance [32]. Alternative definitions for the voltage unbalance can be found in a number of standards. Hereafter, three different parameters are pointed out. These are defined within respective National Electrical Manufacturers Association (NEMA) and IEEE standard documents. These methods should however be handled with caution, since they are all referred as “unbalance”. Therefore, the different definitions may derivate significantly different results, by making it difficult to interpret and compare the results, especially if the zero-sequence component is present [32].

• The maximum deviation from the average phase voltage, referred to the average of the phase voltage (IEEE 112 [137] and IEEE 1159 [35]);

• The ratio of negative- and positive-sequence voltage (IEEE 1159, EN 50160 [8] and IEC documents);

• The maximum deviation from the average line voltage, referred to the average of the line voltages (NEMA definition).

The NEMA and IEC definitions gives similar results [32]. The other mentioned definitions deviate by a factor of 2 to 3, which does not mean that these parameters are less correct. The choice of parameters depend on the application, its characteristic and on the interests of an operator [32]. The basic characteristic of voltage unbalance is the ratio between negative- and positive-sequence voltages calculated over a 200-ms interval [32]. Aggregated values are obtained over the standard intervals of 3-s, 10-min, and 2-h [32]. IEC standard 61000-4-30 suggests over the one-week measurement period for contractual applications the following site indices [32], [62]:

• The number or percentage, of 10-min or 2-h values that exceed contractual values; • The highest (worst-case) 10-min or 2-h value compared to contractual values;

REPORT

of 158

• The 95% (or other percentage) of the 10-min or 2-h probability weekly values. In standard EN 50160, the 95% of the 10-min mean RMS values of the negative-sequence component of the supply voltage during one-week period is considered as unbalance index. National standard in South Africa NRS 048-2 [138] states for the voltage unbalance that for each phase, the highest 10-min RMS value, which is not exceeded for 95% of the week is noted, with the assessment period should be a minimum of 7 continuous days. In the aim of developing a voltage unbalance measurements processing procedure along the same line for voltage distortion and flicker (IEC TR 61000-3-6 [139]; IEC 61000-3-7 [72]), the issue of voltage unbalance was addressed in CIGRE working group C4.07 [62]. It is mentioned that the observation period should at least be few days, including weekend. The final recommendation is to use the following site indices [32], [62]:

• The 95% of the 3-s values over one day; • The 95% of the 10-min values over one week; • The 99% of the 3-s values over one day (in exceptional cases).

In [62] the summary of indices relevant to negative-sequence unbalance factor has given considering the international documents and regional or national standards and guidelines. Most commonly, the 10-min values are used. Although different equations may be used for calculating voltage unbalance factor, results should be similar for a given integration time. The voltage unbalance is described by the VUF, as discussed previously. The European standard EN 50160 defines the acceptable levels of unbalance in normal operating conditions, in terms of parameters and their permissible limits at supply terminals. The observed values are the 10-min values of the negative-sequence component of the supply voltage. In each period of one week, 95% of these values should be in range of 0-2% of the positive-sequence component depending on the voltage level [136]. In addition, the IEC standard series 61000 concerned with EMC issues, sets the voltage unbalance limit to 2% (LV) [73], [140]. However, in some cases the content of voltage unbalance may reach up to 3% level [141], according to the system constructional characteristics (e.g. in older parts of network and if the network is predominantly single-phase [62]). In many EU countries, the compatibility level of LV is VUF≤2%. The planning limits are VUF≤2% for MV and VUF≤1% for HV [142]-[143]. These values can variate in different countries by their local standards and grid codes. IEEE 241 [144] allows voltage unbalance to be in a range of 2-2.5% and IEC up to no more than 2%. ANSI C84.1 suggests under no-load conditions, set the maximum voltage unbalance limit up to 3% [141], since unbalance greater than this can result in significant motor heating and failure [132]. For example, the Australian National Electricity Code (NEC) specifies that the voltage unbalance be limited at 0.5% for systems operating at or above 100 kV, 1.3% for systems operating between 10-100 kV and 2% for 10 kV and lower voltage systems (over a 30-min averaging period) [145]-[146]. However, the most relevant standards for unbalanced single-phase loading is considered European EN 50160 and IEC 61000 series by taken as reference in many supply systems. In [62] the more detailed comparison of existing limits has been presented considering international documents and regional or national standards and guidelines. It can be noted that at MV level the 2% voltage unbalance limit is almost the same in every document. The exceptions up to 3% in some area have been made, usually in cases where the networks are predominantly single-phase (single-

REPORT

of 158

phase traction loads, and single-phase distribution). At HV and EHV levels, limits vary from 1% to 2% [62].

2.6.3 Consequences Power system is primarily designed to respond the positive-sequence voltage. An unbalance of the three phase voltages (negative-sequence component) have an effect on the PQ within the network and may cause major damage to power system equipment and loads [133], [147]. A small unbalance in the phase voltages cause a disproportionately large unbalance in the phase currents. Under unbalance conditions, the power system will incur more losses and heating effects, and be less stable because when the phases are balanced, the system is in a better position to respond to emergency load transfers. The effect of voltage unbalance can also be severe on equipment such as induction motors, power electronic converters and ASDs. Under unbalanced voltage condition, the operating of rotating machines experiences a reduction in efficiency. Rotating machines have a small negative-sequence impedance. Therefore, a negative-sequence voltage will result in additional current, which increases the losses and hereby the heating of the machine [32]. When the machine is operating at its rated power for longer periods, some derating is needed to prevent the overheating. Increased heating effect could lead to premature motor failure [133]. In induction motor, the positive-sequence voltage produces the desired positive torque, whereas the negative-sequence voltage produces an air gap flux rotating against the rotation of the rotor, thus generating an unwanted negative (reversing) torque. The result is a reduction in the net torque and speed, and the possibility of speed and torque pulsations and increased motor noise [133]. Overall, the net effect of the voltage unbalance is reduced efficiency and decreased life of the motor. Power electronic converters serve the interface for many large electronic loads, from three-phase UPSs to motors operating at variable speeds through the use of ASDs. Most of these converters contain a diode rectifier front-end and DC-link capacitor to convert the incoming AC voltage to a low-ripple DC voltage [133]. Contrary to machines, in case of three-phase rectifiers, there is no simple relation between the negative-sequence voltage and the negative-sequence current. This is related to the non-linear nature of these devices. The concept of impedance is hard to apply to power electronic converters. For a DC current source the effect of voltage unbalance is small. For AC voltage source the effect is much bigger. The current pulse through the diodes is determined by the difference between the peak AC voltage and DC voltage. For a three-phase rectifier the current shows 6-pulses, two for each phase-to-phase voltage. An unbalance in voltage thus causes a significant unbalance in current [32]. Under the conditions of utility voltage unbalance, also the input current harmonics are not restricted to the converter characteristic harmonics, and uncharacteristic triplen harmonics can appear, such as the 3rd and 9th harmonics [133]. In transmission network, for example the SVC compensator normally controls all three phases of output currents together. In unbalanced voltage conditions, the compensating currents in each phases would become different [148]. The harmonic content would be different in each phase and their normal cancellation through delta connection would not take place [148]. The consequences of the operation of STATCOM under unbalanced system conditions are similar. In unbalanced situation, the STATCOM will draw a negative-sequence current component as well as a 3rd harmonic current component, which are clearly unwanted [148].

REPORT

of 158

Therefore, unbalance condition of the power supply affects the converter performance of STATCOM, causing additional distortion of waveforms [149]. Considering the HV grid, the presence of negative-sequence in the three-phase voltages and currents also causes the operational behaviour of converters to suffer from oscillation [150]. The oscillation can lead to over-currents in the power converter and the failure of devices. In case of failure of a power converter, the stress level of any other PE device will rise. This may lead to further failures of power converters and critical stability issues of the electric network. The injection of negative-sequence currents in the transmission network has a negative contribution to the transmission network in overall by decreasing the maximal active power transmission. Unbalance in voltages and currents leads to an uneven heating of cables and lines, and therefore increasing the losses in the transmission lines [32]. The excessive current in one or two phases of the system may force the overload-protection devices to trigger [133]. The increased current can also decrease the lifetime of the capacitor or require the use of a larger capacitor [133].

2.6.4 Mitigation Methods The best manner for maintaining voltage balance in the network would be to prevent the voltage unbalance in the first place. Unfortunately, sustaining an exact voltage balance on all three phases at the point of use is virtually impossible for the following reasons [140]:

• Single-phase loads are continually connected to, and disconnected from, the power system; • Single-phase loads are not evenly distributed between the three phases; • Power system may be inherently asymmetrical.

As previously identified, the voltage unbalance is more likely a greater concern in distribution networks, rather than in transmission networks. In case of unbalanced loads, the great deal can be gained by attempting to distribute single-phase loads equally across all three phases [151]. The distribution systems can also be balanced by changing the system configuration through manual and automatic feeder switching operations to transfer loads among circuits. This reconfiguration can be performed to reduce losses, and has the natural tendency to balance loading among circuits [133]. Voltage unbalance contributed by unbalanced impedances, such is transformers and their connections, can be diminished by proper selection of distribution transformers. Particular attention should be payed to the balancing of open wye and open delta transformer banks. It should also be noted that open wye, open delta banks can significantly magnify the voltage unbalance of the primary system as it converts primary system zero-sequence voltage into secondary system negative-sequence voltage [152]. Voltage unbalance contributed by not systematically transposed transmission lines asymmetrical impedances can be mitigated by implementing the correct transposition of the lines, which in several cases [146] brings the existing voltage unbalance level down. To compensate the discussed effects of unbalanced three-phase voltages and currents in transmission and distribution networks, and to ensure a higher PQ it is requisite to avoid the injection of negative-sequence currents and balance the three-phase voltages. Unbalance compensation can be done by implementing filtering. The load current can be balanced by adding reactive elements in parallel to the load [133]. For variable loads, voltage unbalance in AC supply system can be corrected

REPORT

of 158

by means of a shunt connected thrysitor-controlled SVC, where again the load current is balanced by adding reactive elements in parallel to the load [133]. There are also some disadvantages related to the implementation of compensation filters, which include harmonic injection into the AC system [133]. Since voltage unbalance is often developed within user facilities, then much for its mitigation can be done close by the source. First, all overload equipment should be corrected. In addition, passive power filters and shunt connected SVCs can be used for unbalance compensation in user facilities also. Mitigation of the adverse effect of unbalanced voltage on ASDs can be achieved through the use of properly sized AC-line and DC-link reactors. Connecting both the AC-line and DC-link reactors to the ASD has the greatest effect on phase-current unbalance, reducing it more than half. Reactors will also improve the performance of Pulse Width Modulation (PWM) rectifier ASDs in unbalanced voltage conditions [133]. Power system overload conditions should always be corrected as soon as possible for protection and safety reasons as well as unbalanced compensation. Furthermore, the principle of load balancing with electrical vehicle chargers and PV power plants can be used to achieve a balanced three-phase system [153].

2.6.5 Practical Studies There are numerous available voltage unbalance practical studies covering low-voltage and medium-voltage systems, however there is lack of research results available for transmission level. One of the reasons for this is that voltage unbalance mostly relate to equipment associated with distribution network. In addition, the distribution network costumers are directly and more easily affected by voltage unbalance in the system. Therefore, the distribution system operators have higher responsibility to assure PQ requirements, which include PQ monitoring and control. Nevertheless, some research include also the transmission network due to special load types (single-phase loads). Report [62] presents PQ data gathered from several different countries. One of the objectives of [62] was to collect different measurement data from survey results from past or ongoing surveys across a number of monitoring points. The gathered available statistical data covers 99 sites at MV, 76 sites at HV and 25 sites at EHV level. Voltage unbalance measurement data revealed that most commonly available index for voltage unbalance average over 10-min periods is the 95% for a weekly statistics. For some surveys, statistical parameter for 99% over time are also available. In [62], it is concluded that system-wide unbalance surveys and routine measurements are uncommon and the available statistical parameters are limited. Regarding the measurement results, the following conclusions were made [62]:

• None of the reported MV sites exceeded 2% voltage unbalance level; • For HV sites, more than 11.8% of sites reached or exceeded 1% voltage unbalance, while

less than 4% of sites exceeded 1.5% voltage unbalance; • At EHV, only 1 out of 25 sites showed voltage unbalance exceeding 1.11%.

As an example from Australian sub-transmission networks, the voltage unbalance as a PQ problem has been observed and data covered in [146]. The 66 kV interconnected sub-transmission network has been noted to exhibit voltage unbalance levels up to 2% at some busbars during the peak demand periods. The voltage unbalance levels in the system have decreased after balancing the

REPORT

of 158

loads at some of these busbars, but the improvement has not been significant enough to respond the code requirements [146]. The 66 kV sub-transmission network under the study is connected to the EHV transmission network where the voltage unbalance was measured to be negligible. As mentioned previously the highest influence on voltage unbalance in transmission network level originate from heavy single-phase loads that are directly connected to the transmission network, such most common loads are railway traction supply systems. Therefore, available studies made in transmission level covering the railway systems are further discussed. There are variety of railway system applications as in Europe and around the world. Implemented traction system applications in different countries and regions manly originate from historical development and availability of the technology. In European counties, like Sweden, Norway, Germany, Switzerland and Austria operate their railway systems with 15 kV 16.7 (162/3) Hz, decoupled from 50 Hz system by rotating or static converters [136]. Many countries in Europe are using 3 kV DC system, for instance most of the Spain, Italy, Czech, Poland, Baltic States and part of Russia. In these cases, the unbalance induced by the railway make no concern. On the other hand, significant part of existing European railways and new high-speed trains are supplied with 25 kV 50 Hz AC. These countries are for example Finland, Denmark, parts on Great Britain, France and Spain (high-speed trains) and most of the east and south-east countries. The 25 kV 50 Hz AC traction system is connected between two phases of the transmission network, where the third phase of the transmission network remains non-affected by the traction load. Such power supply system is the source of voltage unbalance in the transmission network level, where the voltage unbalance degree depends on the train motion, load condition and power supply system configuration. In addition to the applicable standards, the voltage unbalance level requirements are regulated by each system and its operator. For instance in France in case of high-speed trains, the proposed values are as follows: respectively VUF≤1% for periods higher or equal to 15-min and VUF≤1.5% for periods less than 15-min [143]. The voltage unbalance limits in England and Wales transmission network are discussed in [154]. On voltage levels above 150 kV in England and Wales there is 1.5% limit and in Scotland 2%. On lower voltages, the voltage unbalance limit in Great Britain is 2%. In case of traction system, it is important to notice that trains are not behaving as usual loads in transmission network. Beside the fact that traction load is a single-phase load, it is also variating and moving in time, therefore causing the voltage unbalance situations in the connection points to the utility grid. There are several studies available from different countries in order to determine the unbalance problems originated from traction systems. In [136], there have observed voltage unbalance levels caused by railway loads in two geographical areas of Finland, Kainuu and Helsinki. Kainuu represents weak transmission network area, whist Helsinki, as the opposite, is electrically strong system area. Besides the considerably weak transmission network, the Kainuu area can be described as light loading density and heavy, punctual ore trains. During train traffic, the observed voltage unbalance level arose typically to 0.5-1% (non-traffic situations the level was between 0.1-0.3%). In the research [136] there were also considered some worst case scenarios (due to transmission network connections, train loading and speed), where the voltage unbalance average 10-min values was between 0.2-1.9%. In addition to the 10-min values, 15-s recording were captured, which showed that in worst case the momentary voltage unbalance can exceed the 2% requirement. The railway traction power supply system in Helsinki is

REPORT

of 158

similar to the Kainuu area. On the other hand, the railway in Helsinki is used only passenger traffic and no heavy good trains are passing these substations. In Helsinki city area the transmission network is connected meshed, therefore the network can be considered as strong. The measures of voltage unbalance were taken during the rush hour when the train traffic was dense. The voltage unbalance variated between 0.1-0.5%. During the non-traffic in night-time, the measured unbalance levels stayed between 0.2-0.25%. The transmission network stiffness strongly explains the various unbalance observation in Kainuu and Helsinki. In Kainuu, the short-circuit power in the measurement point was 180-720 MVA and in Helsinki 5,000-6,000 MVA. Nevertheless, the research [136] indicates that in both cases the voltage unbalance caused by traction load is not exceeding the permissible 2% voltage unbalance level in transmission network connection point. In order to study the voltage unbalance propagation in a power system, the test network with 63- buses (called CIGRE 63) was considered and the results are represented in [142] and [143]. In the test system, the railway connections to the 220 kV and 150 kV voltage levels were considered. In different case studies in [143], the results showed that voltage unbalance occurrence depend on the peak demand of the traction loads and on the short-circuit power of the grid. In all cases, the VUF in transmission network substations was mainly between 0.03-1.55%, with some exceptions where the VUF reaches to the values 2.05% and 4.25%. This high value of the voltage unbalance was because of the weak power system in those observed points [143]. In Denmark, the impact of single-phase railway system to the 132 kV transmission network is studied in [155], where the analyses base on simulation results. Generally, the voltage unbalance for traction applications in Denmark is limited to 1.4% [155]. The voltage unbalance was recorded in four utility substations and results showed that in half of the substations, the unbalance level were in acceptable range. During the maximum load demand in two of the observed substations, the voltage unbalance limit were exceeded, respectively 1.89% and 2.42%. In North American power networks, the voltage unbalance is regulated by standard ANSI C84.1 [141]. In the standard, it is stated that 98% of electrical utilities operate less than 3% voltage unbalance and 66% operate less than 1% voltage unbalance [156]. ANSI C84.1 also recommend that electric supply systems should be designed so that the maximum voltage unbalance level is limited to 3% (end user) measured at the electric-utility revenue meter under no-load conditions [156]. In New York City, the single-phase traction load impact to 115 kV transmission network is evaluated in [157] for urban transportation and as well intercity trains between New York and Boston. The evaluation in [157] expresses the 17 days measurements of voltage unbalance at 115 kV, where unbalance average value was 0.3%, and the maximum value was 0.8%. The measured voltage unbalance in these cases stayed within the acceptable limits. In Taiwan, the high-speed railway system is using single-phase supply from 161 kV transmission network. In [158] there is analysed the Taiwan transmission network voltage unbalance in railway system connection points based on simulation results. Under normal operation conditions, the unbalance factor remained in range 0.19-0.35% (considering different types of traction transformers) and 0.15-1.52% when considering the N-2 criterion and most serious unbalance impact scenario [158]. The result conclude that even in worst case scenario the voltage unbalance remained in permissible range.

REPORT

of 158

China heavily loaded railway system is analysed in [159]. The traction system is supplied via single-phase transformers from 220 kV transmission system. The VUF results are calculated based on the actual operating data, where the traction power supply is integrated into strong power system with large short-circuit capacity. If there is large traction load in the system, the VUF reaches up to 0.25%, which satisfy the required less than 2% criterion demand [159]. The study represented in [160] analysis the voltage unbalance in Korean high-speed railway with commercial operation conditions, where the voltage unbalance measurements have been taken from traction system connection point substations for two different railroads. The mean value of 10-min measured voltage unbalance was between 0.15-0.69%. The limitation for the voltage unbalance in Korea is 3% as the mean value of 2-h (at that time in 2014). From results presented in [160], the voltage unbalance is satisfied in the limit value. It is confirmed that there are no problems for voltage unbalance. In India, the typical single-phase railway supply system is connected to the transmission network in 110 kV voltage level [161]. Reference [161] represents the study made in Indian traction system, where the voltage unbalance values are mentioned in two observed substations. The results conclude that in one of the observed substations the voltage unbalance works out to 1%. In the second substation, the results were slightly higher than the specified 2% limit. The higher voltage unbalance persist for only short period i.e. less than 10-min.

2.7 Harmonic Distortion

2.7.1 Background Harmonics are sinusoidal voltages or currents with frequencies which are integer multiples of the fundamental frequency in the network [162]. Beside the mostly occurring (odd) integer harmonics in the power system, there can be also a non-integer multiples of the fundamental frequency in the network. The amplitude of the non-integer ones, in normal cases are relatively low. With a high penetration of PE devices and non-linear loads, current waveforms are distorted by the injected current harmonics. It can be stated that with a higher amount of injected harmonics, the PQ within the transmission network suffers [163]. A brief history of electrical systems shows that problem of harmonic components or non-sinusoidal waveforms in electrical networks persist since the beginning of AC networks. The issues like heating of motors, influence on the communication lines etc., were indicated and analysed already in the end of the 19th century. Although the measurement systems were not sophisticated yet, the knowledge about harmonics was expanding. At that time, the harmonics were caused mainly by saturable iron cores and non-linear motor loads [164]-[165]. Nowadays, the complexity of the harmonic issues grows as the electrical networks are becoming more and more saturated with non-linear loads, based on cost-efficient but harmonically rich PE devices. The use of power electronics has been widely accepted at all voltage levels, starting at low voltage with household appliances, computers, ASDs etc., at medium voltage with traction drives and lately more and more at the high voltage levels as the FACTS devices, PE based HVDC systems, converters for larger loads or generators, arc furnaces etc. Beside to the saturable iron cores of the transformers, these modern electronic devices have become the main reason of the harmonic

REPORT

of 158

distortion in both, the transmission and distribution level. The effect of the harmonically distorted loads to the network voltage is not limited to the characteristics of the load itself, but will be heavily affected by the “strength” of the network, or in other words by the characteristic impedance of the network. The PQ of the transmission network is influenced by sources of harmonic emission, such as non-linear loads and power injection by converters [166]-[167]. Power converters inject harmonics due to their switching operations. Since power injection by renewable energy and thus, the operation of the associated PE devices is dependent on the daytime and weather conditions, a time and spatial dependency of the level of injection harmonics can be observed. Furthermore, the total harmonic distortion of the transmission network may also change over time due to changing load and weather conditions. Consequentially, also the time and spatial dependency of the PQ within the electric network can be detected. Capacitive and inductive elements in power systems interact with each other and create resonances at selected frequencies. These resonances can create potential risks of high currents or voltages at a certain frequency, which can lead to severe damages of the electrical equipment. Transmission networks are particularly exposed to harmonic resonances, because of their low resistance values causing low resonance damping. Systems with long cable lines or multiple parallel lines need an extra attention, since the system capacitance is increased, and consequently it lowers the system resonance frequency. In critical cases, the system impedance characteristic can exhibit resonances near integer harmonics. In addition, due to possible presence of broadband distortion spectrum, included by non-linear loads (e.g. SVC, HVDC, arc furnaces, UPS, etc.) or generic control strategies of PE devices, the interharmonic resonances could lead to severe stability issues [168]. Additionally, in case of PE devices based networks, where the generators and loads are fed through converters, the converters do not provide resonance damping, since the natural resistance of the system is low. The main sources of harmonic distortion are as follows:

• Saturable devices, due to physical characteristics of the iron core (transformers, rotating machines, non-linear reactors);

• Arcing devices, due to the physical characteristics of the electric arc (furnaces, welders, fluorescent lighting);

• Power electronics, due to the semiconductor switching which occurs within a single cycle of the power system fundamental frequency (VSC, DC motor drives, electronic power supplies, rectifiers, inverters, SVCs, HVDC transmission).

Harmonics are typically characterised using Total Harmonic Distortion index (THD), though other indices are also in use in different applications (e.g. voltage or current crest factor, telephone interference index, number of zero crossings per cycle, etc.). As the influence of harmonic distortion at the low voltage and distribution level has been widely analysed and the emissions standardised (IEEE 519 [169], IEC 61000-3-6 [139]) there is still lack of insight to the harmonic matters of the “future” modern transmission networks. This is as well the main topics of the CIGRE JWG C4.24, which mainly addresses issues with PE interface equipment at different voltage levels, evaluates new measurement equipment and techniques, and among other, evaluates how these PQ issues impact the transmission system. In addition, CIGRE/CIRED JWG C.40 is currently reviewing IEC TR 61000-3-6, IEC TR-61000-3-7 [72] and IEC TR 61000-3-13 [170].

REPORT

of 158

The number of high PE devices that are connected to the transmission network is drastically increasing. Especially high power HVDC links that interconnect isolated or remote areas, systems with different operation frequency and offshore wind parks are in the last decade one of the major contributors to the expansion of the PE devices at high voltage levels. Due to the large impact on the reliability, the modern networks with high amount of PE devices require new researches, studies, and technical recommendations, which will help the TSOs to cope with more severe operating conditions of future networks.

2.7.2 Evaluation Indices Standards dealing with the harmonics in electrical networks are more focused on distribution networks, where the majority of consequences due to harmonic distortions take place. At the distribution level, the focus is on the compliance of voltage harmonics with standardised values, compatibility levels for disturbances in low voltage and medium voltage networks (EN 50160 [8], IEC 61000-2-2 [171]) and compliance of a connected equipment (IEC TR 61000-3-6 [139], IEEE 519 [169]). However, at the transmission networks (HV-EHV), there are IEEE standard, IEC technical recommendation and EN standard (up to and including 150 kV level) that define the following:

• IEC TR 61000-3-6 defines indicative planning levels for HV and EHV for each individual harmonic order and total harmonic distortion, assessment of emission levels, principles for summation of harmonic disturbances within the network, methods of sharing planning levels and allocating emission limits in meshed HV systems;

• IEEE 519 standard offers the recommendations for individual consumers defining exact current distortion limits that are applicable to individual device connected to different voltage levels; in the recommendations for utilities, the voltage distortion limits for longer and shorter periods is defined as well as other principles of harmonic summations, interferences with communication lines etc.;

• EN 50160 gives an indicative values of individual harmonic voltages at the supply terminals for harmonic orders up to 25;

The Engineering Recommendation (ER) G5/4 [172] is technical document, which gives recommendations to the harmonics in the electrical networks (used in most of the Commonwealth countries). The approach for defining the emission limits to new customer connections is slightly different from the IEC principle, i.e. G5/4 uses the “first-come-first-serve” apportioning method whereas IEC relies on an “equal rights” principle where emission limits are apportioned on a per-MW pro-rata basis. The planning as well as compatibility levels are defined for all voltages up to 400 kV. It should be noted that G5/4 is currently under review in UK. The IEC technical recommendation defines the levels of harmonics on the basis of the compatibility levels (for LV and MV only) and planning levels (for all voltage levels), where planning levels are lower or equal to compatibility levels. Planning level is defined as the level of a particular disturbance in a particular environment, adopted as a reference value for the limits to be set for the emissions from the installations in a particular system, in order to co-ordinate those limits with all the limits adopted for equipment and installations intended to be connected to the power supply system. The planning levels defined in standards are stated as the “indicative values” and are therefore not compulsory for the use (e.g. by the TSO). The limit values of individual harmonic, are in comparison to the limit values at the MV levels lower, due to expected full propagation of the harmonic voltages from (E)HV to MV networks with addition of harmonic distortion that is generated at the lower voltage

REPORT

of 158

levels. Planning levels provide the reference for the limits for emissions pertaining to a certain system giving the TSOs chance to set limits of emissions for individual consumer. The general methodology for calculating the system harmonics indices can be divided into three steps. First, calculating the different spectra of the voltage and currents over a window of time, then calculating the required indices from the spectra for different sites, and finally calculating the total system indices from the sites indices. Several indices are developed to describe the harmonics phenomenon. The most common indices are the THD for the voltages and currents, and they are calculated by following Equation 2-6 and Equation 2-7.

123� = �∑ �4�546��7 (2-6)

1238 = �∑ 84�546�87 (2-7)

Where h is the harmonic number, Vh and Ih are the harmonic voltages and currents, and V1 and I1 are the fundamental voltage and current. Other harmonics indices are developed for more specific applications. The Total Demand Distortion (TDD) is developed to describe the harmonics performance in case of low fundamental current, where the THD could be misguiding. In addition, telephone interference factor is developed to describe the harmonics performance when it affects the audio and communication system (high harmonics orders). The factor-K index (similar to K-factor used in USA) is describing the derating of transformers under the harmonics presence. There is also developed a zero crossing factor to describe the impact of harmonics in the equipment that works on the concept of waveform changing from positive to negative or vice versa (e.g. contactors and electronic clocks). The main steps in the harmonic level evaluation process are summarised in the following points [62]:

• Calculating the spectrum over a 10 cycle period (for 50 Hz system), with consideration for synchronizing the window with the actual frequency;

• The RMS values of the obtained spectrum are aggregated over a 3-s periods to create the “very short time index” (Vh, vs);

• The 3-s RMS values are aggregated into a 10-min value, which gives the “short time” index (Vh, sh);

• The “short time” and “very short time” indices are evaluated over longer periods of study, typically a day or a week.

The calculated indices can be described statistically and evaluation of the voltage can be carried out by using one or more indices. Especially during shorter periods of higher distortions (e.g. bursts or start-up conditions) additional indices may be used:

• The 95% weekly value of Vhsh (RMS value of individual harmonic over 10-min period) should not exceed the planning level;

• The greatest 99% probability daily value of Vhvs (RMS value of individual harmonic component over 3-s period) is defined only for low and medium voltages and does not apply to the (E)HV networks

REPORT

of 158

The harmonic voltage planning levels are equally provided for the HV and EHV levels. Indicative planning levels of harmonic voltages are brought out in Table 2-7 below. Table 2-7: Indicative planning levels for harmonic voltages (in percent of the fundamental component) in HV and EHV [139]

Odd harmonics non-multiple of 3

Odd harmonics multiple of 3

Even harmonics

Harmonic order

Harmonic voltage

Harmonic order

Harmonic voltage

Harmonic order

Harmonic voltage

5 2% 3 2% 2 1% 7 2% 9 1% 4 0.8% 11 1.5% 15 0.3% 6 0.4% 13 1.5% 21 0.2% 8 0.4%

17…49 1.2 ∙ 17ℎ % 21 … 45 0.2% 10…50 0.19 ∙ 10ℎ + 0.16%

The planning level for HV and EHV levels is THDHV-EHV=3%, while for medium voltage THDMV=6.5%. Apart from the IEC TR 61000-3-6 and IEEE 519 there are also other national standards that define the voltage harmonic level limits as well as current harmonic limits for individual device to be connected to the network at different voltage levels [173]. To reduce the overall voltage harmonic distortion in the transmission network emissions of individual devices connected to the network needs to be analysed and limited. The IEC TR 61000-3-6 does not directly define the emission limits for individual customer (device) whose agreed power is less than 0.2 % of the short-circuit power (stage 1), while for the installations exceeding 0.2% value (stage 2), the emission limits need to be set in accordance to the share of the planning level that is dedicated to the specific customer. In this assessment, it is necessary to consider also other factors that contribute to the future variation of the harmonic distortion, especially other planned and unplanned high power distorting devices, new network topologies that could drastically change the impedance characteristics of the system and consequently increase (or reduce) the emissions from connected devices and amplify the system voltage harmonics. Still, the installations that do not comply with the emission limit set in stage 2, may not be connected to the network except in case when the customer and the system operator agree on the detailed study of the actual and future system characteristics which determines the special conditions, expected emissions and influence on system and other customers. The standard IEEE 519 on the other hand defines fixed limit values of the individual harmonic current component that are based on the ratio between maximum short-circuit current and the maximum demand load current at the point of common coupling. Similar approach is presented in [173]. The practice shows that the TSOs to allocate the emissions between new customers is becoming more and more important and complicated. The emission limits for the future installations will definitely be stricter than they are now when the number of PE devices is still relatively small. Some TSOs allow the first “distorting” installations higher emission limits, leaving less space for emissions of the “newcomers”.

REPORT

of 158

2.7.3 Consequences The presence of harmonics in the electric power system causes problems, which are, according to the IEC 61000 technical recommendation, separable into two categories:

• Harmonic currents injected into the system by converters and other sources, which give rise to harmonic voltages;

• Harmonic currents, which induce interference in the communication systems; higher order harmonics are more problematic in this aspect.

The consequences of the harmonics at the transmission network can be observed directly at the equipment connected to the transmission network, as well as on the lower network levels where harmonic distortion occurs due to the harmonic propagation from the high voltage levels. The increasing integration of PE devices in the transmission network causes harmonic pollution and leads to distorted phase voltages, phase currents, additional power losses and overheating of electric equipment [174]. Propagation of harmonic currents results in additional heat loss and as a result, the need to de-rate power system equipment, such as conductors, transformers, motors etc. The capacitive network elements can be affected even more. Due to lower impedance at higher frequencies or in resonant combination with other inductive elements, the harmonic current can increase dramatically, which contributes to additional heat losses and reduction of network elements’ lifetime. The risk of damaged power system equipment causes stability problems in transmission network. In the case of electric machines (generators and motors), the operational behaviour is sensitive to non-sinusoidal waveform. The injected harmonics cause rotor overheating, pulsating torques and noise [175]. The most important harmonics to mention in this case are the 5th and 7th number. For instance, in [176] it has been noted that harmonics in the transmission network can deteriorate the performance of DFIGs by introducing inaccuracy in the generation of controlled active or reactive power. Capacitors, when supplied with non-sinusoidal voltages suffer higher heat loss as they provide low impedance to harmonic currents. Even the incandescent lamps suffer definite loss of life when operated with distorted voltage. In extreme cases current resonance conditions been reported to have caused power outage and damage to circuit breakers [177]. Other examples are factor-K, which is applied to de-rate standard transformers in the presence of harmonic currents [178], and the de-rating conductors due to skin effect [179]. Performance of energy meters is reported to be affected by harmonics of current THD>80% and voltage THD>2% [180]. All these effects have a major impact on the performance and lifetime of the electrical network equipment. Thereby, the possible malfunction of equipment affect highly production processes. More detailed overview of the problems that may arise from harmonics are following [165]:

• Lowered efficiency of generation, transmission and utilisation of electric energy; • Increased heating of system components; • Harmonic resonances resulting in excessive harmonic currents and voltages; • Increased crest factor (max/RMS value); • Aging of the insulation due to increased voltage and temperature; • Stress on capacitor banks; • Multiple voltage zero crossings, which may affect synchronization and lower accuracy of

metering;

REPORT

of 158

• Effect on electromechanical and electronic protective relays; • Induced voltages and currents in the nearby metallic structures and communication lines; • Degrading communication system performance; • Vibration of power system components causing audible noise (transformers with laminated

core); • Disturbing moments and noise in rotating machines; • Current flow in ASD through stray capacitance and motor bearings causing failure; • Triplen harmonics may overload neutral conductor in LV networks; • Possible misoperation of the protection systems based on different analogue or digital

technologies. In [165], the author specifically mention that harmonics cause operational malfunction of some of the measuring and protection equipment as well as an overloading of a harmonic filter. The costs that arise due to consequences of harmonics at the transmission network arise at the side of the supplier (TSO) and the customer (industry, etc.). The costs can be separated into the following points:

• Costs due to equipment failures; • Costs due to equipment decreased life span expectation; • Costs due to customer’s interruption (industry, services, etc.); • Costs due to contractual penalties due to non-optimal operation (high THD, low power

factor); • Costs of new harmonic mitigation installations.

2.7.4 Mitigation Methods Mitigation of harmonics in the transmission network is the process that involves in-depth system analyses of the network topology and deployment of active and passive compensation devices capable of harmonic filtering [181], resonance damping [182]-[183] or other control-based operations. Harmonic distortion emerges from non-linear network elements or loads that consume non-linear current. Uncompensated harmonic currents can flow through network impedances and cause different voltage distortions at different nodes, based on the specific impedance characteristic. Due to the nature of electrical networks (combinations of network shunt capacitances and series inductances), the resonant points (series or parallel) of the network impedance can lead to severe amplifications of harmonic voltage or current at specific element or node. The mitigation of harmonics can be based on network or device mitigation solutions. For the network solutions, the harmonics can be mitigated by shifting the resonance frequencies to safe bands. This can be performed by changing capacitor (main cause of resonance) locations and/or sizes. Providing higher short-circuit capacity at the disturbing load connection node also helps to shift resonance frequencies in network. Equation 2-8 [162] shows the relationship between the short-circuit level and the size of connected capacitor at a bus.

ℎ = �CDDE&FGE& (2-8)

As it can be seen from the Equation 2-8, the increased short-circuit capacity SCC pushes the resonance frequency hr to a higher frequency band, which in a normal situation has lower levels of injections. Contrary, connecting large capacitors Qc leads to reduction in resonance frequency.

REPORT

of 158

The device based harmonic mitigation solutions can be divided into two main techniques, installing devices that are able to block or trap harmonic currents (filters) or reconfiguring the existing devices to have improved immunity and performance in terms of harmonics. In general, active and passive filters can be used to damp harmonics. It is well proven solution if filters are properly designed and located. However, the usage of filters causes additional resonances and economic costs, for what reason filters are not always desirable. Passive filters are week elements, as capacitors cannot be overloaded, because of this they could lead to premature ageing or failures. The second category involves increasing neutral size (reduce zero-sequence harmonic losses), utilising transformer connections (delta and zigzag connections) to block harmonic currents flow, improving converter harmonic injections (by increasing the number of pulses and adopting PWM techniques for harmonic cancellation), installing smoothing reactors and tuning and/or detuning of capacitor banks [1], [162]. It is worth mentioning that the power converters are purely controller driven and the production of harmonic components follows the strategy of controller. Therefore, there is no guarantee that the triple harmonics, which in passive systems are mostly symmetrical, will be symmetrical in PE devices or that no even harmonics will be generated under steady-state conditions. These facts show that in order to reach the stable operation of PE-based devices it is necessary to improve operation of the controller, which should guarantee the stability under all kinds of operation scenarios. One of the common approaches used to mitigate harmonic voltage distortion and to control the propagation of harmonic currents is by installing passive harmonic filters at strategic network buses [162], [184]-[185]. Passive harmonic filter is constructed from passive elements such as resistors, inductors and capacitors. There are generally two approaches to suppress undesired harmonic currents using passive harmonic filters. The series filter uses series impedance to block harmonic currents whereas the shunt filter diverts harmonic currents by means of low impedance shunt path [185]. Series filters are not commonly used, as they must carry full load current making them more expensive compared to shunt filters. Such filters in the applications of converters harmonic mitigation are usually single-tuned to the 5th, 7th or 11th harmonic frequencies [186]. Different combinations or parallel and series RLC components can be used to form wideband or high pass filters to eliminate high frequency harmonics [186]. In order to increase the capacitance of power converter applications in transmission systems, the switching devices are combined in series and in parallel. For the harmonic mitigation purposes, it is more beneficial to combine converter units rather than switching devices. The simplest method for eliminating harmonics by combing converters is by using coupling transformers with different angle shift. It is performed by using n transformers with primary or secondary windings shifted by 60°/n. The most widely used is the 12-pulse connection, which consists of two 6-pulse converters connected via a star connected and a delta connected secondary windings transformers. The 30° phase shift between the transformers results in the elimination of the 5th and the 7th harmonic currents in the primary side, and elimination of the 6th harmonic in the DC side of the rectifier [186]. Multi-stepped voltage waveform can be obtained by using multi-level converters. In such structure, the DC bus capacitor is split into number of levels, typically 3 to 7, and applying topologies such as Diode Clamped Capacitor Multilevel Inverter (DCMLI). The number of levels can be selected to improve waveform distortion and output voltage control [187].

REPORT

of 158

Another technique used for harmonic elimination is the PWM. Controlling the harmonic contents in the voltage waveform can be performed by increasing the notches in the current and voltage waveform. The harmonic contents of a converter pattern depend upon the position, width and the number of notches. Typically, one notch per quarter a cycle allows the elimination of one harmonic component. This technique is known as Selective Harmonic Elimination (SHE). In literature [188]-[190], several mitigation techniques for injected harmonics have been presented. In [188] there have implemented a discrete observer control scheme to reduce voltage harmonics in three-phase systems with non-linear loads. The control loop of the converter is extended by a harmonic feedback, which allows reducing output voltage harmonics and mitigates voltage unbalance caused by for instance, unbalanced loads or unbalanced power injectors. In [189], the design of shunt passive filter and shunt hybrid power filter for reducing harmonics caused by non-linear loads is presented. In [190], a sliding mode harmonic compensation is proposed to improve the performance of grid-connected converters. The presented method includes a fourth order band pass filter in the reference frame to detect harmonics without phase delay. Furthermore, a sliding mode harmonic current control with fast dynamic response is implemented to suppress harmonic currents generated by electric devices in the grid. Some of the techniques applied to reduce or eliminate harmonics can be listed in the following [162]:

• Utilising higher pulse number converters for phase cancellation and elimination of lower frequency harmonics;

• Relocate and resize capacitor banks to avoid resonance magnification; capacitors can also be utilised for harmonic elimination by the means of tuning and detuning;

• Delta and zigzag transformer connection to block the propagation of zero-sequence harmonics;

• Chorded coils and redistribution of windings can reduce the harmonics in synchronous machines;

• The use of different types of filters and chocks.

2.7.5 Practical Studies Harmonic studies in transmission networks typically fall into one of the following categories:

• Planning assessment of new customer connection methods and allocation of emission limits; responsibility for these studies normally lies with the system operator;

• Design and technical assessment of a new facility to demonstrate compliance with the allocated emission limits, responsibility for these studies normally lies with the customer;

• Forensic investigation studies in response to incidents or customer’s complaints; responsibility for these studies normally lies with the system operator, although collaboration with customers may be required to identify the dominant source of harmonic distortion.

Each of the above categories has different inputs and outputs and there are several possibilities on how harmonic studies could be performed. Furthermore, the scope, roles and responsibilities for each type of study may differ from country to country depending on specific grid code requirements and applicability of PQ standards or internal policies. As an example, [191] describes roles and responsibilities and the process for exchange of harmonic data between the system operator and the customers in Ireland, which aligns with IEC TR 6100-3-6

REPORT

of 158

[139]. A similar, but not identical, process is described in G5/4 [172] for customer connections in the UK. There is large number of recent publications reporting practical studies related to harmonics in transmission networks. This trend, along with the creation of numerous international working groups, reflects the increasing attention on PQ at transmission voltage levels and particularly the need for assessment of the harmonic voltage distortion effects of HVDC, wind parks, PV and long HVAC cable connections into transmission systems. As an example, references [192] and [193] provide thorough descriptions of the assessment methodologies adopted by the system operators in Ireland and UK, respectively, for the allocation of emission limits to wind parks and HVDC connections. Some examples of the type of technical studies performed by customers to achieve compliance with the given emission limits are given in [194]-[197]. Furthermore, examples of technical assessments aimed to mitigate the impact of reactive power compensation devices and long cable connections are given in [198]-[203]. Some general discussions on the key data requirements and common approaches for harmonic studies are included in [204] and [205]. One of the key issues is the consideration of the extent and level of detail of the transmission network model. The grid could be represented in whole detail, as a single frequency dependent impedance value or with impedance loci. As it is impractical to hand over the complete grid data to customers, this option is unimportant and not normally used in practice other than for studies performed by the system operators themselves. Very common is the provision of a frequency dependent impedance curve that can be integrated in a simulation tool to perform harmonic load flow simulations. However, the use of only one frequency dependent impedance value needs to be taken with caution as it may not reflect the true conditions during a range of real operation scenarios. In reality, normal operation conditions include circuit maintenance outages, transformer tapping, changing status of reactive compensation devices, generation unit commitment, load variations etc. These changing grid conditions will affect the harmonic impedance of the grid, therefore it cannot be considered as a single value for each harmonic order. In addition, slight deviations between the real system and simulations models will also appear. To consider all these conditions and uncertainties, many system operators provide the harmonic impedance data in the form of boundary shapes, such as circles, pie-shape sectors or polygons, in the R-X plane. These boundary areas are normally referred as harmonic loci and can be provided individually for each harmonic order or grouped in banded frequencies. The costumer is then required to demonstrate compliance with the allocated emission limits for all boundary points at each harmonic order. An algorithm for the systematic assessment and identification of the most onerous conditions for harmonic distortion at the connection point is detailed in [206] Independent from the format and extent of data exchanged between system operator and customers, the accuracy of the underlying models used to represent all assets is of big importance. Thus, cables, overhead lines, transformers, loads and generation units have to be modelled with high accuracy to capture their frequency dependant behaviour and damping. For example, thanks to modern personal computers with high computation capability there is no need or justification for the use of lumped pi-equivalents instead of distributed parameters. By using a pi-equivalent model, only the first resonance could be estimated and all higher resonance frequencies will be not visible in calculation.

REPORT

of 158

These modelling aspects are the focus of CIGRE JWG C4/B4.38 “Network Modelling for Harmonic Studies”. The outcome of this WG, expected to conclude by end 2017, will be a detailed study and modelling guidance document to aid power system engineers when performing harmonic voltage distortion analysis. Another important aspect in the practical assessment of harmonic emissions into the transmission system is related to the pre-existing harmonic background distortion at the connection point (or at the point of evaluation). Depending on the approach adopted by the system operator for assigning responsibility of background harmonic amplification, this information is provided (or not) to the customer. It should be noted that there is no harmonized approach amongst transmission system operators with respect to the responsibility for mitigation of background distortion amplification caused by interactions between the customer’s internal network and the main transmission network. In cases where this responsibility is assigned to the customer, the pre-connection background must be provided, although an alternative approach is the definition of a maximum amplification factor. One of the main challenges for system operators is the availability of sufficient measurements to provide a representative (or worst case) value to the customer. In most cases, this involves the installation of portable measurement devices a few months before the expected energisation date. A new challenge arises when the connection point is a new electrical node in the system with no background distortion information available and hence a need to estimate it based on measurements from nearby nodes. The minimum duration of the pre-connection measurement period is difficult to define but it should capture a wide range of operating conditions and system outages with significant impact on resonant frequencies, especially in the low harmonic order range. Besides background harmonic distortion, which is already in the grid, other harmonic sources have to be modelled. As harmonic sources are often modelled as ideal current source, it has to be mentioned that this approach does not generate reliable results for modern PE converters associated with WTGs. In contrast, a generator or PE device is never an ideal harmonic source. It contains at least frequency dependent impedance that interacts with the feeding network and should therefore be modelled as Thevenin- or Notron-equivalent. For a PE device, the input impedance of the converter should be modelled separately from the filters, series reactor and transformer to ensure results, which reflect the reality as best as possible. There are still several open questions, which have not been covered in the past in sufficient detail and require further research work. One of them is the applicability of the IEC summation laws [139] for aggregating multiple harmonic sources from modern PE converters. It is questionable if this summation law could be applied for example to large wind or solar parks where all harmonic sources are based on exactly the same technology. Furthermore, at the moment only worst case calculations are normally performed. As long as the limits are higher than the expected harmonic distortion this approach is suitable, but in extended cable grids with strong resonances and therefore high harmonic amplification, a different approach based on probabilistic may be necessary to avoid unnecessary investment in mitigation devices. In transmission networks, there are published studies available that present actual numerical results on harmonic distortion. CIGRE WG C4.07 [62] has gathered available harmonic measurement data. Analysis of the results show that the basic statistical parameter for comparing the data is the 95% 10-min value (Vhsh95) for the maximum site and 95%-site. From the available measurements results at HV, the planning levels for HV systems are exceeded in many cases, especially at 5th harmonic,

REPORT

of 158

where the harmonic level was often between 2.4-3.2% compared to the planning level 2% (IEC TR 61000-3-6 [139]). The harmonics measurement results can be seen on the Figure 2-10.

Figure 2-10: Low order harmonic voltages from 7 surveys at HV for Vhsh95 [62] Another available study on harmonics in the transmission level (150-400 kV) can be found in [165], for the Danish TSO. It mentions malfunctions of measurement and protection equipment and overloading of certain filters. The maximum levels of most intense harmonics measured are presented in Table 2-8. Table 2-8: Magnitudes of measured harmonics [165]

Harmonic order

Harmonic magnitude

3 0.5% 5 0.2% 7 0.3% 11 0.9% 13 0.65%

Other notable available studies on harmonics in the transmission network are [207], which deals with the impact of capacitor banks on the harmonic content in part of Australian system. Reference [208] handles with 5th order harmonics in a portion of the Japanese system, whilst proposing a new measuring method using state estimation. Reference [209] examines PQ in Brazilian transmission network, where can also see some measurements of the THD.

2.8 Harmonic Stability

2.8.1 Background This section is dedicated to the specific harmonic characteristics of modern transmission power networks. With the continuous development and changes of the transmission systems, TSOs are faced with the increased number of PE-based devices, as for example HVDC systems which are used

REPORT

of 158

to interconnect transmission systems, large WPPs connected to the remote grid by long cables [210] etc. It can be expected that the higher number of HVDC systems will be followed by more frequent use of FACTS devices for establishing optimal operating conditions, e.g. the STATCOMs or SVCs for power factor correction or voltage control. In addition, converter based generation systems are connected in large scale and through cables (representing large shunt capacitances), which in combination with the high power transformers (large inductances) additionally increases complexity of system impedances, forces system resonances to lower frequencies and contributes to amplification of harmonic components. As it is already experienced by many TSOs, the new network conditions show the disadvantages, issues, and challenges that are caused by the higher level of system complexity. Most of these are perceived in areas with high number of modern PE-based WTGs [211]-[212]. The PE-based devices serve as a source of harmonics, which are additionally influenced by the electrical network characteristics. In addition, the mutual interferences could cause unpredicted operation scenarios. The main difference between the PE-based devices and classic synchronous or asynchronous generators that are directly connected to the electrical system is their impedance-frequency characteristic. The classic drives are based on wound machines with windings, therefore they represent mostly inductive characteristic with higher damping of the high harmonic components, which reduces susceptibility to distortions from the outer network. On the other hand, the devices that are based on full-converter or semi-converter power transmission (e.g. DFIG) exhibit the full controller based impedance characteristic. In other words, the entire control process, from signal acquisition, processing to the converter switching, defines the impedance characteristics of the PE-based device. The main part in the controllable bandwidth is covered by the control algorithm that is considered as highly confidential for all manufacturers. The detailed insight to the control algorithms is mostly not available for third parties therefore a risk of unknown controller interactions is possible for all kinds of devices. The stability issues that can occur in the systems with high amount of PE-based devices (e.g. generators, HVDC) can initiate current/voltage components of different frequencies (sub- or super-harmonics). Sub-synchronous oscillations are known and can often be damped by power system stabilizers. However, oscillations in high frequency range are mostly unknown to TSOs. Beside the interaction between the grid impedance and controlled PE device, the harmonic stability issues could occur also due to the control interactions between two or more controlled PE devices (e.g. two parallel WTGs or SVCs or some other combinations). Since the control interaction could lead to stability issues, the co-ordination between controllers and the tuning needs to be done simultaneously to effect an overall positive improvement of all control schemes [213]. Interaction of control algorithms could lead to oscillations in different frequency regions, from 0 Hz to multiple of kHz, i.e. from steady-state interactions to high dynamic, harmonic interactions. In steady-state, the interactions could occur between system related controls which do not involve the controller dynamics (e.g. interaction between two voltage controllers, steady-state at operation limitations etc.). The oscillations in the range 2-15 Hz largely depend on the network strength and control parameters of PE devices. The case with multiple SVCs in voltage control mode has been presented in [214]. The oscillations due to sub-synchronous resonances, usually in frequency range 10-50 Hz, can as well

REPORT

of 158

interact with the controls of PE devices. On the other hand, when properly tuned, these could serve as active damper of sub-synchronous oscillations. Instabilities due to high harmonic interactions of PE controllers (>50 Hz) are likely to occur because of the amplification of harmonics within controller loops. The harmonic instability is caused mostly due to improper dynamics of the synchronization, voltage measurements system, main current control loops or other sections of the control algorithms [215]. High Frequency Oscillations

High frequency oscillations or harmonics are often considered as integer multiples of 50 Hz, which result from assets with a non-linear voltage-current characteristic or from PWM from power electronic assets. In addition to this, high frequency oscillations can also occur due to controllers, which interact with grid resonances [211]. This phenomenon is called harmonic stability as the symptoms are harmonics and can be mixed up with steady-state harmonics, but in this case controllers get more or less instable. Resonances in the grid cannot be avoided. They are defined by inductive and capacitive elements. The frequency range of the first resonance is mostly defined by the characteristic of the grid. Grids consisting of overhead lines will have a resonance with a higher frequency, because there are no major capacitive elements and overhead lines have often a higher resistance than cables. Thus, resonances are more damped and do not reach very high or low values. On the other hand, in cable-dominated grids, the first resonance frequency can be at around 100 Hz or even below, as the high capacitance values of the cables forms resonances with inductive elements, like transformers or other series reactances. If these less damped cable grids connect several converters, which are also not providing additional damping, then the amplitudes at resonance frequancy can reach very high or low values. As converters need to have several fast controllers to operate on a grid, they can go into resonance with the natural frequencies of the grid. In this case, harmonics will be generated by the converter, which are near or at the resonance frequency of the grid. The harmonic amplitudes will grow up often very fast within a few periods and may reach up to several 10%. As no harmonic protection is used within the grid, the grid cannot protect itself at this moment. If converters are not tripped due to a harmonic protection or some other internal abnormal voltage/current protection, this phenomenon could reach steady-state operation and stay for hours or days. This phenomenon could appear after a single switching operation, which shifts the resonance frequency into a critical range resulting in high harmonics generated by converters. However, harmonics will disappear without further measures if the resonance is shifted again due to a switching operation out of the critical range. This phenomenon contains a high risk for grid planning and operation as it is not well known. There is no evaluation method available, which is commonly accepted by universities and industry. In addition, grid control centres cannot calculate in advance, if a controller of a converter will get instable after a switching operation. Furthermore, there is no harmonic protection within the grid. So, this instability will not be solved automatically by protection systems. As these harmonics can reach very high values, there is a great danger that equipment will get damaged or even destroyed.

Sub-synchronous Oscillation

Beside the higher harmonic resonances produced by the interaction between individual devices and power system, also the sub-synchronous oscillations have been noticed, for example in the WPP

REPORT

of 158

applications. In [216]-[217], it has been reported that an HVDC system with DFIG-based wind parks can experience oscillations. The frequencies of the resonances reported are in the sub-synchronous range. The work in [216] postulates that there are two main causes for the sub-synchronous oscillations occurring when WTGs are connected to a system. Firstly, the induction generator effect, which occurs because the rotor is rotating faster than the synchronous speed of the sub-synchronous frequency, and secondly the sub-synchronous control interaction. As reported in [216], a typical frequency of the sub-synchronous current present in the DFIG is close to 20 Hz and can be influenced by the control strategy and the control parameters. The amplitude of the current is expected to grow as the output power from the wind park is increased, according to [216]. The authors examined a case of oscillatory phenomenon when WPP were connected via MMC-HVDC, to the system. The source of oscillations was a 20 Hz current injection present at the location of the wind park’s DFIG. At load AC side, the rectifier experienced currents as shown in the following frequency characteristic in Figure 2-11. Figure 2-11 shows that sub-synchronous oscillations at the DFIG side can turn into super-synchronous oscillations at the load AC side, where the system is connected. The publications [216]-[217] still mention the phenomenon as not fully explored, so a complete theoretical explanation to the mechanisms involved is not yet available. The sub-synchronous oscillation can occur also due to the sub-synchronous control interaction between the grid and WTG. The interaction can cause damage or incorrect operation of WTG. This phenomenon is described in [218], where also the preventive control techniques are proposed.

Figure 2-11: Frequency analysis of the current at the AC side [216]

2.8.2 Evaluation Indices The symptoms of harmonic instability are the sub- and super-synchronous oscillations in the grid. These are evaluated by different indices and compared to the limitations described in the previous Subsection 2.7.2. However, if a harmonic instability occurs, current limits are reached very fast and it is obvious that high distortion of over 10% is not acceptable.

0

10

20

30

40

50

60

70

0 20 30 50 70 80 100 130 150

Mag

nitude

(% o

f fu

ndam

enta

l)

Frequency (Hz)

REPORT

of 158

2.8.3 Consequences Harmonic instability consequences are similar to steady-state harmonics described in Subsection 2.7.3. As this phenomenon should be visible only for short-term, the long-term damage of equipment, for example due to overheating, is most likely not very crucial. However, high harmonics can cause over-voltage and over-current and over-voltage in single assets. For example, the currents and voltage within a grid filter are significantly influenced by harmonic distortion in the grid. In addition, other assets, which react very vulnerable to high frequencies, like surge arrestors, capacitors in relays or other auxiliary systems, could be damaged or destroyed.

2.8.4 Mitigation Methods By occurrence of a harmonic instability, converters are actively generating harmonics and feeding active power at frequencies above 50 Hz in the grid. Thus, standard passive filters, which try to absorb harmonic current, could not be a solution for this phenomenon. Grid filters have minimum impedance at the frequency that the current will reach values, which trigger the protection system or damage the filter circuit. To prevent the occurrence of a harmonic instability the following mitigation methods are known:

• Shift of grid resonance; • Tuning of controls.

As the phenomenon is triggered by a grid resonance and the converter control, it is obvious that the solution can be found by changing these two factors. A shift of grid resonance is difficult to achieve, as it is defined by the characteristic of the used assets. If this issue is not discovered during planning phase, the grid resonance could be influenced by installing a filter. This filter must not be designed for absorbing harmonics, but only for changing the frequency dependent impedance. However, this solution is costly and if the grid topology changes, the filter might be useless. Thus, it is more sufficient to tune the control of converters. By changing some parameters, it might be possible to solve issues for a defined frequency range, but this might shift them to other frequencies. In general, it should be helpful to slow down the fast controllers within a converter, but these might lead to an unfavourable transient behaviour.

2.8.5 Practical Studies During system planning, studies are performed to evaluate possible problems, which might occur in the power networks. As the penetration of PE devices in power networks has grown in the past few years, still no certain investigation methods are available, which both scientists and industry accept. From theoretical point of view, it is possible to do an eigenvalue analysis and identify critical resonances. However, there is no approach available in literature, which uses an eigenvalue analysis to solve high frequency interactions of converter controllers with the grid or among each other. In some papers [219]-[220], the Nyquist Criterion is used to evaluate the stability of PE devices when connected to a grid. Some companies use this method to identify harmonic instability phenomenon [211], [213]. Additionally, Electromagnetic Transient (EMT) simulations could be performed by using detailed vendor models. However, EMT simulations are time consuming and it has to be known where problems might be expected. Otherwise, it is very difficult to identify possible issues. It seems to be feasible to use the Nyquist Criterion evaluation to scan the system for possible problems and use afterwards EMT simulations to verify the simulation results and identify possible countermeasures.

REPORT

of 158

3 PQ in Transmission Systems

3.1 TenneT

3.1.1 Introduction TenneT is a leading European electricity transmission system operator, with activities in the Netherlands and Germany. TenneT strive to ensure a reliable and uninterrupted supply of electricity for some 41 million people in their high voltage grid. Thereby making every effort to meet the stakeholders’ needs by being responsible, engaged and connected. TenneT ranks among Europe's top five TSOs and works closely with governments, Non-Governmental Organisation (NGOs), suppliers and investors all over the world. The aim is to ensure essential high-voltage infrastructure development, realised and managed efficiently, now and in the future. This covers onshore and offshore grids, as well as cross-border interconnections. TenneT is keen to pursue the further development of the North West European (NWE) electricity market. In Germany, TenneT owns the 220 kV and 380 kV grids in the middle of Germany from north till south. In the Netherlands, TenneT owns also the 110 kV lines. In both countries, TenneT is the responsible TSO to connect offshore wind parks to the onshore grid. The key figures of TenneT in Netherlands and Germany are as follows in Table 3-1. Table 3-1: Key figures of TenneT in Netherlands and Germany in 2015

Netherlands Germany

Imports 30,759 GW/h 52,289 GW/h Exports 22,013 GW/h 54,255 GW/h Total grid length 10,248 km 12,127 km Number of transformer substations 325 129 Number of end-users 16.70 million 24.11 million

Besides the connection of offshore wind parks mainly by using HVDC systems, TenneT is also operating and building several HVDC interconnectors to neighbouring countries. At the moment the HVDC systems BorWin1, BorWin2, HelWin1, HelWin2, SylWin1, DolWin1 and DolWin2 (in commissioning) with a total transmission capacity of 5 GW are in operation to connect offshore wind parks. The Netherlands are connected to Norway, Britain and Denmark while Germany will connect to Norway by HVDC systems. The overview of TenneT grid can be seen in Figure 3-1.

REPORT

of 158

Figure 3-1: TenneT grid in Netherlands and Germany in 2015

3.1.2 Generation Notable energy generation in Germany is oriented in renewable energy sources, which are integrated into the system. The proportion of energy generation in Germany by sources is given in Table 3-2.

REPORT

of 158

Table 3-2: Energy generation in Germany year 2015

Energy source Generated energy

Natural gas 9.1% Nuclear power 14.1% Hard coal 18.1% Brown coal 23.8% Renewable energy 30.1% Others 4.8%

In 2016 around 195 GW installed generation capacity is available in Germany while the peak load is only about 85 GW. This is due to the high number of renewable energy sources, which have low number of full load hours within a year.

3.1.3 Consumption In Germany TenneT grid operates only at EHV level. Therefore, most of the customers are distribution system operators, which are connected to TenneT grid. Only a few production companies, which require a 220 kV or 380 kV connection, are connected directly to TenneT grid. The energy import and export levels in TenneT grid in Germany and in Netherlands can be seen in previous Table 3-1.

3.1.4 PQ Overview TenneT has the responsibility in Germany to connect offshore wind parks, which are nearest to the control area of TenneT. For that purpose, several HVDC and also some AC connections have been built. There have been observed different harmonic phenomena during the operation of HVDC systems. They are mainly divided into steady-state harmonics and harmonic instabilities. Steady-state harmonics result, for example, from switching frequencies of converters or assets with a non-linear current-voltage characteristic. In contrast, harmonic instabilities can occur if the controls of converters start to interact with grid resonances. Steady-state harmonics are well known in the literature and their sources are mostly known, as well as calculation methods. In the strong grid of continental Europe, extensive problems with harmonics are mostly unknown, as overhead lines are used in the EHV grid. However, the use of long cable systems to connect offshore wind parks causes resonances with much lower frequencies than known in traditional systems. In addition, new technology is chosen to reduce losses. This is cost-efficient, but on the other hand worsens harmonic problems, as resonances are poorly damped. As the connection or disconnection of cable systems influences the resonance frequency, it is possible to hit integer harmonic frequency. It results in high harmonic distortion and within the system connected by HVDC, it may lead to a security shut-down of the HVDC system. Figure 3-2 represents the harmonic distortion in offshore grid due to steady-state harmonics.

REPORT

of 158

Figure 3-2: Steady-state harmonic in offshore grid Stability problems in low frequency range are well known in transmission networks. They result primarily from slow controllers of the generators in combination with grid resonances. In converter-dominated grids, PE devices take care of frequency and voltage stability. To ensure this, for example to stay synchronized with the grid and control the current, they need fast controllers. As these controllers can have a bandwidth of several hundred Hz they may interact with grid resonances in this frequency range. It may result in a high harmonic distortion with a frequency near the grid resonance frequency. As this is mostly an interharmonic frequency, a first indication for this phenomenon can be obtained by determining the interharmonic frequency. In the recorded measurements below in Figure 3-3, a cable was connected at 0.1 s. Connecting a cable changes the system resonance frequency significantly. Instantaneous harmonics occur in the system, which may lead to HVDC system trip. In Figure 3-4, the time frame from 0.2-0.35-s is displayed. Clearly, it can be seen that there is a high harmonic content in the system. The identified frequency is about 451 Hz. As the peak of the harmonic is shifting, it is not the exact 9th harmonic. Furthermore, all available harmonic models of connected assets do not show such high harmonic emission at the 9th harmonic. This indicates that controllers were instable due to the changed resonance frequency.

-150

-100

-50

0

50

100

150

0 0.02 0.04 0.06 0.08 0.1

Vo

lta

ge

, k

V

Time, s

-1500

-1000

-500

0

500

1000

1500

0 0.02 0.04 0.06 0.08 0.1

Cu

rre

nt,

A

Time, s

REPORT

of 158

Figure 3-3: Voltage (155 kV, L1) during harmonic instability in offshore grid

Figure 3-4: Voltage (155 kV, L1) during harmonic instability in offshore grid in 0.2-0.35-s time frame

3.1.5 PQ Monitoring Systems Within the TenneT grid in Germany several PQ measurement systems are installed in different regions. They are connected to servers to continuously store data, like harmonic distortion. For the harmonic measurements, the RC-dividers are installed at critical substations, such as HVDC system connection points. In other substations, the transformer bushings are used. Within HVDC systems (onshore and offshore), every node is monitored. The used measurement units also provide a transient recorder, which is for example, triggered if the faults occur. On HVDC systems, the additional attention is paid by using a separate transient recorder, which monitors the complete HVDC system, like AC-side, DC-side, arm currents etc.

3.2 ELES

3.2.1 Introduction The public company ELES, Ltd., Electricity Transmission System Operator (ELES) has the exclusive right to perform the public service of the transmission network system operator in Slovenia. The founder and the sole owner of the company is the Republic of Slovenia.

-200

-100

0

100

200

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Vo

lta

ge

,k

V

Time, s

-200

-150

-100

-50

0

50

100

150

200

0.2 0.3

Vo

lta

ge

,k

V

Time, s

REPORT

of 158

Transmission network is designed to transfer electrical energy from utilities to distribution networks and direct consumers connected to transmission network, and to enable electrical energy exchange with neighbouring TSOs. Transmission network in Slovenia consists of three high voltage levels, namely 400 kV, 220 kV and 110 kV. At the end of year 2015, the length of 400 kV overhead lines reached to 669 km, while 220 kV lines totalled 328 km. Moreover, 110 kV voltage level lines were altogether far the longest, with the length of 2,842 km from which 1,862 km were owned by ELES. In Slovenian network there are four different types of substations with 400/110 kV, 400/220 kV, 220/110 kV and 110/35 kV transformations. Transformers with rated power of 300 MVA and 400/110 kV transformation are installed in substations Maribor, Krško, Okroglo and Divača. In addition, transformers with rated power of 400 MVA and 400/110 kV transformation are installed in substations Podlog and Beričevo while transformers with rated power of 150 MVA and 220/110 kV transformation are installed in substations Cirkovce, Podlog, Beričevo, Kleče and Divača. Alongside already mentioned transformers, there is also a power shifting transformer with rated power of 2×600 MVA installed on 400 kV voltage level in Divača substation. It was installed in year 2010 as a response to the uncontrolled flows over Slovenian network just a decade ago. Excessive power flows were very frequent, which resulted in higher losses and lower operational security. The situation has significantly improved with the installation of the named power shifting transformer in Divača substation, which is designed and operated to control power flows at Italian border. Another important aspect of international cooperation is also the leading role of ELES in Slovenia-Croatia-Bosnia and Herzegovina control block. ELES has to ensure that the systems of all three countries act as one towards the rest of Europe. The Slovenian transmission network is expressed in Figure 3-5.

Figure 3-5: Slovenian 400 kV and 220 kV transmission network

REPORT

of 158

3.2.2 Generation At the end of year 2015, the total installed capacity in Slovenia was 3,700 MW, of which 33% corresponds to hydro, 19% to nuclear and 41% to thermal power plants. The remaining 7% are installed at the distribution level, mainly PV and negligible amount of wind power. Except from one pumping hydro power plant of capacity 180 MW, all hydro units are run of river while thermal power plants use lignite or gas. Altogether, power plants produced 13,957 GWh of electrical energy in year 2015 (38% nuclear, 28% thermal, 29% hydro and 8% distributed energy sources).

3.2.3 Consumption In Slovenia, consumers are divided into three groups, Closed Distribution Systems (CDS), direct consumers and the last but not least, five distribution companies that are responsible for distribution of power from substations to consumers (households or companies). In 2015, they totally consumed 14,026 GWh of electrical energy and the largest consumers, were distribution systems that consumed approximately 80%. The peak load in 2015 was 2,086 MW. As a system operator, ELES is also responsible for cooperation with neighbouring TSOs, therefore it has to control cross-border power flows. Because Italy is a big net importer of electrical energy and since Slovenia is also a net importer of electrical energy, the energy usually flows across Slovenia from north (Austria) and south (Croatia) towards west. In 2015, Slovenia imported 9,045 GWh and exported 8,975 GWh. Both values are relatively big compared to both, production and consumption in Slovenia. Just a decade ago, uncontrolled flows over Slovenian network were often, which resulted in higher losses and lower operational security. The situation has significantly improved with the installation of a power shifting transformer in Divača substation. The transformer is designed to control power flows on Italian border.

3.2.4 PQ Overview The main PQ problem, faced by ELES specifically at some parts of the 110 kV level of the transmission network, is flicker phenomenon. Values of other parameters (harmonic distortion, harmonic voltage, RMS voltage, voltage unbalance and power frequency deviations) are within the limits defined by the EN 50160 standard. ELES is also obligated for regular reporting to the national regulatory agency on PQ parameters. However, reports provided to the national regulatory agency are of little or no value for scope that MIGRATE project deals with. Supply Voltage Analysis in ELES Network

Figure 3-6 and Figure 3-7 show the trend of the RMS voltage at 400 kV and 110 kV voltage level. For the 400 kV system, representative data sample from the substation’s busbars (4 out of 7 in total) was used. Values at 110 kV level were measured in 6 Slovenian network substations, which are the most critical considering voltage flicker. Values from the measured sites have been analysed together with regard to the flagged data, which were analysed further and invalid measurement results were discarded from the analysis. For the presentation of obtained measurements, the average 10-min RMS voltage (VRMS) and short time flicker (Pst) values are used, provided from permanent monitoring of PQ. As it may be depicted from trends for 400 kV level, the 1% percentile is increasing, while 99% percentile is decreasing at

REPORT

of 158

slower rate. In total, gradual increase of the voltage during the observation period may be noticed. In the years 2012 till 2014 a gradual increase of the voltage values is seen due to regional operation of the 400 kV network in the wider Balkan area (long and low loaded lines). The problem was noticed and tackled in 2015. This phenomenon does not influence the voltage values on the 110 kV level since the 110 kV voltage level has its voltage control through the control of the transformers in the substation. In 2016, the voltage variation is in the 400 kV network is in range ±2% Vn that means from 402.8 kV to 418.8 kV (measured line-to-line) and the median voltage value is 102.7% Vn.

Figure 3-6: Trend lines of VRMS at 400 kV voltage level

Figure 3-7: Trend lines of VRMS at 110 kV voltage level

REPORT

of 158

VRMS trends on the 110 kV level show gradual decrease of the voltage during the observed period. It must be however emphasised that the 8-year change of VRMS is in a range of 1 kV (less than 1% Vn), which is rather small decrease. Voltage variation in 2016 at the 110 kV level is ±1.6% Vn (from 105,0 kV to 108.1 kV measured line-to-line) and the median voltage value is 106.7% Vn. Flicker Analysis in ELES Network

The short-term flicker values, 99% value, at the 400 kV level have decreased significantly, from 0.7 to 0.5, in last 8-years, as can be seen in Figure 3-8. The main reasons for improvements, are the gradual strengthening of the transmission network with new power lines (e.g. new 2x400 kV line 2013) and new transformers (additional 400/110 kV transformer).

Figure 3-8: Trend lines of Pst at 400 kV voltage level On the other hand, the short-term flicker on the 110 kV level is relatively constant, on average around 0.4 and with 99% values up to 2.0. The trend lines of Pst at 110 kV voltage level can be seen in Figure 3-9. In addition, it can be seen that the 99% value is way above the average in comparison to the 1% values, which are rather close to the average. This is mainly due to the nature of the flicker causing events (relatively short lasting, but violent load changes due to operation of arc furnaces). The flicker is generated on the 110 kV level and then propagated through the 400/110 kV substations to the 400 kV level. The attenuation factor of the propagation can be approximately around 2.

REPORT

of 158

Figure 3-9: Trend lines of Pst at 110 kV voltage level The average short-term flicker values for the weekly profiles have also analysed in ELES network. The average weekly profile changes through period of 2009-2016 are expressed in Figure 3-10 and Figure 3-11, respectively for 400 kV and 110 kV voltage level. It can be observed that values throughout the week are increasing during the weekdays at 110 kV level, while at the 400 kV they are decreasing. The flicker level during weekend compared to the weekdays is relatively high on both, 400 kV and 110 kV voltage level. This is due to the operation of the arc furnaces, as well as the lack of rotating masses at loads or generators in the network. Statistical observation (disregarding the yearly differences and evaluating the average values of all years together) shows attenuation coefficient 1.4 for Monday, 2.7 for Thursday and 2.3 for Sunday. In the following Figures 3-10 through 3-14, the data and analysis for year 2016 contains only the first half of the year (6 months) measurements.

Figure 3-10: Weekly average profile of Pst at 400 kV voltage level

REPORT

of 158

Figure 3-11: Weekly average profile of Pst at 110 kV voltage level Analysis for the daily flicker profiles has been made using the same PQ data as for the trend analysis. For a given PQ measurement site, the data from e.g. 10:00 to 10:50 (6 values) was counted into the 10:00 interval and averaged. Data from all available measurement sites for a given voltage level was then again averaged to produce a single point on the graph. (Single point in the graph represents the average of the 6(values)x6(sites)x365(days)=13,140 measurement points at 110 kV level for each hour of the year. The following Figure 3-12 and Figure 3-13 represent the flicker phenomenon daily profiles in years 2009-2016 at 400 kV and 110 kV voltage level.

Figure 3-12: Daily average profile of Pst at 400 kV voltage level

REPORT

of 158

Figure 3-13: Daily average profile of Pst at 110 kV voltage level Short-term flicker values in the 110 kV network depend on the operation of the arc furnaces connected thereto. Operation of the arc furnaces is tightly linked to the tariff of the electric energy, which is usually high from 6:00 till 14:00, hence arc furnaces operate very rarely at this period. The normal operation time for arc furnaces an iron works is usually during weekends, with non-stop operation (except breaks for inserting and pounding of iron). The working period can vary, depending on the iron works site, the time of the year and the situation on market. It can be observed from Figure 3-12 and Figure 3-13 that the shape of daily flicker profile does not change significantly through the years in both 400 kV and 110 kV level. Highest hourly values appear after 14:00 and lasts approximately until 17:00. It can be seen that the Pst daily profile is flattering through the years at the 110 kV level, while this cannot be seen at 400 kV level. The ratio between Pst at the 400 kV and 110 kV level is approximately 2, with the emphasis that flicker daily profile at 400 kV level is not exhibiting very deep morning and early afternoon dip. Harmonic Distortion Analysis in ELES Network

Following covers the harmonic distortion in ELES grid. For the presentation of measuring results, the average 10-min values of THD and individual harmonic voltage Vh provided by permanent monitoring of PQ are used. The values of 1%, 50% and 99% of THD and Vh at 400 kV and 110 kV voltage levels are presented in Figure 3-14 and Figure 3-15. Values used in the following analysis were measured in 6 optional substations in the Slovenian 110 kV system. Nevertheless, the arc furnaces are located near the measurement sites. For the 400 kV system, representative data sample from the substations busbars (4 out of 7 in total) was used. When analysing voltage harmonics on HV level, one must take into consideration that there can be issues regarding the performance (amplitude and phase accuracy) of the HV voltage transformers, when measuring harmonic voltage. IEC TR 61869-103 [221] and reference [222], based on the measurements, are pointing out that there could be a non-negligible ratio error even below the 15th order harmonic voltage. Nevertheless, this issue is not taken into consideration for this analysis. Values from all measured sites were analysed together

REPORT

of 158

with regard to the flagged data, which were analysed further and invalid measurement results were discarded from the analysis.

Figure 3-14: Trend lines of THD at 400 kV voltage level The THD values for the 400 kV system show rather constant values through the last 8-years, averaging between 0.83-0.94%. Extreme values have a tendency of gravitating towards the average values through the observation period. These results reflect the previously described changes in the 400 kV network.

Figure 3-15: Trend lines of THD at 110 kV voltage level Surprisingly, the average THD values at the 110 kV level are smaller than the values at 400 kV. They range from 0.27% to 0.44%, through the last 8-years. The 99% values are considerably higher than the average values and range from 1.76% to 1.94%. Whereas 1% values are extremely low, by not

REPORT

of 158

exceeding the 0.1%. This means that harmonics are generally very low at 110 kV level, but can be high in case of present disturbing loads in the network. This performance resemble short-term flicker performance in the 110 kV network discussed before. Nevertheless, none of the harmonic values at 400 kV or 110 kV network exceed the planning levels defined by the IEC TR 61000-3-6. Analysis of the individual harmonic voltages (up to the order of 15th) on the 400 kV and 110 kV level in ELES network showed expected results. From the trends it is difficult to draw sound conclusions whether certain harmonic values are increasing or decreasing. Judging from the visual analysis of the individual harmonics trends, the observations are listed in below Tables 3-3 and Table 3-3. Even harmonics values are very low, as can be expected for the symmetrically operated HV network. The 1% values, as well as the average values are not very different from each other. The 99% values indicate that the main harmonic sources in the network are occasional events, such as energising network components. Odd harmonics (including triplens) are more consistent. Their values are as expected higher than those from the even harmonics. Table 3-3: Voltage harmonic trends for the Slovenian 400 kV network

Harmonic order

Trend of the average value Consistence of the minimum and

maximum values

3rd Constant Converging 5th Constant Converging 7th Rising Constant 9th Falling Constant/Converging 11th Falling Converging 13th Rising Diverging 15th Constant/Rising Diverging

Table 3-4: Voltage harmonic trends for the Slovenian 110 kV network

Harmonic order

Trend of the average value Consistence of the minimum and

maximum values

3rd Constant Converging 5th Constant Converging 7th Rising Diverging 9th Falling Converging (strongly) 11th Rising Diverging 13th Rising Diverging 15th Rising Diverging

From Table 3-3 and Table 3-4, it can be seen that the 3rd and 5th order harmonic voltages are rather constant in both networks and the 1% and 99% values (e.g. minimum and maximum values) are converging towards average values. The 7th harmonic voltage values are rising in both networks, either with constant minimum and maximum values and diverging values. The 9th harmonic voltage values are falling with the converging trend and the same can be seen at the 11th harmonic voltage level at 400 kV. On the other hand, 11th harmonic voltage in the 110 kV network, as well as the 13th and 15th harmonic voltages in both networks, are showing the trend of rising with the divergent feature of the minimum and maximum values. The reason for the described harmonic characteristics is yet to be analysed. One of the reasons may be that there is a shift in the system impedances. Another reason may be the change in the ration of the 6- and 12-pulse devices in the network. Most likely, it is a combination of both. Nevertheless, no values at 400 kV and 110 kV network exceeds the planning levels defined in the IEC TR 61000-3-6.

REPORT

of 158

3.2.5 PQ Monitoring Systems In 2008 ELES started the project of permanent monitoring of PQ at 110 kV, 220 kV, 400 kV voltage levels. Currently, the permanent monitoring of PQ is established for 186 measuring points at the most important transmission system busbars. For this purpose approximately 100 PQ analysers are used. All PQ analysers are remotely connected with the measuring centre, which is located at ELES headquarters in Ljubljana. Data is stored in SQL database and analysed by program package. The PQ monitoring system in ELES is represented in Figure 3-16.

Figure 3-16: ELES PQ monitoring system

3.3 Elering

3.3.1 Introduction Estonian Transmission System Operator Elering manages the Estonian transmission system in real time. Elering is responsible for planning the functioning of the system and managing it to ensure safe and reliable operation of the network to all participants. As well as enabling the electricity transmission, Elering is also responsible for managing the balance of the system at all times. The Estonian transmission system connects power plants (synchronous and converter connected), distribution networks and large power consumers. It is part of one synchronous area together with

REPORT

of 158

Latvia, Lithuania, Belarus, and Russia. Estonian power system is connected with Russian power system via three 330 kV lines, two from Narva to St. Petersburg and Kingissepa, and one from Tartu to Pskov (with a total transmission capacity of approximately 800-1,000 MW). Estonia is connected to the Latvian system with two 330 kV lines, one between Tartu and Valmiera, and one between Tsirguliina and Valmiera (with a total transmission capacity of approximately 800-1,000 MW). Estonia is also planning a third connection to the Latvian system with a capacity of 1,400 MW, which should be operating in 2020. Additionally, Estonia has two HVDC connections with Finland. EstLink 1 (VSC technology, 350 MW) has been in operation since the end of year 2006 and bears great symbolic significance for the connection of the power systems of the Baltic and Nordic. Estlink 2 (LCC technology, 650 MW) became operational in year 2014. The transmission capacity between Estonia and Finland is 1,000 MW. Estonian transmission network consists of:

• 1,702 km of 330 kV lines; • 158 km of 220 kV lines; • 3,479 km of 110 kV lines; • 61 km of 35 kV lines; • 309 km of HVDC lines; • 146 substations.

Estonian transmission network is shown below in Figure 3-17.

Figure 3-17: Estonian transmission network

REPORT

of 158

3.3.2 Generation Generation capacity in Estonia is about 2,677 MW (November 2015). Most of the generation is located in the northeast part of Estonia, where most of the energy is produced from Estonian main natural resource oil shale. Largest oil shale power plants are Eesti (1,355 MW), Balti (322 MW) and Auvere power station. Auvere power station is the newest and is in operation since January 2016 with output power capacity of 270 MW. Recently, there has been a significant rise in the use of wind energy. In 2016 there is 375 MW of wind power capacity installed and more projects are planned (over 4000 MW including offshore WPP) for the future. Most of the wind parks are located in the northern and western part of Estonia. Another development in the recent years is the use of different biofuels, wood, peat etc., for power and heat co-production. In Estonia, the hydro resources are limited. The largest hydro power plant is Linnamäe HPP with 1.2 MW of output power. Other renewables, like geothermal or solar energy, are not widely used. However, in the last two years, the solar energy is becoming more and more popular on a smaller scale (power production of 11-300 kW). First applications to connect solar to power system have been delivered with output power of 20-60 MW. The Estonian power plants with their location can be seen in Figure 3-18.

Figure 3-18: Power plants in Estonia

REPORT

of 158

3.3.3 Consumption In 2015, the total consumption in Estonia was 7.9 TWh and the maximum peak was 1,553 MW. Main Estonian consumption is aggregated near the large cities and technological parks. Most of the consumption is located on northern part of Estonia (near Tallinn and Ida-Virumaa), but also in Pärnu and Tartu. Figure 3-19 illustrates the consumption in Estonia, where the maximum loads of Estonian geographical parts can be seen. It must be said that those loads are maximum peaks of each region and they do not occur at the same time. Comparison between the consumption and generation of the last 5 years is shown in Table 3-5 and Figure 3-20. Estonia has mainly been an exporting country regarding electrical energy.

Figure 3-19: Energy consumption in Estonia Table 3-5: Energy consumption and generation in Estonia

Year Consumption Generation

Total, TWh Total, TWh Renewable, TWh

2010 8.2 10.44 0.94 2011 7.9 11.39 1.16 2012 8.1 10.46 1.37 2013 7.9 11.66 1.15 2014 7.8 10.91 1.36 2015 7.9 9.01 1.51

Figure 3-20: Comparison between energy consumption and generation in Estonia

5

6

7

8

9

10

11

12

13

2010 2011 2012 2013 2014 2015

En

erg

y,

TW

h

YearConsumption

REPORT

of 158

3.3.4 PQ Overview Estonian power system is continuously evolving, steadily incorporating features of a modern system, such as HVDC links, converter-connected generation (mainly WPP), etc. The relatively small size of Estonian power system can allow even the slight changes to have an extensive impact on PQ. Thus, special care must be applied in TSO activities regarding grid connection of new generation and consumption. Currently Elering observes PQ characteristics levels in Estonian network. The following are under observation:

• Over- and under-voltages; • Voltage unbalance; • Flicker; • Harmonics (resonances, stability).

Over- and Under-voltages

First factor, which contributes to over-voltages, is the increase of generation in weak areas of the transmission network. The main areas of WPP development (mostly Western Estonia) have been historically less evolved economically and population-wise. Thus, the available transmission network, its local short-circuit power and voltage regulation capabilities are quite modest and are gradually drained up by the areas’ development. Another factor that can cause voltage problems in the future is possible lack of synchronous generation in the power system. Currently the backbone of Estonian generation is based on old designs using oil shale as a primary energy source. All of this capacity is situated in North-Eastern Estonia, which is most developed transmission grid-wise. Dismantling this capacity because of intolerable emission levels can have a very complex negative impact on the power system – lowering rotating mass, reactive and active power regulation capability, short-circuit power, which altogether resulting in worsening PQ levels, especially voltage.

Figure 3-21: Seasonal variations of 99% voltage measurement values at 110 kV level

REPORT

of 158

Figure 3-21 shows aggregated 99% values for voltage measurements in three different measurement points. It can be seen that less developed areas in A and C steadily show higher voltage values, while B measuring point exhibits less voltage variations and a larger margin regarding allowed voltage level. Estonian TSO must therefore foresee adequate transmission network reinforcements and measures to maintain enough short-circuit power as well as sufficient regulation levels for active and reactive power. This can also include new requirements for non-synchronous generation in terms of voltage regulation and short-circuit current generation capability.

Voltage Unbalance

There is currently no problem with voltage unbalance levels in Estonian transmission network. The matter can arise due to electrified railway development, as the plans foresee new AC-fed railway track (“Rail Baltic”) to pass areas with relatively weak transmission network. The proposed amount of asymmetrical railway load can cause unallowable voltage unbalances in these areas. Elering has already foreseen activities needed to strengthen the transmission network in case of such development in the future. Figure 3-22 presents seasonal variations of measured 99% asymmetry factor ka values. The allowed limit for asymmetry is 2%. Observed values are well under the limit.

Figure 3-22: Seasonal variations of 99% asymmetry ka values at 110 kV level

REPORT

of 158

Flicker

Output of WPPs follows the local wind speed and depending on network strength can influence local voltages. There are examples where regular voltage fluctuations (e.g. following semi-regular wind squalls) cause noticeable voltage flicker. This has been the case in some regions in Estonia, where increased level of WPPs’ production is combined with relatively weak grid connections. Figure 3-23 shows seasonal variations of 95% long-term flicker (Plt) values. The network planning limits for Plt is 0.85 (for individual WPP Plt=0.35).

Figure 3-23: Seasonal variation of 95% Plt values at 110 kV level Harmonics

Voltage and current harmonics can become a very prominent issue, if not properly addressed. Harmonic current generating devices are rapidly introduced on a large scale in both generation and consumption. In the last 10 years there has been a significant increase on the number of PE devices (e.g. WPP, HVDC links, etc.). In general, the levels of harmonics have increased over the years and are moving close to the planning limits. There are even cases where these limits are exceeded. Figure 3-24 illustrates recorded seasonal variations of 99% THD values (allowed level THD=3%). Parts of the network that are weaker, exhibit higher levels of voltage distortion. Recorded 95% levels for individual harmonics are shown in Figure 3-25. Most notable harmonics are 3rd, 5th, 7th, 11th and 13th.

REPORT

of 158

Figure 3-24: Seasonal variation of 99% THD values at 110 kV level

Figure 3-25: 95% values of 2nd-13th voltage harmonics (as a fraction of normal voltage;

blue line expresses allowed levels) during years 2010-2013 at 110 kV level

REPORT

of 158

3.3.5 PQ Monitoring Systems Elering has developed a central PQ monitoring system (PQMS) that comprises of around 130 stationery PQ analysers (IEC 61000-4-30 class A compliant), central data servers and working stations with proper visualisation and reporting software (used for data queries, visualisation and reporting). Additional data is also available from around 300 fault recorders, as well as more than a thousand feeder terminal disturbance recorders placed all over the Estonian transmission network. Main aims of the Elering’s PQMS are:

• PQ monitoring and troubleshooting; • Verification of customers’ contractual compliance; • Transmission system performance analysis.

In order to fulfil these goals, measurement of both, voltages and currents, is performed. The main recorded indices are the values of voltages, currents, active and reactive power, voltage and current harmonics (including THD and TDD), system frequency, voltage unbalance, voltage dips and temporary power frequency over-voltage (with disturbance recording). Complex recording of virtually all parameters related to PQ enables assessing PQ related issues much more quickly and efficiently. Measurements are carried out on 35 kV, 110 kV and 330 kV voltage levels in Elering’s transmission network. All used PQ analysers are stationary; there are currently no portable ones. The measuring points for PQ analysers are chosen mainly on points of common coupling of customers (usually the voltages are taken from substation busbar or, if available, customer bay VTs and currents are taken from customer bay CTs) or on other important sites (e.g. HVDC station bay). Capacitive VTs with corrective sensors, inductive VTs and RC-dividers are used. For current measurements regular inductive type CTs are used.

3.4 EirGrid Group

3.4.1 Introduction EirGrid Group is a state-owned company that manages and operates the transmission network across the island of Ireland. EirGrid’s task is to deliver a safe, secure and reliable supply of electricity now, and in the future. The EirGrid Group achieves this goal using a family of organisations: EirGrid, SONI and SEMO.

• Since 2006, EirGrid has operated and developed the national high voltage electricity grid in Ireland. EirGrid, it its role of TSO, operate the flow of power on the grid and plan for its future, while ESB networks, a separate entity in the role of Transmission Asset Owner, is responsible for carrying out maintenance, repairs and construction on the grid in Ireland.

• SONI is the electricity System Operator for Northern Ireland. Since 2014, SONI is responsible for planning for the future of the grid, while Northern Ireland Electricity is the Transmission Asset Owner, responsible for maintenance, repairs and construction of the grid in Northern Ireland.

• SEMO is the Single Electricity Market Operator for the island of Ireland. This organisation runs the wholesale market for electricity.

REPORT

of 158

• EirGrid Interconnector Limited (EIL) owns the East West Interconnector (EWIC). EWIC is a 500 MW HVDC link between the electricity transmission networks of Ireland and Great Britain. EIL sell capacity on the EWIC through auctions.

• EirGrid Telecoms Limited (ETL) provides opportunities for telecommunications trading and associated activities across fibre networks integrated into the EWIC.

The transmission system in the island or Ireland is operated at 275 kV and 110 kV in Northern Ireland and at 400 kV, 220 kV and 110 kV in Ireland. The two transmission systems are currently connected by one 275 kV double circuit and two 110 kV single circuits. A new 400 kV cross-border circuit is planned. There are two HVDC interconnections with Scotland (Moyle 500 MW LCC) and Wales (EWIC 500 MW VSC). A geographical overview of the transmission network is shown in Figure 3-26.

Figure 3-26: Transmission system in the isle of Ireland

REPORT

of 158

The 400 kV network in Ireland provides a high capacity link between a large coal fired generation plant in the west coast with Dublin on the east, the largest demand centre. The 275 kV network in Northern Ireland is comprised of a double circuit ring with two spurs to a large combined-cycle gas turbine generation plant (CCGT) and to the cross-border link with Ireland. The 110 kV circuits provide parallel paths to the 220 kV, 275 kV and 400 kV networks and is the most extensive element of the all-island transmission system. The transmission system is mostly comprised of overhead lines, with the exceptions of the main city centres of Belfast, Dublin and Cork where underground cables are used. Table 3-6 below shows the total lengths of overhead lines and cables used in the all-island transmission system at different voltage levels. Table 3-6: Total length of transmission circuits

Voltage level Total line lengths, km Total cable lengths, km

400 kV 439 0 275 kV 825 <1 220 kV 1786 129 110 kV 5877 372

The total transformer capacity installed in the transmission system is summarised in Table 3-7. In addition, there are three phase shift transformers: Carrickmines (Dublin) 220 kV 350 MVA, Enniskillen (cross-border) 110 kV 125 MVA and Strabane (cross-border) 110 kV 125 MVA. Table 3-7: Total transformer MVA capacity

Voltage level Capacity, MVA Number of transformers

400/220 kV 3 050 6 275/220 kV 1 200 3 275/110 kV 3 840 16 220/110 kV 11 679 58

The reactive power compensation devices are summarised in Table 3-8.

Table 3-8: Reactive power compensation devices

Voltage level Type Capacity,

Mvar Number of

devices

400 kV Line shunt reactor 160 2 Voltage source converter interconnector +/- 175 1

275 kV Shunt capacitor 236 4 220 kV Shunt reactor 100 1

110 kV Static var compensator 90 2 Shunt capacitor 921 42

3.4.2 Generation In Ireland, renewable energy policy is driven by a binding European legal requirement to meet 16% of the country’s total energy consumption with renewable sources by 2020. The Irish Government aims to meet this target with 40% renewable electricity, 12% renewable heat and 10% renewable transport. It is estimated that 3,200-3,800 MW of installed wind generation will be needed to meet

REPORT

of 158

37% of demand, while 3% is expected to be sourced from hydro, solar, bioenergy and renewable Combined Heat and Power (CHP) generation. The Strategic Energy Framework for Northern Ireland also sets a target of 40% of renewable electricity consumption by 2020. This target translates into approximately 1,600 MW of renewable generation capacity in Northern Ireland. The main contributor to this figure is wind generation, with approximately 1,200 MW. Currently there is 9,088 MW of generation capacity installed in Ireland, with 7,741 MW connected directly to the transmission system and the remaining 1,347 MW connected to the distribution system. Figure 3-27 shows the connected and expected growth of wind capacity in Ireland up to 2024. Note the high levels of wind generation connecting at the distribution system. Currently there is only one offshore wind park in Ireland, with 25 MW capacity, connected at the distribution system at 38 kV.

Figure 3-27: Expected growth in wind capacity in Ireland There is 3,201 MW of generation capacity connected in Northern Ireland. 2,365 MW is connected to the transmission system and 836 MW is connected to the distribution system. The generation connected to the transmission system is mostly conventional synchronous plant and only one wind park is connected at transmission level (73.6 MW). Currently there is no offshore generation connected in Northern Ireland.

3.4.3 Consumption The all-island demand for 2015 and 2016 is shown in Table 3-9. While most of the demand is currently connected at distribution level (bar a reduced number of large industrial customers connected at 110 kV), Ireland is experiencing an unprecedented level of interest for connection of large data centres to the transmission network. At present there is approximately 250 MW of

REPORT

of 158

connected data centres. Beyond this, there is approximately 550 MW data centres with either contracted capacity or in the connection offer process. In addition to this, threre is approximately 1,000 MW of enquiries for further connections. To put this trend in context, if all the enquiries were to connect, the data centre load could account for 20% of the all-island system peak demand. Table 3-9: All-island demand

Year

All-island winter peak, MW

All-island summer peak, MW

All-island summer valley, MW

Autumn peak, MW

Ireland NI Ireland NI Ireland NI NI

2015 6 505 5 240 2 212

1 472 4 831 1 694 3 865 1 290 1 691 477

2016 6 532 5 265 2 221

1 473 4 856 1 696 3 885 1 291 1 700 477

3.4.4 PQ Overview The main area of concern is harmonic voltage distortion. In recent years, EirGrid have observed increasing levels of voltage distortion at low harmonic orders. While still well within the established planning levels (as per IEC TR 61000-3-6), on-going monitoring and data processing is in place to understand and anticipate any possible risks to the security and quality of supply. This harmonics trend has been caused by the proliferation of wind turbines using electronic converters and the increased number and length of radial underground cables associated with these wind parks. Most of these connections are located in remote and weak parts of the transmission network, where natural frequencies of resonance and system damping are low. An example of this trend is presented in Figure 3-28, which shows the percentage variations of individual harmonic orders and THD measured during the same two-month period over two consecutive years (2014 and 2015). The main observation from this graph is a sharp increase in the 5th harmonic order, i.e. 23% increase in just one year. A reduction in triplen and high harmonics (above 13th) has also been observed. The trend suggests that the sources and sinks of harmonic currents are shifting down in the harmonic spectrum. Also, the percentage rise in the characteristic harmonics (5th, 7th, 11th and 13th) of power electronic converters is evident.

Figure 3-28: Observed trend in harmonic voltage distortion in the Irish transmission system (2014-2015)

REPORT

of 158

The direct impact of an underground cable connection can be observed in Figure 3-29 (10-min average) and Figure 3-30 (95 percentile). In this case, a 110 kV underground cable shifted the natural frequency of resonance towards the 7th harmonic order and resulted in a large amplification of background distortion even before the wind park started exporting power (i.e. without the active harmonic current injections from the converters in the wind turbines).

Figure 3-29: Harmonic voltage distortion measurements (10-min average values) before and after underground cable connection

Figure 3-30: Harmonic voltage distortion measurements (95 percentile values) before and after underground cable connection

REPORT

of 158

3.4.5 PQ Monitoring System Over the past number of years, EirGrid has developed an extensive network of PQ recorders supported by a TCP/IP wide area network. There are currently over 70 devices installed in more than 50 different 110 kV, 220 kV, 275 kV and 400 kV transmission stations in Ireland and Northern Ireland. A small number of portable devices are also available for quick deployment when needed. The geographical location of PMUs and Disturbance Recorders is shown in Figure 3-31 below. Both types of devices are used for PQ monitoring.

Figure 3-31: PQ monitoring devices in Ireland Most monitoring devices are permanently installed at the interface point between the TSO and the customer on revenue class instrumentation cores. Inductive CTs and inductive VTs are predominantly used. The PQ devices can record both voltage and current harmonics. Such measurements can be obtained on a per phase basis or as an average of the three phases. In EirGrid, the recorders are configured by default to automatically report the three phase average voltage harmonics at the start of each

REPORT

of 158

month. These monthly reports contain the harmonic data in both physical units (kV) and in percentage. The reporting period has been reduced to weekly or daily in specific locations that need closer monitoring. Furthermore, the PQ devices are configured to issue an alarm if the instantaneous level of THD exceeds 2.0%. The assessment criterion against planning levels is based on 95% measurements on 10-min averages.

3.5 Questionnaire

3.5.1 Introduction The European Energy Supply for Electricity is undergoing fundamental changes. This includes strong moves away from heavy reliance on fossil fuels as the primary energy source mainly provided by large synchronous generators connected to the transmission systems, towards a decarbonised future supply relying increasingly on variable Renewable Energy Sources (RES) using non-synchronous generation predominantly connected to the network via PE. Some countries in Europe have already experienced times in which in some periods the national demand for electricity has been exceeded by the RES production alone. As this phenomenon continue to extend, the development of Europe wide markets and system operation will facilitate greater sharing of resources. European Network of Transmission System Operators for Electricity (ENTSO-E) expects that renewable energy sources will have a dominant role in generation capacity mix over the next 10-20 years. Installed wind and solar capacities are forecast to increase by 80% and 60%, respectively over that time-period. These fundamental changes in power system as a whole also affect the PQ in the transmission networks. In order to understand and deal with future PQ related challenges this topic is an integral part of MIGRATE project. One of the first tasks of this MIGRATE work package WP5 team was to identify present transmission network practice with regards to PQ issues in order to establish the state of the art. For that purpose, a questionnaire on range of issues of PQ in current transmission networks was developed and distributed to a significant number of transmission system operators in Europe. This section summarizes some of the key findings of that questionnaire, based on 23 responses received including responses from all MIGRATE TSO partners and responses from each synchronous zone in Europe. The aim of this section is to identify the range of existing issues related to PQ in transmission network and not to recommend any specific approaches or principles. The term “power quality” in this questionnaire covers the following phenomena:

• Fluctuations in the RMS voltage level (flicker, dips, surges); • Imbalance between the phases (voltage or current); • Distortions in the voltage or current waveform (harmonics, transients); • Frequency variations.

3.5.2 The Survey Organization of the Questionnaire

The questionnaire contained only 12 multiple choice questions, which were sent to all participants in order to facilitate high response rate and provide basic, but comprehensive overview of the PQ issues among European TSOs. This section analyses responses to the questions from questionnaire listed in Table 3-10.

REPORT

of 158

Table 3-10: Survey questions

The questions asked can be roughly grouped in 5 categories related to different aspects of PQ issues in transmission networks, namely, (1) whether PQ is monitored or not, (2) what type, how many monitors and how PQ is monitored, (3) what quantities are measured, (4) what are the main issues of PQ, and (5) how these issues are mitigated. Survey Participants

Potential survey participants were identified by the MIGRATE WP5 members and contacted by e-mail. Objective was to receive feedback from all MIGRATE partners and as much as possible from other European TSOs. The questionnaire was sent to the contact by e-mail. In total, 23 responses were received, including all MIGRATE partners and at least one response from each of the European synchronous areas.

3.5.3 The Results of the Survey Q0: Power System Characteristics

European synchronous areas are presently characterised by mostly running synchronous units and using overhead lines to transmit power. Corresponding results from the survey participants are shown in Figure 3-32. It can be concluded that most of the TSOs have more than 60% of synchronous based generation connected to their network and the level of wind and solar generation is generally less than 50%. For example, from the Figure 3-32 it can be seen that 13 TSOs in Europe have responded that the level of wind and solar in their network is in between 1-19%. Based on the answers from the survey more than 85% of circuits in Europe are overhead lines and use of cables is not common at transmission level. From the Figure 3-32, it can be seen that 14 TSOs in Europe have responded that the level of circuits with overhead lines is between 80-99%.

No Question CategoryQ1 Are you monitoring power quality in your networkQ3 Why do you do the monitoringQ2 How many power quality monitors do you haveQ6 What type of transducers are used for power quality monitoringQ9 Do you use or plan to use PMUs for power quality monitoring

Q10 How many PMUs in total do you have installed in your network at different voltage levelQ4 Which aspects of power quality to you monitorQ5 What electrical parameters do you monitorQ7 What are the main issues of power quality in your networkQ8 What are the main sources that cause power quality issues in your network

Q11 Do you use power quality mitigation in your networkQ12 How many of each of the following power quality mitigation devices do you have in your network

1

2

3

4

5

REPORT

of 158

Figure 3-32: Response to Q0 – What are your system characteristics with respect to generation mix and type of circuits?

Q1: Are You Monitoring Power Quality in Your Network?

The summary of the answers to this question is shown in Figure 3-33 for all respondents. It can be

seen that TSOs in different European synchronous areas seem to prefer to monitor PQ in their

network by using mobile units (43%) and somewhat less using fixed units installed at <10% of sites

(39%). It can be seen that most of TSOs perform regular PQ measurements in their network. Based

on the answers it can be concluded that various approaches are in use in Europe. For example, some

TSOs install monitoring devices at all sites with new connections to transmission level voltages and

some at TSO/DSO interface points. Some TSOs measure PQ only before and after WPP connections

and in case of railway connections. Those TSO that do not measure PQ at present stated that the

reason is the lack of regulatory requirements.

Figure 3-33: Response to Q1 – Are you monitoring power quality in your network?

Q3: Why Do You Do the Monitoring?

The summary of the answers to this question is shown in Figure 3-34 for all respondents. The analysis of the results showed that the most common reason for PQ monitoring is maintaining compliance against standards followed (stated by 65% of respondents in all responses) by desire to respond to requests and complaints from connected customers (57% and 56%, respectively). Other reasons include expansion project preparations and auxiliary supply related PQ measurements.

REPORT

of 158

Figure 3-34: Responses to Q3 – Why do you to the monitoring?

Q2: How Many PQ Monitors Do You Have?

The summary of the answers to this question is shown in Figure 3-35 for all respondents, indicating whether portable of fixed monitors are used. It can be seen that TSOs in general predominantly use fixed monitors, with portable monitors being used but only for solving special issues (most of the TSO own only few of these types of monitors). The median value of number of monitors used is 3 for portable monitors and 33 for fixed monitors. Close to 50% of the TSOs have more than 20 fixed monitors in operation and close to 25% having more than 100 fixed monitors. It was reported that also option to outsource PQ measurements is used.

Figure 3-35: Responses to Q2 – How many PQ monitors do you have?

Q6: What Types of Transducers Are Used for PQ Monitoring?

The summary of the answers to this question is shown in Figure 3-36 for all respondents. The analysis of the results showed that the most common approach to measure PQ in transmission networks is to use inductive type voltage transformers. This is indicated almost by half of the respondents. This is followed by capacitive type voltage transformers. Other solutions based on the answers are not common and used only by few TSOs. Special concern should be taken when measuring harmonics as the accuracy when measuring higher harmonics with regular measurement transducers in high

REPORT

of 158

voltage networks is limited. This has been acknowledged by some TSOs in Europe and it has been stated in the answers by one TSO that all new installations will be equipped with resistive capacitive type voltage transformers and by multiple TSOs that capacitive type voltage transducers are equipped with special parts to enable more accurate higher harmonics measurements. Other possibilities include capacitive type voltage transformers equipped with special technology for enabling more accurate harmonic measurements.

Figure 3-36: Responses to Q6 – What types of transducers are used for PQ monitoring?

Q9: Do You Use or Plan to Use PMUs for PQ Monitoring?

The summary of the answers to this question is shown in Figure 3-37 for all respondents. The analysis of the results showed that more than half of the TSOs use or plant to use PMUs for PQ monitoring.

Figure 3-37: Responses to Q9 – Do you use of plan to use PMUs for PQ monitoring? This question had follow-up questions to those who replied yes. The follow up question was: What do you monitor or plan to monitor, and which type/manufacturer of PMUs are in use? The summary of the answers to this question is shown in Figure 3-38 for all. The analysis of the results showed that more than half of responses indicate that high and low voltages and frequency excursions are monitored. These two are followed by unbalance monitoring. Other PQ phenomena are all considered to be equally important. These results can also be considered as being well in line with the possibilities of PMUs. The other issue singled out by these responses are frequency excursions.

REPORT

of 158

Figure 3-38: Responses to Q9 follow-up question – What do you monitor or plan to

monitor? The answers regarding which type/manufacturer PMUs are is use indicated that most of the TSOs use PMUs from one manufacturer and all together PMUs from 6 different manufacturers are in use in Europe. Some TSOs indicated that they use PMUs from different manufacturers.

Q10: How Many PMUs in Total Do You Have Installed in Your Network at Different Voltage

Levels?

The summary of the answers to this question is shown in Figure 3-39 for all respondents. The question included reference to voltage levels 110 kV, 220 kV, and 400 kV. In order to combine the results from different network they were divided into sub-transmission and transmission networks. Later includes all voltage levels from 220 kV, i.e. 220 kV, 275 kV, 330 kV, 380 kV, and 400 kV. The analysis of the results showed that most of the TSOs use PMU devices at higher voltage levels. Use of PMUs in sub-transmission level can be considered as limited. This is well in accordance with the understanding on PMUs and their applicability in transmission network, i.e. the purpose of the PMUs is to monitor transmission corridors and system stability related indicators. The median value of number of PMUs used in transmission network is 10 and average value is 13. Close to 40% of the TSOs have more than 20 PMUs in operation.

Figure 3-39: Responses to Q10 – How many PMUs in total do you have installed in your network at different voltage levels?

REPORT

of 158

Q4: Which Aspects of Power Quality Do You Monitor?

The summary of the answers to this question is shown in Figure 3-40 for all respondents. The analysis of the results showed that there is no specific PQ phenomenon that prevails. The interest in different PQ phenomena is spread across the board with slightly more interest in harmonics, flicker, high and low voltages, and dips and swells (87%, 83%, 74% and 74% respectively), closely followed by unbalance, frequency excursions, interruptions and voltage dips (52-65%). Other factors in Figure 3-40 include slow voltage variations, mains signalling voltage. This indicates that the TSOs in Europe are interested in complete PQ picture rather than in an individual PQ phenomenon.

Figure 3-40: Responses to Q4 – Which aspects of PQ do you monitor?

Q5: What Electrical Parameters Do You Monitor?

The summary of the answers to this question is shown in Figure 3-41 for all respondents. The analysis of the results indicates that in about 80% of the cases the TSOs tend to monitor all three phases. It is also more common to measure system current than to measure specific customer current. Moreover, the answers indicate that it is more common to measure line-to-neutral voltage than line-to-line voltage. However, this is in contradiction to the recommendations given in EN 50160, CIGRE TB 412, and as well as the guidelines by the European energy regulators, which all agree that the data analysis for PQ purposes should for transmission networks be based on the phase-to-phase voltage. This does not rule out line-to-neutral voltage measurements, but the survey results do indicate discrepancy between recommendations by expert groups and common TSO practice.

Figure 3-41: Responses to Q5 – What electrical parameters do you monitor?

REPORT

of 158

Q7: What Are the Main Issues of Power Quality in Your Network?

The analysis of the results indicates that the mostly highlighted PQ issues in the European transmission networks are high and low voltages and harmonics (both 35%) and sometimes also unbalance (52%). The summary of the answers to this question is shown in Figure 3-42 for all respondents.

Figure 3-42: Responses to Q7 – What are the main issues of PQ in your network? Based on the answers the harmonics above 50th level (4%) are not of concern in current networks. Other answers include harmonic instability which is especially raised by one TSO.

Q8: What Are the Main Sources that Cause Power Quality Issues in Your Network?

Based on the answers it can be concluded that PQ issues in the transmission networks are mostly caused by customers, with industrial customers (43%) having higher percentage than distribution networks (26%). Other devices, including wind and solar power plant and HVDC links, by large do not significantly contribute to PQ issues. In some salient cases though, these contributions are significant and rise above 20% for Type 4 WPP and LCC based HVDC. Other answers include railway loads that mostly contribute to voltage asymmetry. The summary of the answers to this question is shown in Figure 3-43 for all respondents.

Figure 3-43: Responses to Q8 – What are the main sources that cause PQ issues in your network?

REPORT

of 158

Q11: Do You Use Power Quality Mitigation in Your Network?

After identifying PQ issues it is important to find most cost effective solution based on technical and economic criteria to mitigate them. In Europe, 65% of TSOs use power quality mitigation in their network. This number is slightly reduced if only continental Europe TSOs are considered. The summary of the answers to this question is shown in Figure 3-44 for all respondents. This question had follow-up question to those who replied yes to the main question. It was asked to specify which devices or solutions are in use. The summary of the answers to this question is shown in Figure 3-45 for all respondents. The analysis of the results showed that the main solution for mitigating PQ issues is to use passive filters. This was indicated by more than 45% of all those who answered yes to first question. Other devices, as SVC, STATCOM and active filters are used but to a lesser extent. Other devices mentioned were shunt capacitors, shunt reactors and synchronous condensers. Other network related approach mentioned was assignment of connection phases for unbalanced loads based on overall minimization of unbalance.

Figure 3-44: Responses to Q11 – Do you use power quality mitigation in your network?

Figure 3-45: Responses to Q11 – Do you use power quality mitigation in your network? What devices are used for power quality mitigation?

REPORT

of 158

Q12: How Many of Each of the Following Power Quality Mitigation Devices Do You Have in

Your Network?

This question included similar options for devices as question eleven. The summary of the answers to this question is shown in Figure 3-46 for all respondents. The mostly used PQ mitigation devices are passive filters. These devices are mostly installed into the network in conjunction with HVDC links. Based on the survey answers there are more than 40 passive filters available, some SVCs and few STATCOMs. Voltage levels where these units are installed vary from 66 kV to 400 kV. The survey answers indicate that most of the passive filters (82%) are installed at the voltage levels higher than 150 kV, STATCOMs (67%) to networks with voltages up to 150 kV and SVCs (60%) mostly to networks with voltages more than 150 kV. Other devices here include synchronous compensators, shunt reactors and shunt capacitors. Based on the answers the size of the devices is predominantly in a range of 40-99 Mvars (35%) and followed by units with sizes of 100-149 Mvars (25%) and 200+ Mvars (23%).

Figure 3-46: Comparison of responses to Q12 – How many of each of the following PQ mitigation devices do you have in your network? To which voltage level these devices are installed and what is their typical size?

3.5.4 Conclusions This section of the report summarises main findings of a survey on range of PQ related issues in transmission networks. The main conclusions arising from the analysis of the results of survey of 23 TSOs from all over Europe with all synchronous zones presented are as follows:

• European generation mix is characterised by the synchronous units (more than 60%) and by wind and solar generation penetration being generally lower than 25%. Transmission networks circuits in all European synchronous zones are mostly based on overhead lines (more than 80%) and usage of cables is not common.

• European TSOs prefer to carry out PQ monitoring using predominantly fixed, permanently installed monitors. Close to 50% of the TSOs have more than 20 fixed monitors in operation and close to 25% have more than 100 fixed monitors. The median value of monitors used is 3 for portable monitors and 45 for fixed monitors.

• Most of the TSOs in Europe (91%) carry out PQ monitoring. Primary motivation for PQ monitoring is maintaining standard compliance (65%) and customers complaint (57%) followed by defining emission limits (43%).

• European TSOs prefer to use inductive type (45%) and capacitive type (35%) voltage transformers as transducers for PQ monitoring. Challenges for harmonics measurements have been acknowledged by some TSOs and use of CVTs with special components and RC-dividers is increasing.

REPORT

of 158

• There is no single dominant phenomenon influencing the decision to carry out PQ monitoring, and those who do carry out monitoring will observe a variety of phenomena. The interest in harmonics (87% of all respondents), flicker (83%), voltage regulation (74%), dips & swells (74%) is closely followed by unbalance, frequency excursions, interruptions and voltage steps 65%, 65%, 57% and 52% respectively.

• When carrying out measurements in the network, in over 75% of the cases all three phases are monitored and the measurement of system current is more common than measurement of specific customer currents.

• Roughly two-thirds of voltage measurements are line to neutral voltages, rather than line-to-line voltage.

• European TSOs (around 50%) see a possibility to use PMUs for PQ measurements. It is preferred though, to use PMUs for monitoring frequency excursions (62% of all respondents) and high and low voltages (46%).

• PMUs are preferred to be used in transmission networks with voltages above 220 kV (80%). Each individual TSO uses only a few different models of PMUs. For over 80% of the respondents, one of two different PMU suppliers are used.

• Close to 40% of the TSOs have more than 20 PMUs in operation. The median value of number of PMUs used in transmission network is 10 and average value is 13.

• Main PQ issues in Europe are mostly high and low voltages (35% of the respondents) and harmonics (35%).

• It is highlighted by TSOs that the observed PQ issues are mostly caused by the customers (69%) (industry more than distribution networks). Other sources but, but to a lesser extent, mentioned are WPP and HVDC links.

• Over half of the European TSOs use PQ mitigation and preferred device (over 45%) is passive filter. These units are mostly installed in networks with voltage levels higher than 150 kV and typical size varies between 40 and 150 Mvars.

REPORT

of 158

4 PQ Monitoring and Assessment

4.1 PQ Monitoring

4.1.1 Introduction PQ monitoring in electric power systems has increased in recent years. This trend can be seen for all voltage levels from LV up to EHV. One of the reasons for higher interest in PQ monitoring can be the developments in enabling technology (monitoring equipment, communication technology, data storage and processing) [223]. The fast development has made it possible to monitor at a large scale and to record virtually any parameters of interest [224]. Monitoring of voltages and currents gives the network operator information about performance of their network, both for the system as a whole and for individual locations and customers. However, there is also a pressure from the customer side and regulatory agencies to provide more and more detailed information about the actual PQ levels in the grids [224]. The change in components types connected to the network, such as non-conventional, PE-interface connected generators as well as envisaged further increase in non-conventional types of loads/storage (e.g. electric vehicles) additionally call network operators attention to monitor and document various aspects of network performance [223]. Beside verification of limits given by standards or the investigation of customer complaints, the PQ monitoring can be viable as forecasting method to optimise network operation or support investment planning [224]. Monitoring of PQ in a utility is heavily influenced by its regulatory environment or by a utility’s individual practice in specific country. For some parameters (voltage level, harmonics) the requirements are often clearly defined (by standards) and are mandatory. Other parameters may be monitored and reported, but are less often regulated. The same applies at individual sites, where a customer may be sensitive to PQ issues or a customer may need to be constrained to control the pollution and disturbances which it produces [3].

4.1.2 Objectives for PQ Monitoring The main aspects of PQ monitoring are the number of sites to be monitored and their location, the selection of monitoring parameters and the duration of monitoring, as well as the selection of monitoring technology. All aspects of PQ monitoring deployment are influenced by the objectives that the utility is seeking to address. As such, the most important step in deployment of a PQ monitoring system is identification of the system’s objective(s). Utilities presently monitor PQ for several important reasons. In general, the following main objectives for PQ monitoring can be distinguished [3]:

• Compliance verification – compares a defined set of PQ parameters with limits given by standards, rules or regulatory specifications;

• Performance analysis/Benchmarking – is usually an issue for a network operator and results are used for internal purposes (e.g. strategic planning, asset management, etc.);

• Site characterization – is used to describe PQ at a specific site in a detailed way in purpose to answer pre-connection questions on PQ of a specific customer;

REPORT

of 158

• Troubleshooting – the measurements are always based on a PQ problem (e.g. exceeding levels, equipment damage, etc.);

• Advanced applications and studies – cover specific measurements and analyses, sophisticated methods to improve the efficiency of system operation;

• Active PQ management – includes all applications where any kind of network operation control is derived from the measurement results.

The PQ monitoring objectives are described and analysed more detail in CIGRE working group C4.112 report [3], where beside the description of an objective its typical tasks have also been pointed out.

4.1.3 Selection of Monitoring Locations Selection of PQ monitoring locations is dependent on the objectives of the monitoring. In classic power systems, PQ monitoring points are usually located at the frontiers between classic generation, transmission, distribution and customer [3]. Related to the increased penetration of renewable and distributed generation, the monitoring at their connection points is also highly recommended. The PQ monitoring can be permanent or temporary. In some cases, the combination of these two approaches (hybrid) is also used. For permanent monitoring PQ monitors intended for fixed installation are usually used, while mobile monitors are selected for temporary monitoring. Fixed monitors are usually cheaper than mobile ones, but fix installed systems need a powerful infrastructure that may cause significant additional costs [224]. In some cases, a combination of fixed and mobile siting can achieve a good trade-off between cost and benefit. In power industry, several methods are used to select locations for installing PQ monitors. These include the following [3]:

• Random selection of the monitoring locations/site/point; • Monitoring at required sites as defined by the regulator; • Monitoring at a number of sites such that a statistically representative sample is achieved; • Selection based on identified or reported system user quality complaints; • Monitoring at sites where important/sensitive system users are, or will be connected; • Monitoring at sites with expected high levels of PQ disturbances/high probability of future

problems; • Sites that are important for the operation of the system.

The number of sites to be monitored is also an important aspect. The number of sites to be monitored can range from a single site, for example in case of a PQ investigation, to thousands of sites, in the case of a TSO or DSO system assessment [3]. The rapid deployment of smart metering devices with PQ functionality contributes in increasing number of monitored sites. The expanded discussion on the selection of monitoring locations has been brought out in [3]. The selection of the monitoring locations is a complex task requiring a good knowledge of the power system architecture and the PQ disturbances affecting the grid. Generally, the selection of location for PQ monitoring depends on utility’s objectives, the architecture of the power system, the length of the monitoring period and cost related issues [2].

REPORT

of 158

4.1.4 Duration of Monitoring One major factor influencing the monitoring system location as well as the costs is the monitoring duration. In the existing practice of PQ monitoring there are the following approaches regarding to the duration of PQ monitoring [3]:

• Long-term, continuous PQ monitoring, with fixed instrument installation; • PQ monitoring with portable instruments, with a rotating approach, where a monitor stays

at site for a specific period of time to capture a sample of measurements and then is moved to another site;

• Temporary and short-term PQ monitoring, with mobile or handheld instrument, mostly for a period of time sufficient for problem identification (i.e. for troubleshooting).

Each PQ parameter varies differently over time, influencing the structure of the data management and the duration of monitoring as well as the nature of reporting [3]. As a simplification, Table 4-1 below indicates the overall spectrum of timescales for different PQ parameters. Table 4-1: Appropriate timescales for monitoring different PQ phenomena [3]

Parameter Timescale of interest

Voltage transient <20-ms Voltage dip 10-ms to 2-s Frequency excursion Possibly one minute Phase unbalance Possibly one day Harmonics (percentile) 1 week Flicker (percentile) 1 week Dip/temporary over-voltage statistics 3 years

Where portable monitors are used, monitoring must be over a time-period sufficient to determine the normal characteristic operating cycles at the site. In general, one or two full business cycles may be sufficient for this purpose. For most of the commercial or residential loads, the business cycle is one week [3]. Monitoring in EHV and HV networks should be long-term and continuous at all measuring locations and performed by fixed, permanent PQ monitors [3]. The disturbances, which occur only momentary during long periods of time, like transients, voltage dips and temporary power frequency over-voltages or interruptions of supply, cannot be measured in a traditional continuous manner. However, in practice monitoring at least with transient recorders is highly recommended, for example in HVDC stations. In that case, the data can be available if disturbance occur or ignored and deleted over time-period if there is no disturbances in the system.

4.1.5 Selection of Monitoring Parameters The PQ parameters to be monitored depend on the PQ monitoring objectives and the further use of an information gathered with the monitoring. It is neither technically nor economically feasible to monitor all PQ parameters at all locations. Very detailed monitoring may lead to difficulties when analysing data. As general rule, it is better to use aggregate variables such as maximum, minimum, distortion, specific harmonics, flicker, etc. instead of very small averaging windows [224]. Furthermore, it has

REPORT

of 158

to be decided whether measurement of voltage and current, or measurement of voltage only is necessary for the planned monitoring system. The implementation of voltage only measurements is usually the cheapest way. As an example, for compliance verification voltage only measurement can be sufficient, while in case of analysing costumer complaints additional current measurements can be of essential importance. In case of measurement systems for performance analysis, additional current measurements can provide useful information [224]. The most relevant standard is the IEC PQ monitoring standard IEC 61000-4-30 [73], which gives the main basis and recommendations for PQ monitoring parameters. The required parameters for each monitoring objective named in Subsection 4.1.2 is more broadly discussed in [3]. As following, some of the most frequent PQ monitoring objectives and parameters have considered. For compliance verification purpose the parameters and intervals are defined by the relevant standard (e.g. EN 50160 [8], IEC 61000-2-12 [225] or IEC TR 61000-3-6 [139]). In case of performance, analysis and advanced applications all the PQ parameters may not be required. Typically, the utility will select those, which consider being useful or more interesting based on the previous experiences. Often variables with indicative limits are chosen, e.g. dips, over-voltages, as well as shorter integration periods if required. For utility performance assessment, the number of monitored locations may reach a large number. Therefore, a scalable, multi-level approach for data handling is necessary. The selection of parameters and averaging intervals, in case of site characterization, mainly depends on the information demanded by the customer. Variables stated in compliance standards (i.e. EN 50160) are likely to be included. However, certain parameters may be more important depending on the nature of the customer load. Troubleshooting usually requires the most flexible possibilities in parameter and aggregation interval section. Due to the limitation to a certain time interval and few sites only, more parameters including the waveform data at high resolutions should be measured. Including voltage and current is important for most of the troubleshooting cases [223].

4.1.6 Selection of Monitoring Equipment The evolution of new technologies (sensors, intelligent electronic devices, software, and telecommunication) and standardisation of related products are facilitating the PQ monitoring in larger number of sites. The trend toward an integration of PQ functions into devices e.g. smart meters, protection relays, etc., could reduce cost of future PQ monitoring systems [224]. For the whole measurement chain, a certain overall accuracy has to be guaranteed, including the measurement device itself and the required transducers [224]. To determine and quantify the influence of instrument transformers on the overall uncertainty on PQ measurements, it is necessary to simultaneously consider the electrical behaviour of an instrument transformer for a given disturbance and the measurement method [221]. The influence and use of instrument transformers for PQ measurements is handled in IEC TR 61869-103 [221]. Instrument transformers are used to provide PQ measurement instruments’ a signal suitable for their input channels and containing all the relevant information needed for primary signals. Such information may be the accurate reproduction of the primary signal or may give to the instrument the relevant information in order to reconstruct the PQ parameters of the primary signal [221].

REPORT

of 158

One aspect in selection of instrument transformer is the measurement time interval. The signal from the instrument transformer is digitalised and processed over several time intervals [221]:

• Measure of RMS voltage over 1 cycle refreshed each half-cycle (VRMS(1/2); • Measures over a period of 10/12 cycles for 50/60 Hz; • 150/180 cycles aggregation for 50/60 Hz of 10/12 cycles measurement; • 10-min aggregation of 10/12 cycles measurement; • 2-h aggregation of 10/12 cycles measurement.

In Table 4-2, all the PQ disturbances for which IEC 61000-4-30 [73] provides a measurement method and time intervals are listed. Table 4-2: PQ disturbances and measurement intervals [221]

Disturbance 1 cycle 10/12 cycles

150/180 cycles

10-min 2-h Other

Power frequency 10-s Magnitude of voltage x x x x Flicker x x Dips and temporary over-voltage x Voltage interruption x Voltage unbalance x x x x Voltage harmonics x x x x Voltage interharmonics x x x x Main signalling x Rapid voltage changes x U/O deviation parameters x x x x x

Instrument transformer can affect the magnitude of the main signalling voltage seen by the instrument. The measure of PQ disturbances measured requires improved frequency response (magnitude and phase) as well as transient response from the transformer [221]. The relation between the requirements and the PQ disturbances is shown below in Table 4-3 (according to IEC 61000-4-30). Table 4-3: Transformer parameters influencing PQ measurement [221]

Disturbance Magnitude Phase Transients

Power frequency Magnitude of voltage x x Flicker x x Dips and temporary over-voltage x x x Voltage interruption x x x Voltage unbalance x x Voltage harmonics x x Voltage interharmonics x Main signalling x Rapid voltage changes x x U/O deviation parameters x x

The impact of instrument transformers on PQ measurements is related to technology adopted, to design and manufacturing details.

REPORT

of 158

Many contributions are available about frequency response of instruments transformers. Such information can be useful in order to assess instrument transformers behaviour in use for the measurement of PQ parameters. Special attention should be paid when higher voltages and currents are involved, since it may not be representative of the real behaviour of the sensor. The instrument transformer frequency response and measurement behaviour is in more detail discussed in [221]. The approximate overview of the useful frequency range of available instrument transformer technologies can be seen in Figures 4-1 and 4-2. The accuracy of the harmonic measurements and the voltage transformers frequency response in transmission system level has also discussed in [226]-[227].

Figure 4-1: Voltage transformer technologies frequency range [221]

Figure 4-2: Current transformer technologies frequency range [221]

REPORT

of 158

4.2 PQ Assessment

4.2.1 Introduction To get the needed information from PQ measurements, the obtained data must be analysed first. Main problems of PQ assessment are extracting information from enormous amount of data and presenting it in a useful manner. This section describes different aspects of analysing data and gives brief overview how to report results in the most efficient way. Methods to analyse obtained PQ monitoring data depend on different aspects [3]:

• The objective of PQ monitoring; • The standards, grid codes, regulatory policies and existing bilateral contracts, which give

guidance with respect to analysis and reporting of PQ data; • The duration of monitoring; • Monitoring locations and number of monitors, and availability of communication

infrastructure of the system; • The power system being monitored; what, how and where the monitoring is carried out

depending on the location of local production/generation; • Available technologies to manage large amounts of data.

Table 4-4: PQ monitoring data presentation approaches

Objective Indices Reporting intervals

Reporting tools

Reporting ways Correlation of

interests

Compliance verification

Standards, regulators or bilateral contracts

Expected by standards, regulators or bilateral contracts

Internal or external reports, web, mobile devices

Percentiles, indices, tables

Network operational events

Performance analysis

Standards, regulators, coefficients

Medium- and long-term campaigns

Internal reports, papers, web

Indices and their historical trends, contour charts, frequency histograms

Network configuration parameters/ meteorological data

Site characterization

Standards, connection rules

Short- and long-term campaigns

Internal reports, papers

Those expected to characterised pre-connection or existing customers performances

Customers data

Troubleshooting

Standards and bilateral contracts for a specific PQ problem

Few weeks monitoring campaigns

Confidential reports

Those requested by a specific PQ problem involving equipment damages evaluation

Different electric parameters and customers data

Advanced application and studies

According to the objective of application or study

Dependent on the purpose, mainly long-term campaigns

Confidential reports, internal reports (dependent on information)

Those requested by the objective(s) of studies

Operational events, electrical parameters, meteorological data, configuration

Active PQ management

According to the management application

Real-time, short and medium-term campaigns

Confidential reports, internal reports (dependent on information)

Those requested by the objective(s) of application

Operational events, network configuration, electrical parameters

For the simplified comparative overview of different objectives and their approaches, Table 4-4 is composed [3], [223]. Most important factor to PQ assessment is the objective of PQ monitoring. The objectives of PQ monitoring are in detail discussed in subsection 4.1.2. Once the objective of PQ

REPORT

of 158

monitoring is clear, it is easier to establish the importance of other factors. There are simple objectives, like compliance verification that has only the outcome of “pass” or “fail”, but there are also complicated objectives that need deeper analysis methods for system assessment, planning etc. In the following, the analysis and reporting methods for different monitoring objectives are more deeply discussed.

4.2.2 PQ Reporting

In general, PQ disturbances can be classified as continuous or discrete disturbances (also as variations or events and steady-state or non-steady-state disturbances). The classification of PQ disturbance is important factor considering the relation of how to measure, analyse and report PQ disturbances. Steady-state disturbances (e.g. steady-state voltage, harmonics) are described by constant numbers, which means they can be measured with simple hand-held devices. Field measurements have however time variations of indices. In this case, the quantities may be presented by a trend graph as shown in the example on the Figure 4-3. In Figure 4-3, the disturbance value is measured as average values over pre-defined intervals of time. For reporting and presentation of disturbance values, it is necessary to reduce these time varying values to one or more numbers, which can be used to characterise the performance of a single or many sites [3]. They are calculated as statistical measures, such as average or maximum of the measured values. In practice, the average value is usually too optimistic and maximal value too pessimistic. For average, the disturbance level is higher for a considerable part of the time. However, the maximal value occurs very briefly. Considering mentioned circumstances, the practical way to assess PQ indices is to use 95% value (also referred as the 95% cumulative probability level or 95th percentile [3]), the value which is not exceeded for 95% of the time.

Figure 4-3: Trend graph including percentile values [3] Calculating PQ indices for continuous disturbances is straightforward as there is a time varying trend for the disturbance. However, there is a challenge with reporting discrete disturbances, such as voltage dips or transients, which occur as isolated disturbances over a few cycles with a generally long interval of repetition. This means that the concept of “steady-state” is not applicable for these cases, which makes the calculation for discrete disturbance indices significantly more problematic. Discrete disturbance is observable by the presence of a voltage/current cycle, which is different

REPORT

of 158

enough to exceed the limit level programmed in PQ monitor. This event can be captured by saving its parameters together with a timestamp. The lack of widely accepted appropriate indices for discrete disturbances complicates the PQ assessment process. Without distinct indices it is difficult to provide simple and understandable reports on performance at a single site, and even harder for the multiple sites [3]. For example, there have been developed methods to define indices for voltage dips [62], [228]-[229] but no single method has been accepted internationally. For other discrete disturbances, such as transients and temporary power frequency over-voltages, defining indices is a work in progress but even less behind. The following subsections discuss on the PQ monitoring and assessment according to previously mentioned objectives in Subsection 4.2.1.

4.2.3 Compliance Verification Compliance verification compares a predefined set of measured PQ parameters with limits given by standards, rules or regulatory specifications. This is usually done for individual sites with a simple result of “pass” or “fail”. Compliance verification may be done for different reasons [3]:

• Measurements made to check PQ limits to regulatory requirements; • Monitoring for bilateral contracts; • Compliance check for voltage quality at a specific site with given standards; • Compliance check for a connected equipment/installation against given standards.

Compliance verification may require continuous monitoring, which gives a set of data to be examined over a defined time frame in order to produce an index of comparison with the appropriate limit or regulation [3]. No single data analysis interval can be highlighted, because the requirements for the data analysis interval may change depending on the type of compliance assessment being performed or the standards and regulations to be applied. Usually, the reporting period covers long intervals. To assess compliance verification, the indices needed depend on the limits that they are compared with. For continuous disturbances, the general approach is the comparison of the statistical indicator of the disturbance (like RMS voltage, harmonics and flicker severity) with the limit. According to [3], the following parameters, dependent on the limits or regulations to comply with, must be addressed:

• Basic averaging windows are 3-s, 1-min, 10-min and 2-h mean RMS values (usually 10-min values are used);

• Statistical indicators are 5th, 95th 99th percentile values; • Evaluation period for the percentile calculation ranging from 1-h to several years (usually 1-

week period is used); • The period covered by the report is user-defined, but it is usually 1-year.

Reporting of the PQ disturbance should be kept as simple as possible. Key parameter to be expressed is whether the site is compliant or not. For the non-compliant sites, the performance analysis should be made in a way, where the problem or limit that makes the site non-compliant is made clear. For discrete disturbances (e.g. voltage dips, transients) evaluation techniques and limits are much less defined and further work is needed.

REPORT

of 158

4.2.4 Performance Analysis/Benchmarking Performance analysis is mainly done for system operator by a routine monitoring of sites throughout the system. The objective of performance analysis is to gain information about the system overall PQ for network planning, asset management etc., or for benchmarking. The typical reason for performance analysis is to assess average PQ for a specific region or long-term trends for a specific set of sites. Performance assessment data may also be used for benchmarking between system operators. However, it is not widely recommended, since IEC standard 61000-4-30 [73] points out that any attempt to normalise outputs of different PQ surveys will be extremely complex and open to misinterpretation [3]. For performance analysis in PQ monitoring, the recommended reporting period 1-year is suggested. In that case, it is possible to analyse the impact of all seasonal changes throughout the year. For trending the data analysis interval, it is freely selectable by the needs of reporting period (e.g. for monthly trending purposes an interval of month may be selected). One issue in performance analysis is the large amount of data that needs to be assessed and put in a readable format. If a single site is analysed, the assessment and reporting is simple, the basic graphical visualisation and indices can be used, as shown as an example in Figure 4-4.

Figure 4-4: Example of a histogram for continuous disturbances in single site [3] In case, if many sites are of interest the suggested way to effectively present and analyse the data is through hierarchical analysis and reporting structure. A bottom-up methodology should be applied for data analysis [3]. Individual sites should be analysed and reported first. Indices calculated from each site may then be used to build up the indicators, which can be used to characterise the performance of the entire system. In this way, the higher level management of system operator can get highly summarised information about system performance, but PQ specialist can get specific and detailed data for use in PQ strategies and planning. The suggested analysis and reporting methods in interest of the performance of large number of sites is more closely explained in [3].

REPORT

of 158

For disturbance levels, a histogram is a good way to represent large amount of data or using a bar chart may be also very informative, as shown as an example in Figure 4-5. If an overview of the distribution of all sites are required, a graph showing site disturbance level bars distributed based on the percentile ranges, may be used as shown in Figure 4-6.

Figure 4-5: The bar chart of disturbance levels at sites [3]

Figure 4-6: Distribution of site disturbance levels [3]

4.2.5 Site Characterization Site characterization is used to describe PQ at a specific site and in a detailed way. Usually it is initialized by the request from a customer, system operator or a regulator. The selected monitoring parameters and measurement methods are strongly dependent on the objective of monitoring. The required reporting intervals for data analysis vary depending on the characteristics of the site to be assessed. Data analysis interval must be long enough to incorporate at least one full cycle of normal activity at the site, so that periods with high or low disturbance levels may be captured and reported. For most applications, a week is generally acceptable data analysis and reporting interval [3].

REPORT

of 158

For continuous disturbances, it is suggested in [3] to calculate statistical measures and produce them in tabular form, as shown as an example in Table 4-5, and for a graphical presentation as a time varying trend, as in Figure 4-7, or as a histogram. Table 4-5: Presenting tabulated statistical data [3]

Weeks 95th percentile Maximum

Value, % Instant

02/04/2012-08/04/2012 1.37 1.82 08/04/2012 15:40 09/04/2012-15/04/2012 1.34 1.51 14/04/2012 13:50 16/04/2012-22/04/2012 1.35 1.99 22/04/2012 14:10 23/04/2012-29/04/2012 1.99 1.79 28/04/2012 23:30 Number of weeks exceeding the standard limit (8%): 0

Figure 4-7: Time varying disturbance trend plot [3]

4.2.6 Troubleshooting Troubleshooting measurements are always done for a specific PQ problem (e.g. exceeding pre-defined limits, equipment damage, etc.). For industrial customers, poor PQ in connection points may lead to interruption of production, damage to the equipment or producing processes, which leads to a loss of profit. Monitoring for troubleshooting will normally include specific measurements, but a permanent monitoring program may provide input that can be used as a post-problem assessment after a customer complaint. Troubleshooting measurement may follow a compliance verification measurement if limits are not met [3]. Often the cause for equipment failure is a fault in network, which is a non-steady state phenomenon (e.g. voltage dip, temporary power frequency over-voltage, rapid voltage change or voltage transient). However, the problem could also be steady-state phenomenon, like unbalance or harmonics. In some cases, the issues origin from emissions or operating characteristics of the customer devices [3]. Because of that, the main consideration of troubleshooting is to have measurement data from the occurrence of equipment failure. If the cause is non-repetitive, even the longest period measurements will not lead to the cause.

REPORT

of 158

The measuring intervals for troubleshooting given by [3] are minimal of one week up to a number of weeks. If the problem is caused by steady-state phenomena, it may be sufficient to measure for a couple of days. In this case, it is probable that the equipment will fail again and the recorded failure moment will give needed information to identify the cause. Customers to whom PQ problems have caused significant financial loss, it is sensible to install a permanent PQ monitoring device at the point of connection or close to the sensitive equipment. There are no widely accepted methods of reporting troubleshooting investigations, because almost every case differs depending on the troubleshooting problem. Reporting details and level of the analysis is dependent on who conducts the analysis (e.g. system operator, PQ consulting company, impacted customer). However, for troubleshooting it is suggested to use the same principles as for compliance verification and site characterization, as pointed out respectively in Subsections 4.2.3 and 4.2.5.

4.2.7 Advanced Application and Studies Advanced studies include more specific measurements and analyses, which usually are not connected to daily business. Advanced application and studies covers new highly sophisticated methods to improve the efficiency of the system operation. They are becoming more popular due to its higher resolution. For advanced application and studies there are no widely accepted forms of data analysis, reporting interval or reporting requirements, since they are highly dependent on the purpose of the study or application. The analysis and reporting is in the form of providing as much data as required or as possible [3]. For some purposes, the analysis and reporting can be made as already discussed in previous objectives.

4.2.8 Active PQ Management Active PQ management includes all applications where system control is triggered by PQ measurement results (offline or real-time control). The main reason is to identify potential problems before they can become major issues. At present, the most used active PQ management applications are related to active harmonics filtering and dynamic reactive power compensation [3]. There are two separate data analysis and reporting methods with active PQ management. In first method, the real time measurements give information to SCADA (Supervisory Control and Data

Acquisition), where the most important issue is the measurement and data transmission speed. The knowledge and experience regarding to this method is limited for now [3]. In second method, information is reported in post-event, where the monitoring purposes and information needed guide the assessment. In this case, analysis and reporting methods mentioned in previous objectives are applicable.

REPORT

of 158

4.3 PMU Measurements

4.3.1 Introduction Synchronized phasor measurement technology has become available after the processor technologies and computer relaying became more advanced. It was first introduced in 1980s in the United States. Nowadays, many TSOs and research facilities have a system based on Phasor Measurement Units (PMU) to study the possibilities and applications for synchronized phasor measurements. Due to the high resolution and sufficient sampling rate, the PMUs could also be used for assessing PQ in power systems. This has been seen as one promising future application as the increasing number of PMUs is being installed and additional applications are developed. However, the possibilities and applications regarding PQ monitoring still need to be studied and analysed. PQ phenomena such as voltage dips, voltage unbalance, frequency deviation etc., are in theory easy to cover with PMU measurements. Further assessment is needed for harmonic measurements and influence of harmonics on PMU measurements. PMUs are meant to measure voltages and currents and time synchronize the output signals. By time synchronization, it is generally possible to obtain simultaneous measurements from all over the system. In this way, it is possible to create a relatively precise picture about the current state of power system. This compared to the traditional measurements in different points of power system will give an advantage and provides additional possibilities for system security enhancement. Phasor is a method to present steady-state sinusoidal waveforms for AC circuit analysis. It is a complex number with certain magnitude and phase. Sinusoidal signal in general can be presented with the following Function 4-1. H( ) = IJ cos(N + O) (4-1) Where Xm is the signal amplitude peak value and ω and ϕ are respectively the signal frequency and the phase angel in radians. The phasor representation of this signal is following I ≡ Q�√� S.T = Q�√� (UVWO + XWYZO) (4-2)

The magnitude of the phasor in Function 4-2 is given as RMS value of the sinusoid. The sinusoidal signal and its phasor representation are illustrated in Figure 4-8. The illustration implies that the signal remains unchanged at all times, while in the practice there is finite data window. In many PMUs the used data window is one period of the fundamental frequency of the input signal [230]. It should also be noted that the phasor phase angle is strictly connected to the initial time instant used as a reference (t=0). For this reason, the phase angle is a relative concept that has to be referred correctly to the initial time when the phasor measurement is needed [231].

REPORT

of 158

Figure 4-8: Sinusoidal signal phasor representation; (a) sinusoidal signal, (b) phasor representation [230] The phasor, as such have been used in power system analysis and calculations for a long time. On the other hand, the term synchrophasor is relatively new and directly linked to PMUs. Synchrophasor is committed to link together the phasor concept and its measurements in a straightforward way. The synchrophasor estimation is based on the same idea as phasor but the difference is that the phasor and its phase angle are calculated by using Coordinated Universal Time (UTC) as a time reference. This way it is possible to have a reference for all sinusoidal signals measured in a wide geographical area. Therefore, the synchrophasors calculated at different network points can be easily compared and it provides significant information on the current state of power system. Expression of the synchrophasor is the same as for the phasor, presented as Function 4-1. However, in case of synchrophasor the phase angle is defined to be 0 degree when the maximum x(t) occurs at the UTC time instant T. In other words, the synchrophasor measured at UTC reference time instant T, represents the RMS amplitude and the phase angle of the sinusoidal signal at time instant T [231]. General scheme of PMU is illustrated in Figure 4-9. PMU consists of modules for signal conditioning, data digitalisation, time synchronization and communication for measurement outputs.

Figure 4-9: General scheme of PMU [231]

REPORT

of 158

The signal conditioning module is needed to adapt analogue signals to digital acquisition circuits, including the anti-aliasing filter, which reduces the band of electrical signals. The anti-aliasing filter is a low-pass filter, where the cut-off frequency depends on the needed measurements. For PMUs, the relevant information is presented around the fundamental frequency. Accordingly, the anti-aliasing filter in the signal conditioning module should cut all higher frequencies. Its configuration depends strongly on the adopted A/D converter and the functionalities included in the device [231]. It should be noted that the anti-aliasing filter in the input of PMU produces phase delay. However, this is compensated by the PMUs estimation algorithm [230]. The A/D converter produces digital signal to processing unit, where the phasor estimation algorithm is being used for measurement calculations. The sampling frequency of an A/D converter is strictly related to the frequency response of an anti-aliasing filter and its input range depends on the voltage and/or frequency range considered as input signals [231]. For A/D converter module, the important characteristic is the conversion time from analogue to digital, which is the time required to obtain the digital value from analogue signal. In order to get the correct synchronized measurements, the conversion time should be taken into account and have to be compensated. The basis of PMU is the time synchronization, which can be either internal or external of the PMU. The time synchronization relies on satellite receiver. In case of external time synchronization, the protocols like IRIG-B and Precision Time Protocol (PTP) are used [231]. The GPS synchronization gives a time reference (UTC) to the measurements, reporting instant to the PMU and it can also be used as a trigger for some instant (e.g. data acquisition to operator systems, for example SCADA). Processing steps of the phasor signal in the unit are shown in Figure 4-10. The accuracy of the PMU signal processing depends on the characteristics of the low-pass filters before the signal conversion. However, there are also filters available after the signal conversion from analogue to digital and phasor estimation that may have a significant effect on measurement output. These filters differ from manufacturer to manufacturer and depend on the configuration of PMU. They may be used for attenuating harmonics and interharmonics and for reducing the response time of the synchrophasor.

Figure 4-10: PMU phasor signal processing model [232]

4.3.2 Standards for PMU Measurements

The synchrophasor was first time standardised by the IEEE 1344 “Standard for Syncrophasors for

power systems” in 1995, renewed in 2005 to IEEE C37.118 [233], and in 2011, divided into two

C37.118.1 [232] and C37.118.2 [234], defining the static and dynamic performance requirements

REPORT

of 158

for measurement units and data transfer respectively. The C37.118.1 was followed by an amendment

C37.118.1a [235] in 2015.

In C37.118.1 the specific evaluation criterions for PMU performance assessment have been defined.

The synchrophasor measurements are evaluated using the Total Vector Error (TVE) criterion, which

combines amplitude and phase differences for values obtained from a PMU and theoretical values of

a synchrophasor representation. TVE calculation determines the differences between the ideal value

of the theoretical synchrophasor and the estimate given by the PMU. TVE is defined by the following

Equation (4-3).

1[�(Z) = \�Q'] (�)$Q'(�) �^�Q_] (�)$Q`(�) ��Q'(�) �^�Q`(�) � (4-3)

Where I] (Z) and Ia] (Z) are the sequences of estimates given by PMU, and I(Z) and I�(Z) are the sequences of theoretical values of the input signal at the instant of time n assigned by the unit to those values. For frequency and Rate of Change of Frequency (ROCOF) measurement evaluations, two criterions are defined. The errors FE (Function 4-4) and RFE (Function 4-5) are absolute values of the difference between the theoretical and the estimated values given in Hz and Hz/s, respectively. b� == |��d� − �J�e-d�*| = |∆��d� − ∆�J�e-d�*| (4-4) gb� == |(��/� �d�) − (��/� J�e-d�*)| (4-5) In addition to the mentioned TVE, FE and RFE, other PMU measurement evaluation criteria are available [233].The standard C37.118.1 provides clarification on the evaluation of measurement response and delay time, reporting latency, measurement and operational errors, measurement reporting and reporting rates and times. Standard also includes some examples.

4.3.3 Testing of PMUs In order to assess the PMU performance and suitability for PQ monitoring and assessment in conditions where harmonics are presented in systems a study was performed. The purpose was to obtain knowledge about the response of PMUs in conditions where the voltage and current input signals include higher harmonics. The aim was to simulate network conditions with high penetration of PE sources with emphasis on harmonics. For PMU testing, a special testing scheme as shown in Figure 4-11 was composed. The testing scheme includes signal generator, time synchronization module, PMUs and Phasor Data Concentrator (PDC). In this testing campaign, three PMUs from different manufacturers were used. All devices were time synchronized using separate GPS module and corresponding IRIG-B signal. Separate PDC was used for data collection, and MATLAB was used for the data analysis. The PMU testing study included various tests that were grouped as following:

• Test A with the fundamental harmonic and including harmonics from order of 2nd to 49th; • Test B with various simultaneous harmonic combination, based on actual network

measurements taken from the WPP connection point;

REPORT

of 158

• Test C with interharmonics. In addition to the described conditions, separate analysis on the effect of PMU internal filters was made.

Figure 4-11: PMUs testing scheme In PMU testing, the signal generator was used as a testing device, which is generally used for protection relay testing. This signal generator can generate voltage 3x300 V (85 VA) and current 3x12.5 A (70 VA at 7.5 A). Voltage and current outputs can be synchronized with external time source via IRIG-B time input. According to IEEE C37.118.1 [232], the testing device used to verify performance in synchrophasor measurements shall have a test uncertainty ratio, at least 4 compared with test requirements, which is 1% (TVE). Therefore, the signal generator used in PMU testing study with the accuracy for voltage and current signal output ±0.1% for both magnitude and phase, guarantee the required accuracy. In order to test whether different harmonic distortion have effect on synchrophasor measurement, a base test was carried out by excluding harmonic distortion in voltage and current signal. The purpose was to identify possible errors in PMUs under testing by using ideal 50 Hz sinusoidal input signal (fundamental harmonic), which gives a comparison point for synchrophasor measurements in following Tests A, B and C. In all tests carried out, the PMUs measured positive-sequence synchrophasor and frequency. The measurements data was transmitted to the PDC. The measurement error was calculated in MATLAB by comparing measured data with the one inserted into signal generator. Test A

Test A was performed with harmonic distortion in PMUs input signal, in order to observe the PMU measurements compliance with the standard IEEE C37.118.1 [232] requirements. The test was designed for class M PMUs, which shall have 1% maximum TVE for single higher harmonic distortion. The test signal consisted of fundamental harmonic together with single higher harmonic, which magnitude was 10% of the fundamental harmonic magnitude, as required in standard. The higher harmonics included were from 2nd to 49th order. In test, both voltage and current harmonic distortion were observed. Two different PMUs filter settings were used to define the response for harmonic

REPORT

of 158

distortion in input signal. In the first part, the testing device sent out a signal with a pure fundamental harmonic component, and in second part the signal was changed so that it contained the fundamental harmonic together with a single higher harmonic (2nd-49th order). Test B

Test B was carried out to assess various simultaneous harmonic combinations according to measurements taken from real WPP connection point. As in previous test A, also in test B two different PMUs filter settings were analysed. Test B was derived from PQ measurements in transmission network at two different time instants (Case 1 and Case 2). The content of harmonics in test B cases is shown in Table 4-6. In addition, Figure 4-12 illustrates the Case 1 harmonic content. Table 4-6: Harmonic content in Test B Case 1 and Case 2

Case 1 Case 2

Voltage Current Voltage Current

Harmonic order

% of fundamental

harmonic

Harmonic order

% of fundamental

harmonic

Harmonic order

% of fundamental

harmonic

Harmonic order

% of fundamental

harmonic 1 100 1 100 1 100 1 100 3 0.5 3 1.1 3 0.5 3 1.0 5 0.7 5 0.7 5 0.6 5 0.2 7 0.5 7 1.0 7 0.5 7 1.0 11 0.4 11 1.0 11 0.3 11 0.9 13 0.7 13 1.0 13 0.6 13 1.1 17 1.8 17 0.8 17 1.7 17 0.6 19 2.5 19 0.9 19 2.4 35 0.8 22 2.5 25 1.0 23 1.7 37 1.2 25 2.0 33 0.2 25 1.1 47 0.8 29 0.4 35 1.0 49 0.8 35 0.3 37 1.3 47 0.8 49 0.8

Figure 4-12: Harmonic content in Test B Case 1; CP95 – harmonic value 95% of measurement period, CP99 – harmonic value 99% of measurement period, CPMax – harmonic maximum value

REPORT

of 158

Test C

Test C was performed by implementing interharmonic distortion in PMUs input signal. Frequencies of interharmonics deviated ±25 Hz around the fundamental harmonic and 19th order higher harmonic. Therefore, four different cases were analysed. The input signal included fundamental harmonic with interharmonic, which magnitude was 10% of the fundamental harmonic magnitude. As in previous test, also in Test C two different PMU filter settings were used.

4.3.4 Results

Test A In Test A, the PMUs measurement accuracy calculated in TVE stayed below the required 1% limit [5]. For voltage harmonics the TVE was mostly around 0.1% and for current harmonics it was in a range of 0.08-0.25%. The result are shown in Figure 4-13.

Figure 4-13: Average voltage TVE

It can be seen from Figure 4-13 that PMU A and PMU B showed a significantly higher TVE in voltage harmonic insertion response with 19th and 21st harmonic order. However, the voltage TVE stayed still below the required 1% limit. The exact reason for this behaviour is unknown and needs further assessment.

REPORT

of 158

Figure 4-14: Average current TVE

In Figure 4-15, the current harmonic insertion showed also different response accuracy at 19th and 21st harmonic order for PMU A and PMU B. Contrary in this case, for PMU B the accuracy increased at 21st harmonic order. The results show that there is no connection between higher harmonic order and TVE change. There can be seen some relatively minor variations in the TVE change but these are arbitrary and can be caused by the testing device error. Also considering the PMUs filter settings, the results show a marginal difference between the two filter settings. Therefore, the PMUs filter settings does not seem to influence the PMUs measurement accuracy. Test B The results from Test B with various simultaneous harmonic combinations are shown in Figure 4-15 to Figure 4-16. The results shown correspond to all PMUs voltage and current harmonic TVE with the reference curves (dashed lines). Latter indicates the PMUs TVE calculated from the fundamental harmonic (pure 50 Hz signal) input measurement results (base test). For both Case 1 and 2 in Test B, the PMUs voltage and current harmonics TVE was in a range of 0.05-0.26%, which is below the 1% limit required by standard [232]. From the shown results, it can be seen that in some cases the PMU accuracy seems to increase with the higher harmonic insertion (e.g. Figure 4-15). However, the deviation in accuracy could be explained by the testing device error but further research is needed.

REPORT

of 158

Figure 4-15: Current (left) and voltage (right) harmonics TVE in Case 1

Figure 4-16: Current (left) and voltage (right) harmonics TVE in Case 2

Test C

The Test C results represent the interharmonic distortion effect to the PMUs measurements accuracy. In case for 25 Hz interharmonic distortion, the PMU B voltage and current harmonics TVE in the first set of filters reached over 2.5%, which is greater than the required 1% limit. Other two, PMU A and PMU B have voltage and current harmonics TVE in the required range. On the other hand, while changing the filter settings, the PMU A voltage and current harmonics TVE exceeds the required limit by being more than 6%. Whilst the PMU B has in this case the voltage and current harmonics TVE below the 1% limit (0.2%). In case of 75 Hz interharmonic distortion, the Test C results are represented in Figure 4-17 and Figure 4-18. Similar to the previous interharmonic distortion for the first set of filters, PMU B voltage and current harmonics TVE is more than 2.5%. With the second set of filters (Figure 4-18), the PMU A voltage and current harmonics TVE is more than 6%, while PMU B voltage and current harmonics TVE is 0.2%.

REPORT

of 158

Figure 4-17: 75 Hz current (left) and voltage (right) harmonics TVE with first set of

filters

Figure 4-18: 75 Hz current (left) and voltage (right) harmonics TVE with second set of filters In case of 925 and 975 Hz interharmonics, the voltage and current harmonics TVE values for all tested PMUs were below the 1% requirement and did not vary much from the previous test A result. The voltage and current harmonics TVE changes are shown in following Table 4-7. The result indicate that around 50 Hz frequency (fundamental harmonic) component, in some cases the bandwidth of a PMU anti-aliasing filter can let through interharmonics near the fundamental harmonic component, which decreases the PMU measurement accuracy. However, as seen from 25 and 75 Hz interharmonic distortion, the PMUs accuracy can be increased by changing the filters configuration. Nevertheless, it should be noted that the filters configuration possibilities depend on the manufacturer and the PMUs internal calculation algorithms.

REPORT

of 158

Table 4-7: TVE change in percentiles with different interharmonics

Filter

setting PMU

25 Hz interharmonic

75 Hz interharmonic

925 Hz interharmonic

975 Hz interharmonic

Current TVE

change

%

Voltage TVE

change

%

Current TVE

change

%

Voltage TVE

change

%

Current TVE

change

%

Voltage TVE

change

%

Current TVE

change

%

Voltage TVE

change

%

First filter setting

A 65.20 1935.50 72.80 1927.10 -3.10 71.50 -3.30 32.30 B 1216.80 3039.50 1215.40 3042.8 -10.10 17.40 -11.60 12.00 C -33.30 -8.70 -29.30 -8.30 -29.80 -5.60 -32.80 -8.80

Second filter setting

A 2563.2 36335.0 2554.4 36193.0 93.00 229.1 -1.00 41.70 B 6.00 6.70 9.00 5.40 1.00 4.20 1.90 7.20 C 22.90 20.20 14.70 6.90 1.70 2.30 -2.20 3.30

REPORT

of 158

5 Conclusion and Future Work

5.1 Conclusion This report presents a comprehensive overview of critical PQ phenomena and sources of PQ disturbances in transmission networks, with the focus on PE rich power systems. It considers both, the existing industrial practice by TSOs as well as studies made by research institutions, and provides background for critical PQ issues in existing and future power systems. The main results are as follows:

1. The report points out possible causes for traditional type of PQ disturbances (frequency deviation, steady-state voltage variations, voltage dips and temporary power frequency over-voltages, voltage fluctuations and flicker, voltage unbalance, harmonic distortion) in modern and future transmission networks and brings out the consequences, as well as gives recommendations for mitigation methods and solutions, resulting from information presented by academia and from practical experience reported by system operators.

2. Besides traditional PQ disturbances relatively new topic, related to the significant penetration of PE device in transmission networks and their control system interactions, named harmonic stability, is introduced. Discussion given includes the background of the phenomena, what are the consequences, possible mitigation methods and delivers an overview of available studies.

3. The report includes analysis of PQ in modern and future power systems and provides the

main criteria for understanding the impact of PE on PQ phenomena and PQ phenomena effect on PE.

a. PE devices have an impact on PQ phenomena. The main phenomena under

observation are frequency deviation, voltage fluctuations and flicker, harmonic distortion and harmonic stability.

b. PQ phenomena, e.g. voltage unbalance, temporary power frequency over-voltages, voltage dip and steady state voltage variation, seem not to have direct relation regarding the impact of PE on PQ.

c. In case of impact of PQ phenomena on PE it is acknowledged that many devices are susceptible to voltage dips, temporary over-voltages, and harmonic distortion in the supply network and some to voltage unbalance. High impact can be considered in case of harmonic stability phenomena. However, additional assessment is needed to verify these impacts in case of high penetration of PE in future transmission networks.

d. In case of impact of PE to PQ phenomena it is acknowledged that for phenomena as frequency deviation, steady-state voltage variations, voltage fluctuation and flicker further research is required as currently the impact of these is not known or it is limited.

4. More detailed PQ analysis is presented for some European TSOs (TenneT, ELES, Elering,

EirGrid). An overview of the network, its generation and consumption together with PQ

REPORT

of 158

overview (conditions, issues, etc.) and PQ monitoring system, is presented for each TSO. The highest concern seems to be the injected harmonics, which are directly caused by the increase of PE based devices. In addition, harmonic stability issues are also of relevance.

5. The results of the survey that was performed in order to obtain information on range of PQ related issues in transmission networks are presented. Analysis is based on the answers from 23 different European TSOs, including at least one TSO from each of the synchronous area in Europe. The main conclusions are as follows:

a. European generation mix is characterized by the synchronous units (more than 60%

of the respondents) and by wind and solar generation penetration being generally lower than 25%. Transmission networks circuits in all European synchronous zones are mostly based on overhead lines and usage of cables is not common.

b. European TSOs prefer to carry out PQ monitoring using predominantly fixed, permanently installed monitors.

c. Primary motivation for PQ monitoring is maintaining standard compliance (65%) and customers complaint (57%) followed by defining emission limits (43%).

d. European TSOs (around 50%) see a possibility to use PMUs for PQ measurements. e. Main PQ issues in Europe are mostly high and low voltages and harmonics.

6. Detailed discussion and analysis on background and requirements for PQ monitoring and

assessment is given in the report.

7. The report includes results from the study to assess the influence of harmonics to the operation of PMUs. Based on the study it can be concluded that in general the harmonics do not have an effect on PMUs accuracy but there may be cases where significant influence can be seen.

5.2 Future Work The following presents some key areas that are identified as extensions of the work presented in this report with respect to MIGRATE project and especially in the field of WP5. Development of PE Numerical Simulation Models for PQ Studies

To perform simulation studies and to analyse PQ disturbances and their propagation in electric networks with a high penetration of PE devices, numerical models of the integrated PE devices have to be developed. Based on the required accuracy of simulation models and type of simulations, load flow, RMS and EMT models will be considered. Detailed, EMT models of PE devices can include several hundreds of semiconductor switches, wherefore very small numerical integration time steps are needed in order to accurately compute fast switching events. This type of simulations are capable to reproduce detailed transient responses of the device and therefore it requires detailed modelling of the control algorithms to accurately control the operation of the PE device. According to that, EMT simulations are time consuming and demand higher processing power.

REPORT

of 158

The computational expense can be lowered when using reduced models (EMT or RMS), which provide a similar behaviour and dynamic response in comparison with detailed models. These reduced models are called “average models” and can be used to emulate the average response of switching PE devices by using controlled electric sources and simplified control functions. The average models can be used in the three-phase time-domain simulations (EMT) or in a domain of single frequency signals (RMS). The load flow simulations represent the simplest approach to define steady-state conditions in electrical network. The models of PE devices are therefore simplified and do not include any dynamic behaviour. The load flow models are for the harmonic load flow simulation purposes tuned with reference to the detailed EMT models. Numerical simulations of electric networks with detailed average and simple PE models allow the evaluation of PQ disturbances in the grid and furthermore the development of studies for cost efficient mitigation techniques in order to achieve a higher value for PQ. For this purpose, three types of models will be developed for several devices, including VSC-based devices and DFIG-based wind power generation.

Propagation of PE Disturbances Through Power Networks

The proliferation of PE-based devices in power network is expected to continue to grow. These devices have facilitated a significant level of flexibility and efficiency in operation of power networks. However, they are known to be both some of the major sources and “victims” of PQ disturbances. Therefore, the studies about the temporal and spatial variation of multiple sources of PQ disturbance and how these disturbances propagate throughout the network are critical for the planning and operation of future networks. Due to the increased variation in the output of sources of disturbance, spatial and temporal, the conventional deterministic studies may not suffice and probabilistic analysis of PQ performance must be considered. The renewable generation, wind and solar in particular, and various types of energy storage technologies are all PE interfaced. The increased reliance on HVDC transmission and use of FACTS devices in AC networks as well as proliferation of PE interfaced loads at lower voltage levels are further contributing to variation in PQ levels. The level of contribution of majority of these depends on the operating conditions and loading levels. The high variability in the output and hence variability in contribution to PQ levels in the network is already incorporated in a way in the benchmarking standards and codes by considering statistical measures (e.g. 95th percentiles) for the assessment of network performance. Increased deployment of PMUs and new types of power and energy metering devices (e.g. “smart meters”) will facilitate better evaluation of the PQ performance of individual buses and the network as a whole through increased network observability. The increased amount of measurement data, albeit very useful for system wide evaluations, can prove challenging to analyse and to extract useful information from. Therefore, techniques for structuring, analysing and visualizing the data must be investigated and continuously improved. Assessment of the Influence of PQ Disturbance on Operation of PE Rich Power Networks

Based on the developed numerical simulation models in previous tasks, this task focuses on the influence of PQ disturbances on the operational behaviour of future power networks with a high penetration of PE devices. The methodology for the assessment will be proposed in line with the

REPORT

of 158

developed scenarios in MIGRATE project WP1. As a result of this task, an estimation of the influence of PQ disturbances, including harmonics, frequency variation and voltage variation, on power grids with a high penetration of PE devices will be given. Additionally, sensitivity threshold as well as expected levels of PQ disturbances will be determined. Mitigation of PQ and Provision of Differentiated PQ

Based on the results of previous tasks, mitigation techniques for PQ disturbances will be proposed. During this task, penetration limits for PE devices will be assessed in cooperation with MIGRATE project WP1. The proposed mitigation techniques include device based solutions in order to improve the level of PQ in the electric network. Additionally, network based solutions will be proposed with the aim of postponing network reinforcements. Considering the reduction of PQ disturbances, the capability of PE-based devices other appropriate PQ mitigation devices will be investigated. The proposed mitigation techniques will address the estimated increase of harmonics, frequency variations and voltage variations in the electric power network.

REPORT

of 158

References [1] R.C. Dungan, M.F. McGranaghan, S. Santos, H. W. Beaty, Electrical Power Systems Quality,

2nd edition, McGraw-Hill, 2002 [2] CEER, 5th CEER Benchmarking Report on The Quality of Electricity Supply, 2011 [3] CIGRE/CIRED TB 596, JWG C4.112, Guidelines for Power Quality Monitoring, 2014 [4] P. Kundur, Power System Stability and Control, McGraw-Hill, 1993 [5] J. Machowski, J. Bialek, J. Bumby, Power System Dynamics: Stability and Control, Wiley,

2012 [6] W. Winter, K. Elkington, G. Bareux, J. Kostevc, Pushing the Limits: Europe’s New Grid:

Innovative Tools to Combat Transmission Bottlenecks and Reduced Inertia, IEEE Power and Energy Magazine, Vol. 13, No. 1, Januar 2015, pp. 60-74

[7] J. Morren, S.W.H. de Haan, W.L. Kling, J.A. Ferreira, Wind Turbines Emulating Enertia and Supporting Primary Frequency Control, IEEE Transactions on Power Systems, Vol. 21, No. 1, February 2006, pp. 433-434

[8] EN 50160 Voltage Characteristics of Electricity Supplied by Public Distribution Systems, 2010

[9] P. Kundur, J. Paserba, V. Ajjarapu, G. Andersson, A. Bose, C. Canizares, N. Hatziargyriou, D. Hill, A. Stankovic, C. Taylor, T.V. Cutsem, V. Vittal, Definition and Classification of Power System Stability, IEEE/CIGRE Joint Task Force on Stability Terms and Definitions, IEEE Transactions on Power Systems, Vol. 19, No. 3, August 2004, pp. 1387-1401

[10] ENTSO-E, Frequency Stability Evaluation Criteria for the Synchronous Zone of Continental Europe, Technical Report, 2016

[11] J.J. Grainger, W.D. Stevenson, Power System Analysis, McGraw-Hill, 1994 [12] D. Gautam, L. Goel, R. Ayyanar, V. Vittal, T. Harbour, Control Strategy to Mitigate the

Impact of Reduced Inertia due to Doubly Fed Induction Generators on Large Power Systems, IEEE Transactions on Power Systems, Vol. 26, No. 1, February 2011, pp. 214-224

[13] L. Jin, S. Yuan-zhang, P. Sorensen, L. Guo-jie, G. Weng-zhong, Method for Assessing Grid Frequency Deviation due to Wind Power Fluctuation Based on Time-Frequency Transformation, IEEE Transactions on Sustainable Energy, Vol. 3, No.1, January 2011, pp. 6573

[14] C. Luo, B.T. Ooi, Frequency Deviation of Thermal Power Plants due to Wind Farms, IEEE Transactions on Energy Conversion, Vol. 21, No.3, September 2006, pp. 708-716.

[15] H. Banakar, C. Luo, and B. T. Ooi, Impacts of wind power minute-to-minute variations on power system operation, IEEE Transactions on Power Systems, vol. 23, pp. 150–160, Februar 2008

[16] C. Luo, H.G. Far, H. Banakar, P.K. Keung, B.T. Ooi, Estimation of Wind Penetration as Limited by Frequency Deviation, IEEE Transactions on Energy Conversion, Vol. 22, No. 1, September 2007, pp. 783-791

[17] E. Lannoye, D. Flynn, M. O’Malley, Evaluation of Power System Flexibility, IEEE Transactions on Power Systems, Vol. 27, No. 2 May 2012, pp. 922-931

[18] A.M. Jain, B.E. Kushare, P. Gautam, N.N. Jangle, K.K. Wagh, Effect of Change in Frequency on Operation and Performance of Three Phase Induction Motor, International Journal on Recent Trends in Engineering and Technology, Vol. 7, No. 2, July 2012, pp. 117-120

[19] B.S. Abdulraheem, C.K. Gan, Power System Frequency Stability and Control: Survey, International Journal of Applied Engineering Research, Vol. 11, No. 8, 2016, pp. 5688-5695

REPORT

of 158

[20] Y.G. Rebours, D.S. Kirschen, M. Trotignon, S. Rossignol, A Survey of Frequency and Voltage Control Ancillary Services - Part II: Economic Features, IEEE Transactions on Power Systems, vol. 22, No.1, February 2007, pp. 358-366

[21] P.K. Keung, P. Li, H. Banakar, and B.T. Ooi, Kinetic Energy of Wind-Turbine Generators for System Frequency Support, IEEE Transactions on Power Systems, Vol. 24, No. 1, February 2009, pp. 279-287

[22] J. Ekanayake, N. Jenkins, Comparison of the Response of Doubly Fed and Fixed-Speed Induction Generator Wind Turbines to Changes in Network Frequency, IEEE Transactions on Energy Conversion, Vol. 19, No. 4, December 2004, pp. 800-802

[23] GE Energy, Windinertia™ Control, Facts Sheet, 2009, [Online] http://site.ge-energy.com/prod_serv/products/renewable_energy/en/downloads/GEA17210.pdf

[24] ENERCON GmbH, Der Beitrag von Windenergie und angeschlossener Technologien fur ein stabiles Netz, Technical Report, 2015, [Online] https://www.hskarlsruhe.de/fileadmin/hska/EIT/Aktuelles/seminar_erneurbare_energien/Sommer_2015/Folien/2015-04-08_Enercon.pdf

[25] A.A. Girgis, W.L. Peterson, Adaptive Estimation of Power System Frequency Deviation and its Rate of Change for Calculating Sudden Power System Overloads, IEEE Transactions on Power Delivery, Vol. 5, No. 2, April 1990, pp. 585-594

[26] L. Tang, J. McCalley, Two-Stage Load Control for Severe Under-Frequency Conditions, IEEE Transactions on Power Systems, Vol. 31, No. 3, May 2016, pp. 1943-1953

[27] Deutsche Energie-Agentur GmbH, Systemdienstleistungen 2030: Sicherheit und Zuverlassigkeit einer Stromversorgung mit hohem Anteil erneuerbarer Energien, Technical Report, 2014

[28] ENTSO-E, Deterministic Frequency Deviations – Root Causes and Proposals for Potential Solutions, A joint EURELECTRIC – ENTSO-E Response Paper, 2011, [Online] http://www.eurelectric.org/media/26970/frequency_deviations_final_february_2012-2012-030-0186-01-e.pdf

[29] J. Brisebois, N. Aubut, Wind Farm Inertia Emulation to Fulfill Hydro-Quebec’s Specific Need, 2011 IEEE Power and Energy Society General Meeting, July 2011, pp. 1-7

[30] ENTSO-E, Network codes on Operational Security, September 2013 [31] E. Fuchs, M.A.S. Masoum, Power Quality in Power Systems and Electrical Machines, 2nd

edition, Elsevier, 2015 [32] M.H.J. Bollen, I. Gu, Signal Processing of Power Quality Disturbances, Wiley, 2006 [33] C.W. Taylor, Power System Stability. McGraw-Hill, 1994 [34] N. Flatabo, R. Ognedal, T. Carlsen, Voltage Stability Condition in a Power Transmission

System Calculated by Sensitivity Methods, IEEE Transactions on Power Systems, Vol. 5, No. 4, November 1990, pp. 1286-1293

[35] IEEE 1159 – IEEE Recommended Practice for Monitoring Electric Power Quality, 2009 [36] J. Dixit, A. Yadav, Electrical Power Quality, Laxmi Publications, 2010 [37] J.V. Milanović, J. Meyer, R.F. Ball, W. Howe, R. Preece, M.H.J. Bollen, S. Elphick, N.

Čukalevski, International Industry Practice on Power-Quality Monitoring, IEEE Transactions on Power Delivery, Vol. 29, No. 2, April 2014, pp. 934-941

[38] M. Hermann, H.D. Kreye, R. Reinisch, U. Scherer, J.V.H. Berndt, Tansmissioncode 2007 - Network and System Rules of the German Transmission System Operators, Technical Report Verband der Netzbetreiber VDN, 2007

REPORT

of 158

[39] Y.G. Rebours, D.S. Kirschen, M. Trotignon, S. Rossignol, A Survey of Frequency and Voltage Control Ancillary Services - Part I: Technical Features, IEEE Transactions on Power Systems, Vol. 22, No. 1, February 2007, pp. 350-357

[40] M. Adibi, W. Alexander, B. Avramovic, Overvoltage Control During Restoration, IEEE Power Engineering Review, Vol. 12, No. 11, November 1992, pp. 43

[41] X. Jiang, J.H. Chow, A.A. Edris, B. Fardanesh, E. Uzunovic, Transfer Path Stability Enhancement by Voltage-Sourced Converter-Based Facts Controllers, IEEE Transactions on Power Delivery, Vol. 25, No. 2, April 2010, pp. 1019-1025

[42] R.S. Wibowo, N. Yorino, M. Eghbal, Y. Zoka, Y. Sasaki, FACTS Devices Allocation with Control Coordination Considering Congestion Relief and Voltage Stability, IEEE Transactions on Power Systems, Vol. 26, No. 4, November 2011, pp. 2302-2310

[43] S. Ertem J.R. Tudor, Optimal Shunt Capacitor Allocation by Nonlinear Programming, IEEE Transactions on Power Delivery, Vol. 2, No. 4, October 1987, pp. 1310-1316

[44] E. Vittal, M. O’Malley, A. Keane, A Steady-State Voltage Stability Analysis of Power Systems with High Penetrations of Wind, IEEE Transactions on Power Systems, Vol. 25, No. 1, February 2010, pp. 433–442

[45] H.T. Le, S. Santoso, T.Q. Nguyen, Augmenting Wind Power Penetration and Grid Voltage Stability Limits Using ESS: Application Design, Sizing, and A Case Study, IEEE Transactions on Power Systems, Vol. 27, No. 1, February 2012, pp. 161-171

[46] R.R. Londero, C.d.M. Affonso, J.P.A. Vieira, Long-Term Voltage Stability Analysis of Variable Speed Wind Generators, IEEE Transactions on Power Systems, Vol. 30, No. 1, January 2015, pp. 439-447

[47] F. Capitanescu, Assessing Reactive Power Reserves with Respect to Operating Constraints and Voltage Stability, IEEE Transactions on Power Systems, Vol. 26, No. 4, November 2011, pp. 2224–2234

[48] F. Dong, B. Chowdhury, M. Crow, L. Acar, Improving Voltage Stability by Reactive Power Reserve Management, IEEE Transactions on Power Systems, Vol. 20, No. 1, February 2005, pp. 338-345

[49] V. Venkatasubramanian, J. Guerrero, J. Su, H. Chun, X. Zhang, F. Habibi-Ashrafi, A. Salazar, B. Abu-Jaradeh, Hierarchical Two-Level Voltage Controller for Large Power Systems, IEEE Transactions on Power Systems, Vol. 31, No. 1, January 2016, pp. 397-411

[50] C.W. Taylor, Concepts of Undervoltage Load Shedding for Voltage Stability, IEEE Transactions on Power Delivery, Vol. 7, No. 2, April 1992, pp. 480-488

[51] B. Otomega, T.V. Cutsem, Undervoltage Load Shedding Using Distributed Controllers, IEEE Transactions on Power Systems, Vol. 22, No. 4, November 2007, pp. 1898-1907

[52] Reseau de Transport d’Electricite, A.M. Denis, 2013 Reliability Report, 2014, [Online] http://www.rtefrance.com/sites/default/files/system_reliability_report_2013.pdf

[53] IEC TR 61000-2-1 Electromagnetic compatibility (EMC) - Part 2: Environment - Section 1: Description of the environment - Electromagnetic environment for low-frequency conducted disturbances and signalling in public power supply systems, 1990

[54] G. Olguin, Stochastic Assesmenf of Voltage Dips Caused by Faults in Large Transmission System, Thesis for the Degree of Licentiate of Engineering, Chalmers University of Technology, Sweden, 2003

[55] P. Heine, M. Lehtonen, Voltage Sag Distributions Caused by Power System Faults, IEEE Transaction on Power Systems, Vol. 18, No. 4, November 2003, pp. 1367-1373

REPORT

of 158

[56] ENTSO-E News, Solar Eclipse: Europe’s TSOs Teamed Up Successfully, March 2015, [Online] https://www.entsoe.eu/news-events/announcements/announcements-archive/Pages/News/solar-eclipse-europes-tsos-teamed-up-successfully.aspx

[57] S. Eftekharnejad, V. Vittal, G.T. Heydt, B. Keel, J. Loehr, Impact of Increased Penetration of Photovoltaic Generation on Power Systems, IEEE Transaction on Power Systems, Vol. 28, No. 2, May 2013, pp. 893-901

[58] K. Kawabe, K. Tanaka, Impact of Dynamic Behavior of Photovoltaic Power Generation Systems on Short-Term Voltage Stability, IEEE Transaction on Power Systems, Vol. 30, No. 6, November 2015, pp. 3416-3423

[59] IEC TR 61000-2-14 Electromagnetic compatibility (EMC) –Part 2-14: Environment – Overvoltages on public electricity distribution networks, 2006

[60] R. Sedaghati, N.M. Afroozi, Y. Nemati, A. Rohani, A. R. Toorani, N. Javidtash, A. Heydarzadegan, H. Sedaghati, A Survey of Voltage Sags and Voltage Swells Phenomena in Power Quality Problems, International Journal of Scientific Research and Management, Vol. 1, No 9, Decemer 2013, pp. 458-462

[61] IEC 61000-2-4 Electromagnetic compatibility (EMC) - Part 2-4: Environment - Compatibility Levels in Industrial Plants for Low-Frequency Conducted Disturbances, 2002

[62] CIGRE TB 261, WG. C4.07, Power Quality Indices and Objectives, 2004 [63] IEEE 1564 – IEEE Guide for Voltage Sag Indices, 2014 [64] D. Sabin, M. Bollen, Overwiev of IEEE STD 1564-2014 Guide for Voltage Sag Indices, 23rd

International Conference on Elecrticity Distribution, Lyon, 15-18 June 2015 [65] Commission Regulation (EU) 2016/631, Establishing a Network Code on Grid Connection of

Generators, Official Journal of the European Union, April 2016 [66] IEEE 493 - IEEE Recommended Practice for the Design of Reliable Industrial and

Commercial Power Systems, 2007 [67] IEC TR 61000-2-8, Electromagnetic Compatibility (EMC) – Part 2-8: Environment - Voltage

Dips and Short Interruptions on Public Electric Power Supply Systems with Statistical Measurement Results, 2002

[68] ACM News, AMC Introduces Quality Standard for the High-Voltage Grids, December 2013, [Online] https://www.acm.nl/en/publications/publication/12438/ACM-introduces-quality-standard-for-the-high-voltage-grids/

[69] B. Franken, V. Ajodhia, K. Petrov, K. Keller, C. Müller, Regulation of Voltage Quality, 9th International Conference Electrical Power Quality Utilisation, Barcelona, October 2007

[70] G. Stein, F. Ghassemi, Grid Code Review Panel – Issues Assessment Proforma Voltage Fluctuation, Grid Code Limits on Rapid Voltage Changes, National Grid, November 2013, pp. 1-8, [Online] http://www2.nationalgrid.com/UK/Industry-information/Electricity-codes/Grid-code/Modifications/GC0076/

[71] EirGrid, MPID 264 – Power Quality, Recommendation to CET by Eirgrid of Modification to Grid Code (CC.10.13), [Online] http://www.eirgridgroup.com/site-files/library/EirGrid/MPID_264_Power_Quality_Harmonics_Recommendation_Paper.pdf

[72] IEC TR 61000-3-7 Electromagnetic compatibility (EMC) - Part 3-7: Limits - Assessment of Emission Limits for the Connection of Fluctuating Installations to MV, HV and EHV Power Systems, 2008

[73] IEC 61000-4-30 Electromagnetic Compatibility (EMC) - Part 4-30: Testing and Measurement Techniques - Power Quality Measurement Methods, 2015

REPORT

of 158

[74] M. McGranaghan, Dealing with Voltage Sags in Your Facility, EPRI-PEAC Corporation, EC&M Archive, September 2003 [Online] http://www.ecmweb.com/powerquality/electric_dealing_voltage_sags/, 2006

[75] C.J. Melhorn, T.D. Davis, G.E. Beam, Voltage Sags: Their Impact on Utility and Industrial Customers, IEEE Transation on Industry Applications, Vol. 34, No. 3, May 1998, pp. 549-558

[76] CIGRE TB 372, TF C4.102, Voltage Dip Evaluation and Prediction Tools, 2009 [77] S.Z. Djokic, K. Stockman, J.V. Milanovic, J.J.M. Desmet, R. Belmans, Sensitivity of AC

Adjustable Speed Drives to Voltage Sags and Short Interruptions, IEEE Transactions on Power Delivery, Vol. 20, No. 1, January 2005, pp. 494-505

[78] S.Z. Djokic, J.V. Milanovic, D.J. Chapman, M.F McGranaghan, D.S Kirschen, A New Method for Classification and Presentation of Voltage Reduction Events, IEEE Transactions on Power Delivery, Vol. 20, No. 4, 2005, pp. 2576-2584

[79] S.Z. Djokic, J. Desmet, G. Vanalme, J.V. Milanovic, K. Stockman, Sensitivity of Personal Computers to Voltage Sags and Short Interruptions, IEEE Transactions on Power Delivery, Vol. 20, No. 1, January 2005, pp. 375-383

[80] W.Kanokbannakorn, T.Saengsuwan, The Studies of a Transmission System and Distribution System Voltage Sags Correlations, 5th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology, 2008, ECTI-CON 2008, 14-17 May 2008, pp. 1033-1036

[81] J.V. Milanovic, C.P. Gupta, Probabilistic Assessment of Financial Losses due to Interruptions and Voltage Sags - Part I: The Methodology, IEEE Transaction on Power Delivery, Vol. 21, No.2, April 2006, pp. 918–924

[82] J.V. Milanovic, C.P. Gupta, Probabilistic Assessment of Financial Losses due to Interruption and Voltage Sags - Part II: Practical Implementation, IEEE Transaction on Power Delivery, Vol. 21, No. 2 April 2006, pp. 925–932

[83] S.C. Vegunta, J.V. Milanovic, Estimation of Cost of Downtime of Industrial Process Due to Voltage Sags, IEEE Transaction on Power Delivery, Vol. 26, No.2 April 2011, pp. 576-587

[84] J.V. Milanovic, Y. Zhang, Global Minimization of Financial Losses Due to Voltage Sags With FACTS Based Devices, IEEE Transactions on Power Delivery, vol. 25, No.1, January 2010, pp. 298-306

[85] CIGRE TB 467, WG C4.107, Economic Framework for Power Quality, 2011 [86] F.C. Pereira, O.C.N. Souto, J.C. De Oliveira, A.L.A. Vilaca, P.F. Ribeiro, An Analysis of Costs

Related to The Loss of Power Quality, 8th International Conference On Harmonics And Quality of Power Proceedings, 1998, 14-16 October 1998, pp. 777-782

[87] M. Samotyi, The Cost of Power Disturbances to Industrial & Digital Economy Companies, EPRI and the Electricity Innovation Institute, June 2001

[88] D. Chapman, The Cost of Poor Power Quality, Power Quality Application Guide, Copper Development Association, 2001

[89] R.P. Bingham, Sags and Swells, Dranetz-BMI, February 1998 [90] V.M. Atlason, J. Kessel, C.L. Bak, Propagation of Transmission Level Switching

Overvoltages Related to the Distribution Level, Projectgroup EPSH-810, Institute of Energy Technology, Aalborg University, May 2007

[91] M. Florkowski, J. Furgat, M. Kuniewski, Propagation of Overvoltages Transferred Through Distribution Transformers in Electric Networks, IET Generation, Transmission & Distribution, Vol. 10, No. 10, April 2016

REPORT

of 158

[92] M. Florkowski, J. Furgat, M. Kuniewski, Propagation of Overvoltages in Distribution Transformers with Silicon Steel and Amorphous Cores, IET Generation, Transmission & Distribution, Vol. 9, No. 16, 2015

[93] J.Y. Chan, Framework for Assessment of Economic Feasibility of Voltage Sag Mitigation Solutions, PhD in the Faculty of Engineering and Physical Sciences, The University of Manchester, 2010

[94] Y. Zhang, Techno-Economic Assessment of Voltage Sag Performance and Mitigation," PhD in the Faculty of Engineering and Physical Sciences, University of Manchester, 2008

[95] A.M. Sannino, M.G. Miller, M.H.J. Bollen, Overview of Voltage Sag Mitigation, IEEE Power Engineering Society Winter Meeting, 23-27 January, 2000, pp. 2872-2878

[96] R. Lasseter, C. Hochgraf, Unbundling Power Quality Services: Technical Issues, 30th Hawaii International Conference on System Sciences, 7-10 January, 1997, pp. 581-588

[97] C. Alvarez, J. Alamar, A. Domijan, J.A. Montenegro, Z. Song, An Investigation Toward New Technologies and Issues in Power Quality, 9th International Conference on Harmonics and Quality of Power, 1-4 October 2000, pp. 444-449

[98] S. Chen, G. Joos, Series and Shunt Active Power Conditioners for Compensating Distribution System Faults, 2000 Canadian Conference on Electrical and Computer Engineering, 7-10 May, 2000

[99] J. Woodworth, Understanding Temporary Overvoltage Behaviour of Arresters, ArresterWorks, August 2011, [Online] http://www.arresterworks.com/arresterfacts/pdf_files/ArresterFacts_028_Understanding _Temporary_Overvoltage_Behavior_of_Arresters.pdf

[100] C. Srisailam, A. Sreenivas, Mitigation Of Voltage Sags/Swells By Dynamic Voltage Restorer Using Pi And Fuzzy Logic Controller International Journal of Engineering Research and Applications, Vol. 2, No. 4, July-August 2012, pp. 1733-1737

[101] R. Nittala, A.M. Parimi, K.U. Rao, Mitigation of Voltage Sag/Swell with CSI-IDVR, 2015 International Conference on Recent Developments in Control, Automation and Power Engineering (RDCAPE), 12-13 March 2015, pp. 33-37

[102] A. Pandey, Rajlakshmi, Dynamic Voltage Restorer and Its application at LV & MV Level, International Journal of Scientific & Engineering Research, Vol. 4, No. 6, June 2013, pp. 668-671

[103] R. Omar, N.A Rahim, Modeling and Simulation for Voltage Sags/Swells Mitigation Using Dynamic Voltage Restorer (DVR), Australasian Universities Power Engineering Conference (AUPEC'08), 14-17 December 2008

[104] H. Awad, F Blaabjerg, Mitigation of Voltage Swells by Static Series Compensator, 35th Annual IEEE Power Electronics Specialists Conference (PESC 04), 20-25 June 2004, pp. 3505-3511

[105] M. Tesarova, M. Kaspirek, Evaluation of Long-Term Voltage Dip Monitoring In HV, MV and LV Networks, 23rd International Conference on Electricity Distribution, Lyon, 15-18 June 2015

[106] M.R A. Qader, M. Bollen, R.N. Allan, Stochastic Prediction of Voltage Sags in a Large Transmission System, IEEE Transaction on Industry Applications, Vol. 35 No. 1, January/February 2002, pp. 152-162

[107] M. Mohseni, S.M. Islam, M.A.S. Masoum, Impacts of Symmetrical and Asymmetrical Voltage Sags on DFIG-Based Wind Turbines Considering Phase-Angle Jump, Voltage Recovery, and Sag Parameters, IEEE Transaction on Power Electronics, Vol. 26, No. 5, May 2011, pp. 1587-1598

REPORT

of 158

[108] B.Q. Khanh, N.H. Phuc, Prediction of Voltage Sag in The Transmission System of Vietnam, A Case Study, 2011 IEE/PES Power Systems Conference and Exposition (PSCE), 20-23 March 2011

[109] D.J. Figueroa, R.R. Avila, M.B.C. Salles, Analysis of transmission Systems with High Penetration of Wind Power using DFIG Based Wind Farms During Voltage Sags, 2013 International Conference on Clean Electrical Power (ICCEP), 11-13 June 2013, pp. 361-367

[110] CIGRE TB 412, WG C4.110, Voltage Dip Immunity of Equipment and Installations, 2010 [111] M.A.M. Hassan, F.M.A. Bendary, M.H.Shahin, E.M. Saied, Voltage Swell Mitigation in

Practical EHV Network Using Flexible AC Transmission Systems Based on Evolutionary Computing Method; GIGRE C4-107, 2016

[112] D.L. Brooks, R.C. Duga, M. Waclawiak, A. Sundaram, Indices for Assessing Utility Distribution System RMS Variation Performance, IEEE Transactions on Power Delivery, Vol. 13, No. 1, January 1998, pp.254-259

[113] H.H. Schiesser, B. Heimbach, Customer Costs due to Major Voltage Disturbances: Estimations for an Urban Distribution Network in Switzerland, 22nd International Conference on Electricity Distribution (CIRED 2013), Stockholm, 10-13 June 2013

[114] Z. Klaić, S. Nikolovski, Z. Kraus, Voltage Variation Performance Indices in Distribution Network; Tehnički vjesnik, Vol. 18, No. 4, 2011, pp. 547-551

[115] A. Baggini, Handbook of Power Quality, Wiley, 2008 [116] J.C.Das, Power System Harmonics and Passive Filter Designs, Wiley-IEEE Press, 2015 [117] C. Vilar, J. Usaola, H. Amaris, A Frequency Domain Approach to Wind Turbines for Flicker

Analysis, IEEE Transactions on Energy Conversion , Vol. 18, No. 2, 2003, pp. 335-341 [118] B. Barahona, P. Sorensen, L. Christensen, T. Sorensen, H.K. Nielsen, X.G. Larsen,

Validation of the Standard Method for Assessing Flicker From Wind Turbines, IEEE Transactions on Energy Conversion, Vol. 26, No. 1, 2011, pp. 373-378

[119] W. Hu, Z. Chen, Y. Wang, Z. Wang, Flicker Study on Variable Speed Wind Turbines with Permanent Magnet Synchronous Generator, 13th Power Electronics and Motion Control Conference (EPE-PEMC 2008), 1-3 September 2008, pp. 2325-2330

[120] W. Hu, Z. Chen, Y. Wang, Z. Wang, Flicker Mitigation by Active Power Control of Variable-Speed Wind Turbines With Full-Scale Back-to-Back Power Converters, IEEE Transactions on Energy Conversion, Vol. 24, No. 3, 2009, pp. 640-649

[121] G. Chang, S. Lin, Y. Chen, H. Lu, H. Chen, Y. Chang, An Advanced EAF Model for Voltage Fluctuation Propagation Study, IEEE Transactions on Power Delivery, Vol. PP, No. 99, 2016

[122] M. Göl, O. Salor, B. Alboyaci, B. Mutluer, I. Cadirci, M. Ermis, A New Field-Data-Based EAF Model for Power Quality Studies, IEEE Transactions on Industry Applications, Vol. 46, No. 3, 2010, pp. 1230-1242

[123] IEC 61000-4-15 Electromagnetic Compatibility (EMC) - Part 4-15: Testing and Measurement Techniques – Flickermeter – Functional and Design Specifications, 2010

[124] IEC 61000-3-3 Electromagnetic compatibility (EMC) - Part 3-3: Limits - Limitation of Voltage Changes, Voltage Fluctuations and Flicker in Public Low-Votage Supply Systems, for Equipment with Rated current ≤16 A per Phase and Not Subject to Conditional Connection, 2013

[125] UIE, Union Internationale d’Electrothermie, Flicker Measurement and Evaluation, Technical Report, 2nd Edition, 1991

[126] Power Quality in Electrical Systems, Effects of Voltage Fluctuations on Electrical Equipment, September 2011, [Online] http://www.powerqualityworld.com/2011/09/effects-voltage-fluctuations.html

REPORT

of 158

[127] B.N. Ramos, J.L. deC. Parga, An EMTP Study of Flicker Generation and Transmission in Power Systems due to the Operation of an AC Electric Arc Furnace, 9th International Conference on Harmonics and Quality of Power, 1-4 October 2000, pp. 942-947

[128] M. Maksić, I. Papič, B. Blažič, Simple Continuous Assessment of Transmission-NetworkFlickerLevels Caused by Multiple Sources, IEEE Transaction on Power Delivery, Vol. 31, No. 6, December 2016, pp. 2546-2552

[129] B. Blažič, I. Papič, Analysis of measures for flicker mitigation in the Slovenian transmission network, Study in Slovene, Faculty of Electrical Engineering, University of Ljubljana, 2014

[130] D. Stade, H. Schau, A. Novitskiy, Flicker Analysis in the H.V. Transmission System, 10th International Conference on Harmonics and Quality of Power, 6-9 October 2002, pp. 541-546

[131] P. Caramia, G. C. Paola, Power Quality Indices in Liberalized Markets, Wiley, 2009 [132] D.B. Vannoy, M.F. McGranaghan, S.M. Halpin, W.A. Moncrief, D.D. Sabin, Roadmap for

Power Quality Standards Development, IEEE Transactions of Industry Applications, Vol. 43, No. 2, March/April 2007, pp. 412-421

[133] A. von Jouanne, B. Banerjee, Assessment of Voltage Unbalance, IEEE Transaction of Power Delivery, Vol. 16, No. 4, October 2001, pp. 782-790

[134] M.I. Hossain, R. Yan, T.K. Saha, Investigation of The Interaction Between Step Voltage Regulators and Large-scale Photovoltaic Systems Regarding Voltage Regulation and Unbalance,” IET Renewable Power Generation, Vol. 10, No. 3, 2016, pp. 299-309

[135] R. Yan, T.K. Saha, Voltage Variation Sensitivity Analysis for Unbalanced Distribution Networks Due to Photovoltaic Power Fluctuations, IEEE Transaction on Power Systems, Vol. 27, No. 2, May 2012, pp. 1078-1089

[136] P. Heine, M. Hyvärinen, J. Niskanen, A. Oikarinen, H. Renner, M. Lehtonen, The Impact of 50 Hz Railroad Systems on Utility System Power Quality, Electric Power Quality and Supply Reliability Conference (PQ 2010), 16-18 June 2010, pp. 61-66

[137] IEEE 112 – IEEE Standard Test Procedure for Polyphase Induction Motors and Generators, 2004

[138] NRS 048-2 Electric Supply – Quality of supply, Voltage characteristics, compatibility levels, limits and assessment methods, 2003

[139] IEC TR 61000-3-6 Electromagnetic compatibility (EMC) - Part 3-6: Limits - Assessment of Emission Limits for the Connection of Distorting Installations to MV, HV and EHV Power Systems, 2008

[140] EPRI Power Electronics Applications Center, Input Performance of ASDs During Supply Voltage Unbalance, Power Quality Testing Network (PQTN), No. 28, 1996

[141] ANSI C84.1 Electric Power Systems and Equipment - Voltage Ratings (60 Hertz), 1995 [142] G. Burchi, C. Lazaroiu, N. Golovanov, M. Roscia, Estimation of Voltage Unbalance in Power

Systems Supplying High Speed Railway, Journal of Electrical Power Quality and Utilisation, Vol. 11, No. 2, pp. 113-119, 2005

[143] N. Golovanov, M. Roscia, D. Zaninelli, Voltage Unbalance Vulnerability Areas in Power Systems Supplying High Speed Railway, IEEE Power Engineering Society General Meeting, 16 June, 2005, pp. 2509-2514

[144] IEEE 241 – IEEE Recommended Practice for Electric Power Systems in Commercial Buildings, 1990

[145] System Standards, National Electricity Code (NEC) Australia, Version 1.0, Amendment 9.0, S5.1a.7, October 2004

REPORT

of 158

[146] P. Paranavithana, S. Perera, D. Sutanto, R. Koch, A Systematic Approach Towards Evaluating Voltage Unbalance Problem in Interconnected Sub-transmission Networks: Separation of Contribution by Lines, Loads And Mitigation, 13th International Conference on Harmonics and Quality of Power, 28 September-1 October 2008, pp. 1-6

[147] T.E. Seiphetlho, A.P.J. Rens, On the Assessment of Voltage Unbalance, 4th International Conference Harmonics and Quality of Power (ICHQP 2010), 26-29 September 2010, pp. 1-6

[148] N.G. Hingorani, L. Gyugyi, Understanding FACTS: Concepts and Technology of Flexible AC Transmission Systems. Wiley-IEEE Press, 1999

[149] M.T. Bina, M.D. Eskandari, Consequence of Unbalance Supplying Condition on a Distibution Static Compensator, IEEE 35th Annual Power Electronics Specialists Conference, 20-25 June 2004, pp. 3900-3904

[150] R.R. Oliveira, C.A.L. Rocha, A.C. Delaiba, J.C. Oliveira, F.N. Belchior, Three-Phase Synchronous Generators Performance with Unbalanced and Non-Linear Loading Analytical and Experimental Analysis, IEEE International Symposium on Industrial Electronics, 9-13 July 2006, pp. 1744-1749

[151] A. von Jouanne, B.B. Banerjee, Voltage Unbalance: Power Quality Issues, Related Standards and Mitigation Techniques, Palo Alto Electric Power Research Institute, May 2000

[152] D.R. Smith, H.R. Braunstein, J.D. Borst, Voltage Unbalance in 3- and 4-wire Delta Secondary Systems, IEEE Transaction of Power Delivery, Vol. 3, No. 2, April 1988, pp. 733-741

[153] S. Weckx, J. Driesen, Load Balancing With EV Chargers and PV Inverters in Unbalanced Distribution Grids,“ IEEE Transaction on Sustainable Energy, Vol. 6, No. 2, April 2015, pp. 635-643

[154] OFGEM, Modificaion to the Grid Code: Voltage Unbalance (GC0088), January 2016 [Online] https://www.ofgem.gov.uk/publications-and-updates/modification-grid-code-voltage-unbalance-gc0088

[155] V. Matta, G. Kumar, Unbalance and Voltage Fluctuation Study on AC Traction System, 2014 International Conference of Computation of Power, Energy, Information and Communication (ICCPEIC), 16-17 April 2014, pp. 315-320

[156] J. Rossman, T. Laughner, A. Murphy, D. Marler, G. Kobet, Transmission Voltage Unbalance Evaluation, Future of Instrumentation International Workshop (FIIW) 8-9 October 2012, pp. 1-4

[157] P. Sutherland, M. Waclawiak, M. McGranaghan, System Impacts Evaluation of a Single-Phase Traction Load on a 110-kV Transmission System, IEEE Transactions on Power Delivery, Vol. 21, No. 2, April 2006, pp. 837-844

[158] S.L. Chen, R.J. Li, P.H. His, Traction System Unbalance Problem – Analysis Methodologies, IEEE Transactions on Power Delivery, Vol. 19, No. 4, October 2004, pp. 1877-1883

[159] H. Hu, Z. He, K. Wang, X. Ma, S. Gao, Power Quality Impact Assessment for High-Speed Railway Associated with High-Speed Trains Using Train Timetable – Part II: Verifications, Estimations and Applications,” IEEE Transactions on Power Delivery, Vol. 31, No. 4, August 2016

[160] H. Lee, G. Kim, C. Lee, S. Oh, S. Park, Analysis of Voltage Unbalance on Korean Railway System, Electric Power System Research, Paper No. 414, 2005

[161] N. Gunavardhini, M. Chandrasekaran, C. Sharmeela, K. Manohar, A Case Study on Power Quality Issues in the Indian Railway Traction Sub-Station, 7th International Conference on Intelligent Systems and Control (ISCO), 4-5 January 2013, pp. 7-12

REPORT

of 158

[162] G.J. Wakileh, Power Systems Harmonics: Fundamentals, Analysis and Filter Design, Springer, 2001

[163] P. Salmeron, R.S. Herrara, A.P. Valles, J. Prieto, New Distortion and Unbalance Indices Based on Power Quality Analyzer Measurements, IEEE Transaction on Power Delivery, Vol. 24, No. 2, April 2009, pp. 501-507

[164] E. Acha, M. Madrigal, Power Systems Harmonics: Computer Modeling and Analysis, Wiley, 2001

[165] W. Wiechowski, Harmonics in Transmission Power Systems; PhD Thesis, Faculty of Engineering, Science and Medicine, Aalborg University, 2006

[166] A. Pană, A. Băloi, F. Molnar-Matei, Contributions on the Harmonic Analysis of a Transmission Overhead Line - A Case Study, 16th International Scientific Conference on Electric Power Engineering (EPE), 20-22 May 2015, pp. 122-126

[167] D. Sharon, J.-C. Montaño, A. López, M. Castilla, D. Borrás, Power Quality Factor for Networks Supplying Unbalanced Nonlinear Loads, IEEE Transaction on Instrumentation and Measurement, Vol. 57 No.6, pp. June 2008, pp. 1268-1274

[168] S. Zhang, S. Jiang, X. Lu, B. Ge, F.Z. Peng, Resonance Issues and Damping Techniques for Grid-Connected Inverters With Long Transmission Cable, IEEE Transaction on Power Electronics, Vol. 29, No. 1 January 2014, pp. 110-120

[169] IEEE 519 – IEEE Recommended Practice and Requirements for Harmonic Control in Electric Power Systems, 2014

[170] IEC TR 61000-3-13 – Electromagnetic Compatibility (EMC) - Part 3-13: Limits - Assessment of Emission Limits for the Connection of Unbalanced Installations to MV, HV and EHV Power Systems, 2008

[171] IEC 61000-2-2 – Electromagnetic Compatibility (EMC) - Part 2-2: Environment - Compatibility Levels for Low-Frequency Conducted Disturbances and Signalling in Public Low-Voltage Power Supply Systems, 2002

[172] ER G5/4 Planning Levels for Harmonic Voltage Distortion and the Connection of Non-Linear Equipment to Transmission Systems and Distribution Networks in the United Kingdom, Energy Networks Association, 2nd edition, 2008

[173] D.A. Bradley, P.J. Morfee, L.A. Wilson, The New Zealand Harmonic Legislation; IEE Proceedings B-Electric Power Applications, Vol. 132, No.4, July 1985

[174] C. Lascu, L. Asiminoaei, I. Boldea, F. Blaabjerg, High Performance Current Controller for Selective Harmonic Compensation in Active Power Filters, IEEE Transaction on Power Electronics, Vol. 22, No. 5, September 2007, pp. 1826-1835

[175] V.E. Wagner, J.C. Balda, T.M. Barnes, A.E. Emanuel, R.J. Ferraro, Effects of Harmonics on Equipment, IEEE Transactions on Power Delivery, Vol. 8, No. 2, April 1993, pp. 672-680

[176] M. Kiani, W. Lee, Effects of Voltage Unbalance and System Harmonics on the Performance of Doubly Fed Induction Wind Generators, IEEE Transaction on Industry Applications, Vol. 46, No. 2, March-April 2010, pp. 1-7

[177] G. Lemieux, Power System Harmonic Resonance - A Documented Case, IEEE Transaction on Industry Applications, Vol. 26, No.3, May/June 1990, pp. 483-488

[178] IEEE C57.110 – Recommended Practice for Establishing liquid-Filled and Dry-Type Power and Distribution Transformer Capability When Supplying Nonsinusoidal Load Currents, 2008

[179] J. Arrillaga, D.A. Bradley, P.S. Bodger, Power System Harmonics, Wiley, 1985

REPORT

of 158

[180] A. Domijan Jr, E. Embriz, G. Lamer, C. Stiles, C.W. Williams, Watthour Meter Accuracy Under Controlled Unbalanced Harmonic Voltage and Current Conditions, IEEE Transactions on Power Delivery, Vol. 11, No. 1, January 1996, pp. 64-72

[181] S.B. Dewan, F.P. Dawson, Internal Model Current Control of VSC-based Active Power Filters, 30th Annual IEEE Power Electronics Specialists Conference (PESC), 1 July 1999, pp. 155-160

[182] C. Wessels, J. Dannehl, F.W. Wuchs, Active Damping of LCL-Filter Resonance Based on Virtual Resistor for PWM Rectifiers – Stability Analysis with Different Filter Parameters, IEEE Power Electronics Specialists Conference (PESC), 15-19 June 2008, pp. 3532-3538

[183] J. Dannehl, F.W. Fuchs, S. Hansen, P.B. Thøgersen, Investigation of Active Damping Approaches for PI-Based Current Control of Grid-Connected Pulse Width Modulation Converters With LCL Filters, IEEE Transactions on Industry Applications, Vol. 46, No. 4, July-August 2010, pp. 1509-1517

[184] G.W. Chang, S.Y. Chu, H.L. Wang, Sensitivity Based Approach for Passive Harmonic Filter Planning in a Power System, IEEE Power Engineering Society Winter Meeting, 27-31 January 2002, pp. 937-940

[185] E.B. Makram, E.V. Subramaniam, A.A. Girgis, R. Catoe, Harmonic Filter Design Using Actual Recorded Data, IEEE Transactions on Industry Applications, Vol. 29, No. 6, November/December 1993, pp. 1176-1183

[186] Y.H. Song, A.T Johns, Flexible AC Transmission Systems (FACTS), The Institution of Electrical Engineers, 1999

[187] R.W. Menzies, Y. Zhuang, Advanced Static Compensation Using a Multilevel GTO Thyristor Inverter, IEEE Transactions on Power Delivery, Vol. 10, No. 2, April 1995, pp. 732-738

[188] L. Padmavathi, P.A. Janakiraman, Reduction of Voltage Harmonics in Three-Phase Inverters Using Discrete Observers, Joint International Conference on Power Electronics, Drives and Energy Systems (PEDES), 20-23 December 2010, pp. 1-5

[189] K.B. Mohanty, R.N. Mishra, Power Quality Improvement Through Harmonics Mitigation Using Shunt Hybrid Power Filter, 9th International Conference on Industrial and Information Systems (ICIIS), 15-17 December 2014, pp. 1-6

[190] S.W. Kang, K.H. Kim, Sliding Mode Harmonic Compensation Strategy for Power Quality Improvement of A Grid-Connected Inverter Under Distorted Grid Condition, IET Power Electronics, Vol. 8, No. 8, August 2015, pp. 1461-1472

[191] Eirgrid, Power Quality Requirements for Connection to the Transmission System, September 2015, [Online] http://www.eirgridgroup.com/site-files/library/EirGrid/Note-on-Harmonics-Power-Quality-Requirements-for-Connection-to-Transmission-System.pdf

[192] M. Val Escudero, S. Murray, J. Ging, B. Kelly, M. Norton, Harmonic Analysis of Wind Farm Clusters Using HV-AC Underground Cables in the Irish Transmission Network, CIGRE Conference, Paris, 2014

[193] Z. Emin, F. Ghassemi, J.J. Price, Harmonic Performance Requirements of An HVDC Connection: Network Owner Perspective, 9th IET International Conference on AC and DC Power Transmission (ACDC), 19-21 October 2010

[194] L. Kocewiak, Harmonic Analysis of Offshore Wind Farms With Full Converter Wind Turbines, 8th International Conference on Large-Scale Integration of Wind Power into Power Systems, 2009

[195] A. Shafiu, A. Hernandez, F. Schettler, E. Jorgensen, Harmonics Studies for OffShore WindFarms, 9th IET International Conference on AC and DC Power Transmission (ACDC), 19-21 October 2010

REPORT

of 158

[196] A.J. Hernandez M.S. Wijesinghe, A. Shafiu, Hamonic Amplification of the 576 MW Gwynt-y-Môr Offshore Wind Power Plant, 11th International Workshop on Large-Scale Integration of Wind Power into Power Systems, Lisbon, 13-15 November 2012

[197] Z. Emin, F. Fernancez, M. Poeller, G.E. Williamson, Harmonic Distortion Specification and Compliance of An Offshore Wind Generation, 10th IET International Conference on AC and DC Power Transmission (ACDC 2012), 4-5 December 2012, pp. 1-6

[198] J.H.R. Enslin, J. Knijp, C.P.J. Jansen, J.H. Schuld, Impact of Reactive Power Compensation Equipment on the Harmonic Impedance of High Voltage Networks, IEEE Power Tech Conference, Bologna, 23-29 June 2003

[199] J.H.R. Enslin, Y. Hu, R.A. Wakefield, System Considerations and Impacts of AC Cable Networks on Weak High Voltage Transmission Networks. 2005/2006 IEEE PES Transmission and Distribution Conference and Exhibition, 21-24 May 2006, pp.1-5

[200] F.F. da Silva, C.L. Bak, P.B. Holst, Study of Harmonics in Cable-based Transmission Networks, CIGRE C4-108, 2012

[201] M.H.J. Bollen, S.M. Gargari, Harmonic Resonances due to Transmission Cables, CIGRE Conference, Belgium 2014

[202] Y. Fillion, S. Deschanvres, Background Harmonic Amplifications Within Offshore Wind Farm Connection Projects, International Conference on Power Systems Transients (IPST2015), Cavtat, 15-18 June 2015

[203] C.F. Jensen, L. Kocewiak, Z. Emin, Amplification of Harmonic Background Distortion in Wind Power Plants with Long High Voltage Connections, CIGRE C4-112, 2016

[204] R.P.D. Ross, M.P. De Carli, P.F. Riberiro, Harmonic Distortion Assessment Related to the Connection of Wind Parks to the Brazilian Transmission Grid, CIGRE C4-101, 2016

[205] K.L. Koo, Z. Emin, Challenges in Harmonic Assessments of Non-linear Load Connections, CIGRE-C4-111, 2016

[206] V. Myagkov, L. Petersen, S.B. Laza, F. Iov, L.H. Kocewiak, Parametric Variation for Detailed Model of External Grid in Offshore Wind Farms, 13th International Workshop on Large-Scale Integration of Wind Power into Power Systems as well as on Transmission Networks for Offshore Wind Power Plants (WIW2014), 11-13 November 2014

[207] D. Tong, V.G. Nikolaenko, N. Ginbey I. Lau, Harmonic Propagation in Transmission System With Multiple Capacitor Installations, PowerCon 2000 International Conference on Power System Technology, February 2000

[208] N. Kanao, M. Yamashita, H. Yanagida, M. Mizukami, Y. Hayashi, J. Matsuki, Power System Harmonic Analysis Using State-Estimation Method for Japanese Field Data, IEEE Transactions On Power Delivery, Vol. 20, No. 2, April 2005, pp. 970-977

[209] R.J.R. Gomes, D.O.C. Brasil, J. R. de Medeiros, Power Quality Management Issues Over the Brazilian Transmission System, 10th International Conference on Harmonics and Quality of Power, 6-9 October 2002, pp. 27-32

[210] L. Kocewiak, Harmonics In Large Offshore Wind farms, Department of Energy Technology, Aalborg University, 2012

[211] C. Buchhagen, M. Greve, A. Menze, J. Jung, Harmonic Stability – Practical Experience of a TSO, Wind Integration Workshop, Vienna, 2016

[212] M.K. Bakhshizadeh, J. Hjerrild, L. Kocewiak, Harmonic Modelling, Propagation and Mitigation for Large Wind Power Plants Connected Via Long HVAC cables: Review and Outlook of Current Research, 2016 IEEE International Energy Conference (ENERGYCON), 4-8 April 2016, pp. 1–5

REPORT

of 158

[213] L.H. Kocewiak, J. Hjerrild, C.L. Bak, Wind Turbine Control Impact on Stability of Wind Farms Based on Real-Life Systems Analysis, EWEA 2012 Europe's Primier Wind Energy Even, Copenhagen, 16-19 April 2012

[214] M. Parniani, M.R. Iravani, Voltage Control Stability and Dynamic Interaction Phenomena of Static VAr Compensators, IEEE Transaction on Power Systems, Vol. 10, No. 3, August 1995, pp. 1592-1597

[215] R.M. Mathur, R.K. Varma, Thyristor-Based FACTS Controllers for Electrical Transmission Systems, 1st edition, Wiley-IEEE Press, 2002

[216] J. Lv, P. Dong, G. Shi, X. Cai, H. Rao, J. Chen, Subsynchronous Oscillation of Large DFIG-based Wind Farms Integration Through MMC-based HVDC, 2014 International Conference on Power System Technology (POWERCON), 20-22 October 2014, pp. 2401-2408

[217] J. Lyu, X. Cai, Impact of Controller Parameters on Stability of MMC-based HVDC Systems for Offshore Wind Farms, International Conference on Renewable Power Generation (RPG 2015), 17-18 October 2015, pp. 1-6

[218] H.T. Ma, P.B. Brogan, K.H. Jensen, R.J. Nelson, Sub-Synchronous Control Interaction Studies Between Full-Converter Wind Turbines and Series-Compensated AC Transmission Lines, 2012 IEEE Power and Energy Society General Meeting, 22-26 July 2012, pp. 1-5

[219] J. Sun, Impedance-Based Stability Criterion for Grid-Connected Inverters, IEEE Transansaction On Power Electronics, Vol. 26, No. 11, November 2011, pp. 3075-3078

[220] M. Cespedes, J. Sun, Impedance Modeling and Analysis of Grid-Connected Voltage-Source Converters, IEEE Transansaction on Power Electronics, Vol. 29, No. 3, March 2014, pp. 1254-1261

[221] IEC TR 61869-103 – Instrument Transformers - The Use of Instrument Transformers for Power Quality Measurement, 2012

[222] H. Seljeseth, E.A. Saethre, T. Ohnstad, I. Lien, Voltage Transformer Frequency Response: Measuring Harmonics in Norwegian 300 kV and 132 kV Power Systems, 8th International Conference on Harmonics and Quality of Power, 14-16 October 1998, pp. 820-824

[223] J. Kilter, J. Meyer, B. Howe, F. Zavoda, L. Tenti, J.V. Milanović, M. Bollen, P.F. Riberio, P. Doyle, J.M. Romero Gordon, Current Practice and Future Challenge for Power Quality Monitoring – CIGRE WG C4.112 Perspective, 15th International Conference on Harmonics and Quality of Power (ICHQP’12) 17-20 June 2012, pp. 390-397

[224] J. Meyer, J. Kilter, B. Howe, F. Zavoda, L. Tenti, J.M. Romero Gordon, J.V. Milanović, Contemporary and Future Aspects of Cost Effective Power Quality Monitoring – Position Paper of CIGRE WG C4.112, Electric Power Quality and Supply Reliability Conference (PQ), 11-13 June 2012

[225] IEC 61000-2-12 – Electromagnetic Compatibility (EMC) - Part 2-12: Environment - Compatibility Levels for Low-Frequency Conducted Disturbances and Signalling in Public Medium-Voltage Power Supply Systems, 2003

[226] R. Stiegler, J. Meyer, J. Kilter, S. Konzelmann, Assessment of Voltage Instrument Transformers Accuracy for Harmonic Measurements in Transmission Systems, International Conference on Harmonics and Quality of Power (ICHQP2016) 16-19 October 2016

[227] J. Meyer, R. Stiegler, J, Kilter, Accuracy of Voltage Instrument Transformers for Harmonic Measurements in Elering`s 330 kV Transmission Network, 10th Electric Power Quality and Supply Reliability Conference (PQ) 29-31 August 2016

[228] R.A. Barr, V.J. Gosbel, I. McMichael, A New SAIFI Based Voltage sag Index, 13th International Conference on Harmonics and Quality of Power (ICHQP), 28 September-1 October 2008, pp. 1-5

REPORT

of 158

[229] J. Lamoree, D. Mueller, P. Vinett, W. Jones, M. Samotyj, Voltage Sag Analysis Case Studies, IEEE Transactions on Industry Applications, Vol. 30, No. 4, July/August 1994, pp. 1083-1089

[230] J.D. La Ree, V. Centeno, J.S. Thorp, A.G. Phadke, Synchronized Phasor Measurement Applications in Power Systems, IEEE Transactions on Smart Grid, Vol. 1, No. 1, June 2010, pp. 20-27

[231] A. Monti, C. Muscas, F. Ponci, Phasor Measurement Units and Wide Area Monitoring Systems, Elsevier, 2016

[232] IEEE C37.118.1 – IEEE Standard for Synchrophasor Measurements for Power Systems, 2011

[233] IEEE C37.118 – IEEE Standard for Synchrophasor Measurements for Power Systems, 2005 [234] IEEE C37.118.2 – IEEE Standard for Synchrophasor Data Transfer for Power Systems, 2011 [235] IEEE C37.118.1a – IEEE Standard for Synchrophasor Measurements for Power Systems --

Amendment 1: Modification of Selected Performance Requirements, 2014

critical pq phenomena and sources of pq disturbances in pe ...€¦ · name company author/s: j....

Documents