Smart Distributed Control of Power Systems

Deepak Divan Ron Harley School of Electrical Engineering

Georgia Institute of Technology, Atlanta, GA, USA

777 Atlantic Drive deepak.divan@ece.gatech.eduAtlanta, GA

Confidential

US Power Grid: Problems (Opportunities)

• Grid is seeing increasing congestion and degraded reliability

• Uncontrolled power flow is a major issue for transmission & distribution

• Building new transmission/distribution lines is no longer a simple process

• First line to reach thermal rating limits system transfer capacity, even as neighboring lines are under-utilized

• Real situation is worse as reliability has to be ensured with (N-X) contingencies

• Lack of visibility and control leads to conservative operation resulting in significant under-utilization of assets

• Possible cascading failures under contingency conditions

• Reliability, load growth, will be major drivers for new investments

Power flow path from Wisconsin to TVA*

*Courtesy: Tom Overbye, UIUC

j16 Ω

j24 Ω

138kV∠0°

138kV∠7.75°675A

450A

Confidential

Proposed Vision for Electricity Infrastructure Research

The Vision:A fully controllable electricity delivery infrastructure that is cost-effective, provides flexible interconnection of sources and loads, is robust and reliable, is self-healing and degrades gracefully under contingencies, enables operation of market forces, never degrades system reliability to worse than current levels, and is scalable to meet societal requirements of growth and sustainability.

How Will This Be Implemented?• Power delivery assets will have to be infused with intelligence, communications and control capability, using distributed mass-manufactured components for low-cost and high-reliability through redundancy.

• Power grid will be massively-networked for high-reliability, and will allow autonomous local control with local information while ensuring global system level optimization using feedback.

• New techniques for system level monitoring, visualization, protection and control in the presence of massive data streams from widely distributed assets will provide visibility and actionable information to allow operators to avert catastrophic events.

• This approach can dramatically reduce the cost and accelerate implementation of a Smart Grid, enabling system transforming applications such as PHEVs, demand side management, and support proliferation of distributed renewable resources that foster sustainability.

Confidential 4

Thin AC Converters – Making Existing Grid Assets Controllable

• Proposed approach provides static and dynamic control of value of existing grid assets, such as shunt VAR capacitors and transformers

• Allows operation in ‘dispatch’ mode or under local control to regulate voltage or respond to faults

• Direct AC conversion using semiconductor switches, small LC filters, switchgear and minimal energy storage allowing compact size, low volume and low cost

• Imposes minimal additional stresses on the asset allowing use in retrofit applications.

• ‘Fail Normal’ mode of operation, where failure of the thin converter automatically restores normal function of the asset on the grid

• Multi-Level AC Converter allows transformer- less operation in medium voltage applications

Confidential

Thin AC Converters

THIN AC CONVERTERS

MULTI-LEVEL DIRECT AC CONVERTERS

VIRTUAL QUADRATURE SOURCES

Possible ApplicationsPossible Applications

LTC TransformersShunt VAR Capacitors Transmission Lines

Inverter-less STATCOMs Controllable Network Transformers

Smart Wires

http://www.tradesurinc.com/productshttp://www.vatransformer.com

Confidential

Distributed FACTS for Power Flow Control• Distributed FACTS suggested by Divan – Smart Wires

– Provide the functionality of FACTS at lower cost and high reliability

– Series VAR injection controls effective line impedance & real power flow– Large number of modules float electrically and mechanically on the line– Can be incrementally deployed to provide controllable power flow– Standard low-cost mass-manufactured modules– Redundancy gives high reliability and availability– Phase I supported by TVA, Con Ed, DOE and others

D-FACTS

( )δsin2112

sXVVP =

Confidential

Active Smart Wires

• Distributed Static Series Compensator (DSSC)– Active solution employing a synchronous voltage source inverter– Each module rated for 5 KVA (capable of injecting ± 4.6 V @ 1000 A)– Communication interface is required to realize the bi-directional control– Can be made larger for distribution applications (one per line)

Power supply

Main transformer

Filter

PWM Inverter

DC Capacitor

V

ControlsCommunication

Module

Current feedback

Line Current

Confidential

Distributed Series Reactance – Passive Smart Wires

• Simplest implementation of DSI, with inductive impedance injection (Current Limiting Conductor or CLiC) – functions as a current limiting system

• As current in a line approaches the thermal limit, CLiC modules incrementally turn on, diverting current to other under-utilized lines

• Each module is triggered at a predefined set point to reflect a gradual increase in line impedance

• No communication required and the devices operate autonomously

Control

Power Line

Transformer

S1

XM

Power Supply

SM

I0 Ithermal

Line Current (ILine)

N Xm

Confidential

Increase in System Capacity With DSR Modules

0 0.5 1 1.5 2 2.50

0.25

0.5

0.75

1

Time (sec)

Line

Cur

rent

(KA)

Profile of Line 2 Current

DSR Active

548.7 A

750 A

624.2A

GeneratorTaken Off

Simplified Four Bus System

Contingency Condition: Generator Outage

20 40 60 80 100 120 140 160

20

40

60

80

100

120

140

160

180

Load2 (MW)

Load

1 (MW

)

Line2 Overload

Line5 Overload

Line2 Overload

Line2, Line5Overload

Line1, Line2,Line5 Overload

Line2, Line4 Overload

Line1, Line2, Line4, Line5 Overload

A

A'

B

B'

C'

C

System Capacity with DSI/DSR ModulesBlue: Normal, Red: DSR, Green:DSI

Confidential

Increase in Network Utilization

IEEE 39 Bus System

Network Performance With CLiC

Line

2_3

Line

6_5

Line

6_7

Line

9_39

Line

10_1

3

Line

12_1

1

Line

13_1

4

Line

19_1

6

Line

22_2

1

Line

23_2

4

Line

25_2

6

Line

26_2

7

Line

29_2

6

Line

29_2

8

Power Lines

Line

Cur

rent

s (%

Ther

mal

Lim

it)

0

20

40

60

80

100

Line currents with CLiCLine currents without CLiC

Increase in utilization from 59% to 93.3% for 14 lines

G1

G8

G10

230

1

G2

G3

G9

G4G5

G6

G7

39

9

8 7

5

4

3

18

37

25

17

26 28 29

38

24

27

15

14

12

13

1011

32 34

20

19

21 2235

23

36

16

6

31

22.76 MVAR

19.52 MVAR

18.77 MVAR

14.75 MVAR12.64 MVAR

3.52 MVAR

12.86 MVAR

12.04 MVAR

9.15 MVAR

• Increase in Transfer Capacity from 1904 MWs to 2542 MWs (congested corridors and the required MVARs are shown by red lines)

• With (N-1) contingency, capacity is increased from 1469 MW to 2300 MW without building additional lines

• Would require 9 additional lines to realize capacity increase

Confidential

DSR Prototype

Complete module with the casing • Electrical– Operating Level : 161 KV, 1,000 A– ACSR Conductor: Drake (795 Kcmil)– Injection: 10 kVA

• Mechanical– Target weight per module: 120 lb– Packaging to avoid corona discharge,

and other mechanical, thermal and environmental issues

• Simple low-cost design suitable for mass manufacturing

• Suitable for distribution and transmission applications

Confidential

Validation at High Voltage and Current

Corona inception: 125 kV Extinction: 123 kV

Photograph correspond to 166 kV

Confidential

Smart Wires in a Smart Grid

• It is proposed that the use of distributed solutions based on low-power power electronics can allow utilities to move towards dynamically controllable meshed grids, significantly enhancing grid reliability, capacity and utilization. This can enable:

• Improved reliability without having to build new lines

• Improved dynamic coordination between regions

• Reduction in dynamic capacity reserve for generators

• Possibility of moving power along a predetermined contract path

• Can be applied at the transmission, sub-transmission and distribution levels.

• Can be layered incrementally onto the existing infrastructure as desired, and will not degrade the inherent reliability of the existing system.

• Makes the grid self-healing, automatically maintaining safe operating levels even in the face of contingencies.• Funding such investments on the basis of congestion relief is problematic in regulated

environments, new mechanisms may have to be found.