THE ROLE OF SYNCHROPHASORS IN ENSURING ...shukla/tps/Session2/Ilic...THE ROLE OF SYNCHROPHASORS IN...

THE ROLE OF SYNCHROPHASORS IN ENSURING RELIABLE POWER SYSTEM CONTROL

Marija Ilic

Professor, Carnegie Mellon University

milic@ece.cmu.edu

Thorp-Phadke Symposium May 2013

Potential of Measurements, Communications and Control

2

PMU Control

Constrained Line

Line-to-Ground Clearance

Transfer Capacity in Real Time

DLR

Combined state estimation and optimization problem

3

Power System Control

State Estimation

Measurements (2 sec) + PMU (millisecond)

Power Flow Analysis

Optimal Power Flow

Yang Weng

• System Load Curve

0 5 10 15 20 25120

140

160

180

200

220Every 10 min Real Time Load of NYISO in Jan 23, 2010

Time (hours)

Real P

ow

er

Load P

(p.u

.)

0 5 10

180

200

10 min Real Time Load of NYISO in Jan 23, 2010

Time (min)Re

al P

ow

er

Lo

ad

P(p

.u.)

Forecasted Load

Actual Load with Disturbance

Lower Bound

Upper Bound

4

From preventive to corrective management of future electric energy systems

5

Predictable load and the disturbance

0 20 40 60 80 100 120 140 160 180

20.3

20.4

20.5

20.6

20.7

20.8

20.9

21

21.1

Time (mins)

Lo

ad

Re

active

Po

we

r E

vo

lutio

n (

p.u

.)

NYISO August 2006 Load Data for 3 Hours, Power Factor = 0.8

0 20 40 60 80 100 120 140 160 18015

16

17

18

19

20

21

22

23

24

Time (mins)

Lo

ad

Re

active

Po

we

r E

vo

lutio

n (

p.u

.)

NYISO August 2006 Load Reactive Power Data for 3 Hours, Measurement Frequency = 0.50 Hz,Power Factor = 0.8

Pre-planed Load Value

Real Load Evolution, with 0.5Hz Sampling

5

PMUs-enabled grid for efficient and reliable scheduling to balance predictable load

• PMUs and SCADA help more accurate state estimate of line flows, voltages and real/reactive power demand

• AC OPF utilizes accurate system inputs and computes settings for controllable grid, generation and demand equipment to help manage the system reliably and efficiently

• Adjustments done every 15 minutes

• Model-predictive generation and demand dispatch to manage ramp rates

Today static dispatch for scheduling

5

Model-predictive scheduling with wind generation---slow time scale

• 20% / 50% penetration to the system

6

Le Xie

9

Conventional

cost over 1 year *

Proposed

cost over the

year

Difference Relative Saving

$ 129.74 Million $ 119.62 Million $ 10.12

Million

7.8%

*: load data from New York Independent System Operator, available online at http://www.nyiso.com/public/market_data/load_data.jsp

0 50 100 150 200 250 3000

50

100

150

Coal Unit 2 (Expensive) Generation

Time Steps (10 minutes interval)

MW

50 60 70 80 90 1000

50

100

150

Coal Unit 2 Generation: Zoomed In

Time Steps (10 minutes interval)

MW

Conventional Dispatch

Centralized Predictive Dispatch

Distributed Predictive Dispatch

Conventional Dispatch

Centralized Predictive Dispatch

Distributed Predictive Dispatch

BOTH EFFICIENCY AND RELIABILITY MET

Model-predictive dispatch with price-responsive demand

8

• Elastic demand that responds to time-varying prices

J.Y. Joo kWh

$

9

Model-predictive dispatch with EVs

10

• Interchange supply / demand mode by time-varying prices

NiklasRotering

Optimal Control of Plug-in-Electric Vehicles: Fast vs. Smart

14

Large-Scale Nonlinear Grid Optimization for Corrective Actions

Imports can be increased by the following:

Optimal generator voltages

Optimal settings of grid equipment (CBs, OLTCs, PARs, DC lines, SVCs)

Studies have shown 20-25% economic efficiency by implementing corrective (not preventive!) actions

Optimal selection of new equipment (type, size, location)

..Remembering the summer of 1983 when Arun and I wrote our reactive power distribution factors currently used in PJM to explain how reactive power affects voltages.. US National Grid Studies still do not go beyond real power distribution factors.

16

On-line resource management can prevent blackouts….

17

PMU-Based Robust Control –fast time scale (automated)

Zhijian Liu

P

P

• Automated Voltage Control (AVC) and Automated Flow Control (AFC) – Design Best Locations

of PMUs – Design Feedback

Control Gains

P

P

18

Building on the long-ago joint work with Jim

0 10 20 30 40 50 60 70 80 90 1000

0.05

0.1

0.15

0.2

0.25

Time (sec)

Syste

m W

ors

t V

oltage D

evia

tion (

p.u

.)

Automatic Voltage Control for ONE Pilot Point Control Case

No Control

One Pilot Point Control

5% Reliability Criteria

Pilot Point: Bus 76663

• Robust AVC Illustration in NPCC System

19

All load buses are

Monitored

0 10 20 30 40 50 60 70 80 90 1000

0.05

0.1

0.15

0.2

0.25

Time (sec)

Syste

m W

ors

t V

oltage D

evia

tion (

p.u

.)

Automatic Voltage Control for Unlimited Information Control Case

No Control

Full Information Control

5% Reliability Criteria

20

0 20 40 60 80 1000

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Time (sec)

Syste

m W

ors

t F

low

Devia

tion (

p.u

.)

Automatic Flow Control for ONE Pilot Point Control Case

No Control

One Pilot Point Control

5% Reliability Criteria

Pilot Point: Bus 75403

AVC for the NPCC with PMUs

21

Simulations to show the worst voltage deviations

in response to the reactive power load

fluctuations (3 hours)

0 20 40 60 80 100 120 140 160 1800

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Time (min)

Syste

m W

ors

t V

olta

ge

De

via

tio

n (

p.u

.)

1 Pilot Point Secondary Voltage Control with Measurement Frequency = 0.50 Hz,Power Factor = 0.8

No Control

One Pilot Point per Area

5% Criteria

0 20 40 60 80 100 120 140 160 1800

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Time (min)

Syste

m W

ors

t V

olta

ge

De

via

tio

n (

p.u

.)

2 Pilot Points Secondary Voltage Control with Measurement Frequency = 0.50 Hz,Power Factor = 0.8

No Control

Two Pilot Point per Area

5% Criteria

2 Pilot Points Control Performs Better Than 1 Pilot Point!

21

AFC Using PMUs- NPCC System

0 5 10 15 20 25 30 350

0.02

0.04

0.06

0.08

0.1

0.12

Location of Disturbance

Lin

es R

ea

l P

ow

er

Flo

w D

evia

tio

n (

p.u

.)

System Worst Line Real Power Flow Deviation Under Disturbance From Different Buses

No Control

One Pilot Point per Area

0 20 40 60 80 100 1200

0.02

0.04

0.06

0.08

0.1

0.12

Line NumberL

ine

s R

ea

l P

ow

er

Flo

w D

evia

tio

n (

p.u

.)

Worst Real Power Flow Deviation of Each Line Under Disturbance From Different Buses

No Control

One Pilot Point per Area

Control real power disturbance

….Versions of AVC implemented in EdF Italy, China.. It may be time to consider by the US utilities

Liu and Ilic, “Toward PMU-Based Robust Automatic Voltage Control (AVC) and Automatic Flow Control (AFC),” IEEE PES, 2008

23

Pushing the limits to what is doable –transient stabilization ( Selkrik fault with conventional controller)

24

Voltage response with conventional controllers-base case Selkrik fault

Concepts proposed to manage fast phenomena using

non-interacting control; no need for fast communications (Chapman,J; Allen, E)

25

Bus voltages with new controllers

26

Rotor angle response with local nonlinear controllers--an early example of flat control design

Issues with standards for preventing SSR-related safety problems

27

Nonlinear control for storage devices (FACTS,flywheels)

[1] The test system: J. W. Chapman, “Power System Control for Large Disturbance Stability: Security, Robustness and Transient Energy”, Ph.D. Thesis: Massachusetts Institute of

Technology, 1996.

[2] Linear controller: L. Angquist, C. Gama, “Damping Algorithm Based on Phasor Estimation”, IEEE Power Engineering Society Winter Meeting, 2001

[3] Nonlinear controller: M. Ghandhari, G. Andersson, I. Hiskens, “Control Lyapunov Function for Controllable Series Devices”, IEEE Transactions on Power Systems, 2001, vol. 16, no. 4,

pp. 689-694

Linear PI power controller[2]

Nonlinear Lyapunov controller[3]

No controller on TCSC

Use of interaction variables in strongly coupled systems

Interaction variable choice 1:

Interaction variable choice 2:

Must proceed carefully… • The very real danger of new complexity.

• Technical problems at various time scales lend themselves to the fundamentally different specifications for on-line data

• No longer possible to separate measurements, communications and control specifications

• Major open question: WHAT CAN BE DONE IN A DISTRIBUTED WAY AND WHAT MUST HAVE FAST COMMUNICATIONS –Jim—I think this is doable when the grid has localized response…

The persistent challenge: SE to support on-line scheduling implementation (Yang Wang)

Current Power System State Estimation Problems

Nonlinearity Non-convexity

Historical Data are not really used

New devices (i.e. PMU) placement problem

Convexification Semi-definite Programming

Graph-based distributed SDP

SE

Computational Burden

Non-parametric Static state Estimation

Parametric Dynamic state

Estimation

Information Theory based algorithm for

State Estimation

Parallel Computing Algorithm

Load serving entities (LSEs)

Backbone Power Grid

and its

Local Networks (LSEs)

LSE LSE

LSE

LSE

Information flow: MISO

Local Distribution Network (Radio Network)

Multilayer Information for State Estimation

PQ Diesel PQ Wind PQ PQ

Distributed SE

Computation

Physical Layer Online Diagram Information Layer Diagram

LSE

State information exchange

State information

Exchange on the

boundary nodes

Local State Estimation (LSE)

Backbone

Distributed SE

Computation

LSE LSE

Local serving entities (LSEs)

LSE

LSE LSE

LSE

LSE

LSE

LSE

LSE

LSE

LSE LSE

LSE

LSE

LSE

LSE

LSE

LSE

Ideal Placement of PMUs

14 bus example graphical representation

Qiao Li, Tao Cui, Yang Weng, Rohit Negi, Franz Franchetti and Marija D. Ilic, “An information theoretic approach to PMU placement in electric power systems,

IEEE Transactions on Smart Grid, Special Issue on Computational Intelligence Applications in Smart Grids. (Accepted, to appear) 2013

PMU Information Gain Index

Qiao Li, Tao Cui, Yang Weng, Rohit Negi, Franz Franchetti and Marija D. Ilic, “An information theoretic approach to PMU placement in electric power systems,

IEEE Transactions on Smart Grid, Special Issue on Computational Intelligence Applications in Smart Grids. (Accepted, to appear) 2013

Looking ahead- Framework for integrating combination of technologies at value

• Value is a system-dependent concept (time over which decision is made; spatial; contextual)

• Cannot apply capacity-based thinking; cannot apply short-run marginal cost thinking

• Reconciling economies of scope and economies of scale

• Value of flexibility (JIT,JIP, JIC)

• Hardware, information, decision-making software; distributed, coordinated –all have their place and value

35