DISTRIBUTED STATE ESTIMATION A. P. Sakis Meliopoulos Georgia Power - - PowerPoint PPT Presentation

1

PSERC

DISTRIBUTED STATE ESTIMATION

• A. P. Sakis Meliopoulos

Georgia Power Distinguished Professor School of Electrical and Computer Engineering Georgia Institute of Technology Atlanta, Georgia 30332

IEEE-PES – Greek Chapter, May 2007

2

PSERC

Outline

• Background Data
• Limitations of Traditional SE
• Distributed State Estimation/SuperCalibrator
• Implementation and Demonstration
• Next Step: System Wide Implementation
• Future Applications

3

PSERC

The Electric Power Grid

Recognition: The Power Grid is a Critical Infrastructure Vulnerabilities Resulting from System Complexity Vulnerabilities Resulting from Severe Weather Vulnerabilities from the Cyber Space Vulnerabilities from Fuel Supply Lines

4

PSERC

Power Grid Vulnerability

What Did We Learned from Blackouts? 1965: Relay Mis-operation/Lack of Real Time System Info 1989: Vulnerability From Solar Storms 2003: Lack of Situational Awareness/Relay “mis-operation”

• r “Zone 3: Yes or No”

2005: Dynamic Interactions of Electric Loads

• Etc. Etc.

5

PSERC

Consensus:

Real Time Knowledge of the Operating Conditions

• f the Grid Drastically Contributes to Grid Security

Power Grid Vulnerability

6

PSERC

Basic Operational Tool

August 14, 2003 Report: Lack of Situational Awareness

Power Grid Visibility

Basic Tools: SCADA (unfiltered) and SE (filtered) The objective of SE is to provide a reliable real time model How well is it done? Historical performance of SE suggests an Average reliability of 95% Is this performance acceptable? Is this performance acceptable?

7

PSERC

Motivation: Present Operating Model

Real Time Model

State Estimation

Applications

Load Forecasting Optimization (ED, OPF) VAR Control Available Transfer capability Security Assessment Congestion management Dynamic Line Rating Transient Stability EM Transients, etc. Visualizations

Markets:

Day Ahead, Power Balance, Spot Pricing, Transmission Pricing (FTR, FGR), Ancillary Services

8

PSERC

Traditional State Estimation

Power System SE: Basic Assumptions

• Positive Sequence Model
• P, Q, V measurement set
• Near-Simultaneous Measurements
• Single Frequency

Implications:

• Balanced Operation
• Symmetric Power System
• Biased SE
• Iterative Algorithm

9

PSERC

Errors from Imbalance and Asymmetry

10

PSERC

Bus Voltage Magnitude and Phase Errors – Estimated minus Measured Value

Magnitude is Normalized, Phase is Magnified 100 times

Measurement Data: Phase A Only

Displayed Data: Phase A

Max magnitude error: 0.006 pu Phase error: (-0.110 to 0.096)

Example of Imbalance Bias

Measurement Data: Phase A Only

Displayed Data: Phase B

Max magnitude error: 0.018 pu Phase error: (-0.360 to 0.329)

11

PSERC

Voltage Measurement IC Substation A, 115 kV Bus

RG-8 Cable, 500 ft 69kV:69V Wound Type VT

Vcn = 62.33 kV / 147.46 Deg Vbn = 62.96 kV / -92.68 Deg Van = 62.53 kV / 27.52 Deg Vcn = 61.99 V / 147.45 Deg Vbn = 62.61 V / -92.70 Deg Van = 62.19 V / 27.51 Deg Vcn = 61.72 V / 148.00 Deg Vbn = 63.09 V / -92.85 Deg Van = 61.63 V / 27.11 Deg

Instrumentation Errors:

Voltage Measurement Example

Phase A Magnitude Error: 1.46% Phase A Phase Error: 0.41 degrees Phase A Magnitude Error: 1.46% Phase A Phase Error: 0.41 degrees

12

PSERC

Elimination of State Estimator Biases

Eliminates Model Biases

(Full Three-Phase Model with Neutrals, etc.)

Eliminates Imbalance Biases

(Three Phase Measurements)

Biases From Instrumentation Channel Errors

(Augment Model with Instrumentation)

Robustness

(Model Quadratization) 12

13

PSERC

Keeping Things in Perspective…

(a) SCADA Transducers: 1 to 3% (b) Modern Relays: 0.1 to 1% (c) GPS-Synchronized Equipment: Magnitude 0.1% to 1%, Phase: 0.01 to 0.05 Degrees at 60 Hz. (Systematic Errors Can Be Easily Accounted for) (b) System Asymmetries (4 to 6% differences among phases) (c) System Imbalance (0 to 12% among phases – based

• n personal observations)

(d) Instrumentation Channel Errors (0.02 to 3%)

14

PSERC

The SuperCalibrator Concept

The SuperCalibrator is conceptually very simple. The basic idea is to provide a model based error correction of substation data and in particular RELAY DATA. The SuperCalibrator is facilitated by the substation automation technology that makes all substation data readily accessible at a common point. The basic idea is to utilize a detailed model of the substation, (three-phase, breaker-oriented model, instrumentation channel inclusive and data acquisition model inclusive). Then all substation data obtained with any device, PMU, meter, relay, SCADA, etc. is expressed as a function of the state of the detailed substation model. An estimation algorithm determines the best estimate of the substation model state. GPS Synchronized Relays Make the Process Robust and the Results Globally Valid

15

PSERC

Important Point

GPS-Synchronized Measurements Make it Possible to “Distribute” the State Estimation Process The Results of a Local State Estimator Are “Globally” Valid if There is at Least One Valid GPS-Synchronized Datum

16

PSERC

The SuperCalibrator Concept

Phase Conductor

Potential Transformer Current Transformer PMU Vendor A Burden Instrumentation Cables

v(t)

v1(t) v2(t)

Burden

i2(t) i1(t)

i(t)

Attenuator Attenuator Anti-Aliasing Filters Relay Vendor C PMU Vendor C

Measurement Layer

Super- Calibrator

Data Processing

IED Vendor D LAN LAN

• Three-Phase Power System with Explicit Instrumentation Channel Model
• Use Dynamic SE to Filter Phasor Data
• Maintain Streaming Data and Visualizations

Parallel Important Activities: Substation Automation

17

PSERC

SuperCalibrator Implementation & Demonstration

18

PSERC

High Fidelity Power System Model

• Physically Based Power System Modeling
• Explicit Representation of Phase Conductors, Neutrals,

Ground Conductors and Grounding – accounts for ground potential rise

• Explicit Representation of Breakers, Switches
• Explicit Representation of Instrumentation and Relay Inputs

Integrated with the Power System

• Solver Based on the Quadratized Model (improves

robustness)

• Visualization and Animation

19

PSERC

Physically Based Model

Sequence Parameter Model

Not Used – for Info Only

High Fidelity Power System Simulator

Physically Based Models Example: Three Phase Power Line – MSU1

20

PSERC

A/D

A/D

T, 66.4kV:115V Control Cable, RG-8 A/D Converter IC Animator Voltage Meter Voltage Meter

High Fidelity Power System Simulator

Instrumentation Channel Model

( ) ( ) ( )

f I f I f g

in

• ut

i j

~ ~

,

=

Current

( ) ( ) ( )

f V f V f g

in

• ut

v j

~ ~

,

= Voltage

21

PSERC

Substation

V1e ~ V2e ~ V3e ~ V4e ~ V1s ~ V2s ~

State Definition

Definition of State for a Substation

s

V1 ~

n s c s b s a s

V V V V

, 1 , 1 , 1 , 1

~ , ~ , ~ , ~

Vector of dimension 4:

s

V2 ~

n s c s b s a s

V V V V

, 2 , 2 , 2 , 2

~ , ~ , ~ , ~

Vector of dimension 4:

e

V1 ~

n e c e b e a e

V V V V

, 2 , 2 , 2 , 2

~ , ~ , ~ , ~

Vector of dimension 4: … … …

e

V4 ~

n e c e b e a e

V V V V

, 4 , 4 , 4 , 4

~ , ~ , ~ , ~

Vector of dimension 4:

22

PSERC

Pseudomeasurements for Voltages at Next Substation

⎥ ⎦ ⎤ ⎢ ⎣ ⎡ ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ = ⎥ ⎦ ⎤ ⎢ ⎣ ⎡

R S R S

V V Y Y Y Y I I ~

Substation k

IS ~ IR ~ VS ~ VR ~

L i n e i

Given

Measurements if VS and IS of Line I at Substation k Given Line i 3-Phase Model ~ ~ ~

22 21 12 11

Compute Pseudomeasurement

R

V ~

( ) ( )

S S m pseudo

V Y Z Y Z I Z Y Z V

R

~ ~ ~

21 22 1 22 22 21 1 22 22 , − −

− + − = I I

1 22 21 12 11 22 21 12 11 −

⎥ ⎦ ⎤ ⎢ ⎣ ⎡ = ⎥ ⎦ ⎤ ⎢ ⎣ ⎡ Y Y Y Y Z Z Z Z Where:

23

PSERC

Pseudomeasurements from Kirchoff’s Current Law

Substation

I4 ~ I3 ~ I5 ~ I6 ~ I1 ~ I2 ~

level kV : , ~ k I k =

∑

General ~ ~ ~

6 2 1

= + + I I I

For the above example:

Three equations, one for each phase

~ ~ ~

5 4 3

= + + I I I

Three equations, one for each phase

( ) ( )

~ ~ ~ ~

2 1 2 4 3 1

= + + + I I k I I k

Three equations, one for each phase

24

PSERC

Pseudomeasurements for Shield/Neutral Wires

s/n a b c Given Line Model, compute ratio of s/n current over return current:

( )

c b a n s

I I I I ~ ~ ~ ~

/

+ + − = α

Then introduce pseudomeasurement:

( )

c b a m pseudo

I I I I

n s

~ ~ ~ ~

,

/

+ + − = α

25

PSERC

Phase Pseudomeasurements (if measurement of a phase is missing) By Example:

• Assume phase A voltage measurement exists
• Assume phase C voltage measurement does not exist

/

240 ,

~ ~

j a m pseudo

e V V

n s

−

=

26

PSERC

Non-GPS-Synchronized Measurements

By Example:

m c m b m a

V V V ~ , ~ , ~

Given:

α α α j c j b j a

e V e V e V ~ , ~ , ~

Replace by:

m c m b m a

I I I ~ , ~ , ~

Given:

β β β j c j b j a

e I e I e I ~ , ~ , ~

Replace by: If V and I measurements from same IED:

β α = ⇒

27

PSERC

Test System 1

NYPA

(Two Interconnected Subs)

MARCY & MASSENA Model Description

28

PSERC

The SuperCalibrator Concept Description

Substation Model

• Three Phase,
• Breaker Oriented,
• Instrumentation

Inclusive Example: Marcy Substation Inputs: 246 measurements Outputs: 12 State Variables

29

PSERC

The SuperCalibrator Concept Description

Substation Model

• Three Phase,
• Breaker Oriented,
• Instrumentation

Inclusive Example: Marcy Substation Inputs: 246 measurements Outputs: 12 State Variables

30

PSERC

Integrated Power System and Instrumentation Model

31

PSERC

Integrated Power System and Instrumentation Model

32

PSERC

Example of SuperCalibrator

Numerical Experiments Marcy Substation

# Descriptio n State Value Measurement Estimate 1 Voltage Phasor, Terminal 1, Phase A

kV j e j 2 . 50 1 . 436 439

56 . 6

+ =

kV e j

72 . 6

431

436.34, 51.07 2 Voltage Phasor, Terminal 1, Phase B

kV j e

j

1 . 405 8 . 176 442

58 . 113

− − =

−

kV e

j 46 . 113

435

−

• 176.9, -404.23

3 Voltage Phasor, Terminal 1, Phase C

kV j e j 6 . 354 8 . 258 439

13 . 126

+ − = kV e j

41 . 126

430

• 259.14, 355.0

Chi Square Test J=7.8056, Degrees of Freedom=28 Probability = 0.9625

33

PSERC

Test System 2

ENTERGY

(Two Interconnected Subs)

PANAMA & ROMEVILLE Model Description

34

PSERC

ENTERGY: Panama and Romeville Substations

35

PSERC

Panama Substation

36

PSERC

Romeville Substation

37

PSERC

PAN AM A

R O M EVILLE PAN R O M R O M

• PAN
• L

38

PSERC

ENTERGY: Panama-Romeville 230 kV Line - Dynamic Rating project

100m 1 10 100 Confidence Level (% ) 1 10 Parameter k

Error Parameter Versus Confidence Level

Case :

State Estimation Solution Report

Status : Solution Completed 4.290 99.00 12 degrees of freedom

Program W WinIGS-F

• F - F
• Form

rm Q QPFSE_GEN_REPORT

39

PSERC

40

PSERC

41

PSERC

Quantification of SuperCalibrator Output Accuracy

• Chi-Square Test provides a measure of how well the

measurements “fit” the model on a probabilistic basis. Equations omitted

• The SuperCalibrator provides a measure of the

uncertainty of the estimated states. Equations omitted.

• The SuperCalibrator provides a measure of Measurement

error – to be used for remote calibration. Equations

• mitted.

Minimizes Data to be Transferred (very important)

• Communication of Information not Raw Data
• Improve Latencies

42

PSERC

Demonstration

43

PSERC

Next Step: System Wide Implementation

• St. Joh
• St. Thomas

• St. John