Wide Area MeasurementsSystems for Grid Stability

A.M.KulkarniDepartment of Electrical Engineering, IIT Bombay

Mumbai

Organization

The Technology in a Nutshell

Wide Area Phenomena in Power Systems

A simple-to-implement Wide Area

Measurement System – Interesting

Observations of Wide Area Phenomena

WAMS Technology: Important Issues

WAMS Applications

PART – IWAMS Technology in a Nutshell

The Technology in a nutshellTwo Ideas:

1. Accurate Time SynchronizationDirect Measurement of Phase Angular

Differences

2. Faster Rate of Reception of Measured and accurately time-stamped data

Wide Area Measurements: The Technology in a nutshell

Location 1

Location 2

Magnitude of the two phasors can be determined independently but phase angle difference cannot be measured

without synchronization of measurements

Wide Area Measurements: The Technology in a nutshell

Location 1

Location 2

Phase angular difference between the two can be determined if the two local

clocks are synchronized.Synchronizing pulses obtained from

GPS satellites.

Monitoring: Old World

More Nuanced Information Available from WAMS

IEEE Power and Energy Magazine Jan/Feb 2008

Control: Old World

Area 1

Area 2

GovernorsAutomatic Voltage RegulatorsHVDC, Protection SchemesALL use local or near local feedback signals. Governors: system wide co-ordination

Emergency controls

U/f relayingTransfer trip

Ptie

Automatic Generation ControlNon-Local - slow

HVDC link

Power Transient and Control “Spectrum”

CO

NTRO

LSTRAN

SIE

NTS

Lightning/Switching

Slower Network / Torsional Transients

Prime Mover DynamicsSlow Frequency Change

Rotor Relative Angle Dynamics

Aggregate Load Changes

0.01 0.1 1 10010 1000 10000second

Equipment protection, Power Electronic Controls

System Protection Schemes

Governor, Prime Mover controls

Manual Control

PART – II WIDE AREA PHENOMENA IN POWER

SYSTEMS

Wide Area Phenomena RevisitedWhat are Wide Area Phenomena ?

Synchronous Grids: • Frequency same throughout the

grid (in steady state)• Power Flow – a function of phase

angle differences FREQUENCY AND ANGLE STABILITY

Two Synchronous MachinesOpen Circuit Phase 'a' Voltage Waveform

N

S DC

N

S DC

t=0

Machine 1

Machine 2

Snapshot at t=0 for both machines, speeds same

V1

V2

V1

V2

Transmission Line Model (Quasi-Sinusoidal Steady State)

Induction Generators ?

Three phase power is constant for balanced sinusoidal steady state

Three phase power (sinusoidal steady state) is proportional to sin (1-2)

Power

Lumped Two Port ModelV1 1 V2

Synchronous Links

Synchronous machines interconnected with AC lines

N

SDC

N

SDC

TransmissionLine

Induction Generators ?

LoadPower

LoadPower

Mechanical Power

Electrical Power

Understanding Angular / Frequency Stability – Multi-machine system

Relative motion (swing)

Centre of Inertia motion( depends on sum of forces : Fg1+Fg2+Fg3-FL1-FL2-FL3 )

Fg1

Fg2Fg3FL3 FL2FL1

Relative motion (swing)

Plots of Generator Speeds under different situations- multi-machine

Sudden Load Throw OffStable Common and Relative Motion

Sudden Generation TripStable Common and Relative Motion

Large Disturbance Angular Instability : Loss of Synchronism

Small Disturbance Angular Instability : Growing Oscillations

(triggered by any disturbance: big or small)

Frequency Stability: Centre of Inertia Motion

Angular Stability Small Disturbance

Angular Stability Small Disturbance

Synchronous Links

Synchronous machines interconnected with AC lines

N

SDC

N

SDC

TransmissionLine

Induction Generators ?

LoadPower

LoadPower

Mechanical Power

Electrical Power

Loss of Synchronism / Out of Step Operation – Idealized Scenario

Not Acceptable !Distance Relays trip

Uncontrolled System Separation

Loss of Synchronism : Lab Experiment

System SeparationIndian Blackout of 30th July 2012

Picture Courtesy: POSOCO, India

Cut Set of Separation

Loss of Synchronism Indian Blackout of 30th July 2012

PART – IIIA simple-to-implement Wide Area

Measurement System

Interesting Observations of Wide Area Phenomena

PART – IV & VWide Area TechnologyWide Area Applications

WAMS - Technology

Phasor Measurement Units

Computing Phasors Instantaneous waveform sampled at a

high rate

Moving window DFT (half or full cycle)

Magnitude and angle information obtained

Computing Phasors

samples

Moving window

Half cycle DFTComputations performed on the samples within a half cycle window.The window is attached to the latest sample.

Important Notes Phasors continuously computed Magnitude/Angle information transmitted

every 20-40 ms along with time stamp Time stamp obtained from local clock -

synchronized with all other PMU clocks using GPS signal.

Transmission of sample points of instantaneous waveforms is not envisaged (except possibly if protection application is desired)

Important Notes (Standards) Reference : Cos or Sin ? Time Stamp of beginning of

window/middle of window/end of window ? Protocols: UDP/TCP ? Positive Sequence or all three phases ? Measurement Bandwidth ? Maximum Permissible Errors (Total Vector

Error)

Communication Peer-to-peer communication

Typically for protection applications where even instantaneous samples at a high rate (e.g., 50 samples per cycle) may have to be transferred. Communication delays or latencies small.

Communication on a networkPower system monitoring, protection (i.e. system protection) and control

Mix of the both

Integration of PMU data

Communications Though many communication media possible, fiber optic

provides, by and large, the most secure and fast communication medium.

Latencies ? 30-50 ms . Larger delays may preclude design of fast system protection schemes.

Choosing a phasor reporting rate of 60 phasors/sec 1 voltage, 5 currents, 5 MW measurements, 5 MVAr

measurements, frequency, and rate of change of frequency – all reporting as floating point values – bandwidth of 64,000 bps.

A data reporting rate of 12 phasors / second for 1 voltage, 5 currents, and frequency – reported in 16 bit integer format – can be accommodated over a 4800 bps channel.

Latencies between Occurrence of Actual phenomena and reporting of phasors

CT/PT

Window size of the DFT

Processing time

Data size of the PMU output

Multiplexing and transitions

Communication distance

Data concentrators

Large DistancesInvolved ~ 1000 km

IMPORTANT: NOT TO SCALE

WAMS Applications “Low hanging fruit”

Real Time Monitoring(e.g., Angle – time curves)

Better pictureBuild-up of oscillations or monotonous decline

of voltages or frequency. Challenges : Follow-up Action ?

Post_Mortem Analysis(reduction in investigation time)

Online Damping Estimators

Power System“Ambient” Disturbances

WAMSMeasurements

Prony Analysis

EstimateDampingOf Swing Modes

Power System

Step Changes

WAMSMeasurements

System IDEstimateDampingOf Swing Modes

Power System

Multi-sine ProbingSignals

WAMSMeasurements

FFT

EstimateDampingOf Swing Modes

FFT

PACIFIC DC INTERTIE

BRAKING RESISTOR

Offline Method

WAMS Applications Benchmarking and parameter

validation(modeling errors)Systematized procedure required.

Power System RestorationBetter picture – better confidence level – better decisions.More remote actions–Remote synchronization?

WAMS Applications State Estimation

Incremental Enhancement of exist state estimators using additional measurements available

Complete renovation of state estimators with synchronized measurements- requires large scale deployment of PMUs for observability. The future ?

Utility Boundary Conditions: Areas can represent other Areas by equivalents at boundaries

WAMS Applications Islanding Detection Loss of Synchronism detection

The big ideas: Use WAMS for improving system dynamics (EMERGENCY CONTROL)

ClassificationSTABILITY IMPROVEMENT

Preventive ControlOperator Actions based on

Online Contingency Analysis

Slow. Critical contingency has not actually occurred but may potentially occur

Emergency ControlPre-designed

Automatic ControlActions (Discrete)

Fast.Critical Event has occurred, transient is evolving dangerously.

Damping ControlContinuous, Automatic Feedback

Control

HVDC, PSS, SVCetc. --- controllers always in operation.

Some “overlap”

How can WAMS help ? Angular instability :

Predict of out of step in real time -> trigger control actions like gen/load tripping or dynamic brake to prevent loss of synchronism.

OR Allow graceful system separation -- intelligent load/gen

tripping to stabilize frequency and voltage in island

Former is preferable - no resynchronization of systems BUT how does onea) Predict out of step in real time ?

b) Determine quantum of control actions ?

For controlled system separation :Adaptive choice of separation points conceivable

System Separation: Typical Cut Set

System Separation: Controlled Islanding

Courtesy : Tata Power

Predicting Instability is tough!

Predicting Instability is tough!

Faster than Real Time Simulation

A Bit too Ambitious ?

How can WAMS help ? Frequency Control:

Under Frequency Relaying:Local frequency contaminated due to swings (1-2 Hz). df/dt should not trigger on swings but on “common” motion.

Solution: filter, but filtering will involve delay.

Rate of Change of Center of Inertia Speed (NOT LOCAL)

Reflects actual power deficiency. Need to know total inertias (will need to know whether islanded or not, which generators in island)

FREQ

DF/DT

COIDF/DT

COI Speed ?

Issues Loads are voltage and frequency dependent: any

logic for restoring load generation balance should account for this (potential for over-shedding).

Lesser number of PMUs: Observability issues Models and Data

Classification

Purely localNon-local Local with Non-local supervision

Dependability versus SecurityIncorrect Action can be very harmful

Emergency Control: Issues

Dependability versus SecurityGlobally Supervised Local ControlReliability of purely global scheme ?

Comm. Latency + Processing Time available window for action after initiation of critical contingency is small. For transient instability and islanding events (~1s)

Industry Experience with SPS schemesIEEE TPWRS, Vol. 11, No.3 1996Survey of SPS schemes mainly in USA,

Canada, Europe, Japan, Australia

49 utilities, 17 countries and 111 schemes

This survey was done before the advent of WAMS

Industry Experience (IEEE- CIGRE Committee Report and Survey)

12.6Others

11.7Combination of Schemes

1.8VAr Compensation

1.8Generator Runback

1.8Dynamic Braking

1.8Discrete Excitation Control

2.7Out of Step Relaying

3.6HVDC Control

4.5Stabilizers

4.5Load and Generator Rejection

6.3Turbine Valve Control

6.3System Separation

8.2U/f Load Shedding

10.8Load Rejection

21.6Generator Rejection

%Type of Emergency Control

“……. the system condition requiring action does not occur often, but when it does occur, the SPS usually performs its function correctly”

Human Error

Incorrect Setting

Hardware Failure

Software Failure

Inadequate Design Other

0

5

10

15

20

25

30

35

40

45

This is before the advent of WAMS

This is before the advent of WAMS

Operational Statistics

Summary over 1986-1992Successful Operations : 1093Failures : 36Un-successful operations : 20Unnecessary operations : 306

This is before the advent of WAMS

WAMS applications Protection

Power Swing blockingImproved back up protectionCurrent Differential protection

Relay Vulnerability Analysis

Continuous Closed Loop Control ???e.g., Stabilizers using global signals

Require to accurately determine the communication latencies for continuous control

---- Modal speed signals for selective damping

Conclusions

WAMS: Way to the future Better monitoring and control