Synchrophasor Applications

ERCOT STF Meeting, Feb 5, 2014

David Costello

David Costello, P.E.Technical Support DirectorSchweitzer Engineering Laboratories, Inc.Office – 509.334.5741M bil 830 377 9018

li k di /i /d id t ll /

Mobile – 830.377.9018e-mail – david_costello@selinc.com

www.linkedin.com/in/davidacostello/

http://web.ecs.baylor.edu/faculty/grady/ http://web.ecs.baylor.edu/faculty/grady/ T S h h N t k ht lT S h h N t k ht lTexas_Synchrophasor_Network.htmlTexas_Synchrophasor_Network.html

“The MRI of Power Systems”“The MRI of Power Systems”

NERC l Fl id F b 26 2008NERC press release on Florida outage Feb. 26, 2008:

Synchrophasors are “Like the MRI of bulk power systems,” said Rick Sergel, President of NERC.

120V Wall Outlets Work for Synchrophasors (of course, we would prefer to have three-phase grid PMUs)

McDonald Observatory Voltage Phase Angle w.r.t U.T. Austin 120V and Harris 69kV

-10

Voltage Ringdown at McDonald Observatory Observed at the Following Two Locations in Austin: a 120V Wall Outlet on Campus, and the Harris 69kV Substation that Feeds the Campus

-11

10 seconds

20-second window

-12

Deg

rees w.r.t U.T. Austin

w.r.t Harris 69kV

-13

-141600 1700 1800 1900 2000 2100 2200

Sample (30 samples per second)

• The fixed net multiple of 30 degree phase shift between U.T. Austin 120V and Harris 69kV has been removed. The variable but steady power flow phase shift through the substation transformer has also been removed.

Modal AnalysisModal AnalysisSh S t RSh S t RShows System ResponseShows System Response

p

3

4

5

s Ph A l

Grid response to unit trip gives a measure of stability

0

1

2

Deg

rees McD

PanAm

Phase Angleyvia the normalized damping ratio, zeta

Start Sec Stop Sec A B T1 Tau1 C T2 Tau2 Tdamp Fdamp Zeta53 60 2.36 0.81 54.14 0.32 -1.29 54.55 1.96 1.44 0.696 0.116

Exponential Steady State Transition Curve Damped Sinusoidal Term

-1

U.T. Pan Am Voltage Phase Angle wrt Baylor, Sept. 10, 2013, 04:36 GMT3

ActualC Fit

1

2

egre

es

Curve Fit

1

0

D

Copyright © SEL 2013

-152 53 54 55 56 57 58 59 60 61 62

Second

Tuesday, March 31, 2009, 20 Minute Window Beginning 12:53 AM CDT Surprises. Very Unusual Events Occur and Need Investigation

60.1

Superimposed Frequency Measurements Taken at UT Austin (in black) and McDonald Observatory (in red)

60.0

20-minute window with strange 0.067 Hz oscillation

59.90 2 4 6 8 10 12 14 16 18 20

Cause - Throttle valving problem on generator under test. Slow frequency variation had gone unnoticed.

ERCOT’s graph of West-to-North P flow

This is probably the world’s first electricity market-induced grid oscillation!12 hour window

West to North P flow

Price went down, then wind gens responded, then price went up, then wind gens responded, • • •

McDonald Observatory Voltage Phase Angle Relative to Central ERCOT

40Texas Synchrophasor N k’ h f W

20

30

40

ees

Network’s graph of West-to-Central ERCOT Voltage Angle

-10

0

10

Deg

re

12-hour window

Candidate for Ripley’s Believe It or Not?

-200 1 2 3 4 5 6 7 8 9 10 11 12

Hour of March 11, 2011

12-hour window It or Not?

High Wind Generation Levels Present Challenges for Grids,but ERCOT is Handling Them Well

30

35

ERCOT Wind Gen (Percent of Load)Week Starting 00.00 GMT, Sunday, April 21, 2013

7000

8000

9000

ERCOT Wind MWWeek Starting 00.00 GMT, Sunda

15

20

25

4000

5000

6000

0

5

10

0 1 2 3 4 50

1000

2000

3000

0 1 2 3 40 1 2 3 4 5

Day of Week

50

60

Aust

in

0 1 2 3 4

Day of Week

On April 21, 2013, wind generation dropped from

10

20

30

40D

egre

es w

rt U

T A

• 35% to 4% of total generation,• 8500 MW to 1100 MW,in 12 hours

V lt h l (l dfl00

10

0 24 48 72 96

Hour of the Week, GM

Voltage phase angle (loadflow, short circuit, stability, state estimator), dropped 40° and then rose 45°

Surprises. A 2 Hz Mode in Ambient Oscillation Sometimes Forms with High Wind Generation

Big Wind (20%) with 2 Hz Cluster Small Wind (2%) without 2 Hz Cluster

March 18, 02:00 – 03:00 i d G i 20%

March 12, 02:00 – 03:00

2 Hz

Wind Generation > 20% Wind Generation ≈ 2%

Cluster

42 Major Unit Trips, 0.1 Hz or Greater. Conclusion - No Correlation Between Wind Generation and System Inertia

Estimated System H versus % Wind Generation(42 Unit Trip Events, June through November 2010)

y

6 months

10

12

6

8

imat

ed H

2

4Esti

00 5 10 15 20

Wind Generation - % of Total Generation%

Does Wind Generation Impact Grid Inertia?

EPRI Study. Purpose – to compute ERCOT System Inertia Constant H From Frequency Response During 42 Unit Trips Having 0.1 Hz or Greater Freq. Drop.

smsss PJJH

112

1 2

Drop.

ratedsratedrated PdtdPPH

/22

fPfPP smsmsm 21 dtdfP

fPdtdf

fP

PdtdP

PHsrated

sm

s

s

rated

m

s

s

rated

m/2/2

22/2

1

H has the units of seconds. The correct interpretation is that the kinetic energy in the equivalent system machine corresponds to H seconds of rated power. Thus, the machine could provide rated power for H seconds, at which time it would have spun down to zero RPMtime it would have spun down to zero RPM.

Select A DisturbanceSelect A Disturbance

776 MW Generator Tripped*776 MW Generator Tripped*

RealReal--Time PTime Pmm –– PPee CalculationCalculation

Measure Frequency ChangeMeasure Frequency Change

FrequencyFrequency Changes by 0.12

Hz

Measure Time ChangeMeasure Time Change

Ti Ch bTime Changes by 3 seconds

df/dt Calculationdf/dt Calculation

df/dt = 0.12/3

This is 40 mHz / dsecond

Estimated Generation LostEstimated Generation Lost

• df/dt = -40 milliHz/second

H 9 d• H = 9 seconds

• VAbase = 60 GWbase

• fnom = 60 Hz

• Pm – Pe = 720 MW

Already Installed PMUAlready Installed PMU--Capable SEL Capable SEL D i i E h G idD i i E h G idDevices in Each GridDevices in Each Grid

Developing Synchrophasor Solutions Developing Synchrophasor Solutions f O D df O D dfor Over a Decadefor Over a Decade

ries)

lys ys

y (SEL-

400 S

erie

er Web za

tion +

Con

trol

00 S

eries

Rela

ys

oces

sor

Clients

PDCs

00 S

eries

Rela

ys

E®

haso

rs in

a Rela

y

haso

rs in

a Mete

r

haso

rs on

the W

e

ocol

in a R

elay

ynch

roph

asor

s

PDCeiv

es S

ynch

roniz

a

haso

rs in

SEL-30

haso

r Vec

tor P

roc

ports

C37

.118 C

ation

Arch

iving

P

RAdd

s PMU

ntra

l

haso

rs in

SEL-70

RS Add

s PMU

Adds P

MU

d SYNCHROWAVE

hanc

emen

ts

Synch

roph

aSyn

chro

pha

Synch

roph

aIE

EE Pro

toc

SCADA Syn

Hardw

are P

PMU Rec

eiv

Synch

roph

a

Synch

roph

aRTA

C Sup

p2n

d Gen

era

SEL-65

1R A

Cent

Synch

roph

aSEL-

351R

SSEL-

2431

A

PDC and S

Centra

l Enh

a

’02 ’03 ’06 ’07’05 ’08 ’09 ’10 ’11 ’12 ’13

These solutions are in service today, worldwide

SEL Synchrophasor Building BlocksSEL Synchrophasor Building BlocksPh M t U itPh M t U itPhasor Measurement UnitsPhasor Measurement Units

SEL-751

SEL Synchrophasor Building BlocksSEL Synchrophasor Building Blocks

https://www.selinc.com/synchrophasors/https://www.selinc.com/synchrophasors/

PMU SnapshotsPMU Snapshots

Easily Identify ErrorsEasily Identify Errors

Check CT Polarities, Phasing, and RatioCheck CT Polarities, Phasing, and Ratio

Find Problems Find Problems With PT CVTWith PT CVTWith PTs, CVTs With PTs, CVTs

and CTsand CTs• Discover loose

connections inconnections in potential circuits

• Detect CVT drift and future failure

• Prevent future outages andoutages and misoperations

Samples Time Stamped +/Samples Time Stamped +/-- 5 5 microsecmicrosec

NERC PRCNERC PRC--002002--2 2 Di t b R diDi t b R diDisturbance RecordingDisturbance Recording

SEL-3354

SEL-487E

SEL-2401

Advanced Event AnalysisAdvanced Event Analysis

Improve Understanding of EventsImprove Understanding of EventsE lE l 10 23 J l 14 201210 23 J l 14 2012Example Example –– 10:23 p.m. July 14, 201210:23 p.m. July 14, 2012

Improper relaying on Grizzly-Ponderosa 500 kVp p y g y

Eastern Interconnection Transient Eastern Interconnection Transient 3 06 48 J l 21 20123 06 48 J l 21 20123:06:48 a.m. July 21, 20123:06:48 a.m. July 21, 2012

Generator TripGenerator Trip9 21 00 M 14 20129 21 00 M 14 20129:21:00 a.m. May 14, 2012 9:21:00 a.m. May 14, 2012

500 kV DC Line Trip 500 kV DC Line Trip A il 20 2012A il 20 2012April 20, 2012April 20, 2012

750 MW Generation Trip750 MW Generation TripJ l 8 2012J l 8 2012July 8, 2012 July 8, 2012

WideWide--Area VisualizationArea Visualization

Get Instant Feedback After Get Instant Feedback After S t ChS t ChSystem ChangesSystem Changes

Distribution Planning and ImprovementDistribution Planning and ImprovementE lE lExampleExample

Detection of loose fuse caps, ploose connections at safety switch or terminal

cabinets, and animal damage to wiring

Study the Impact of RenewablesStudy the Impact of RenewablesE lE l 7 13 J 7 20127 13 J 7 2012Example Example –– 7:13 a.m. January 7, 20127:13 a.m. January 7, 2012Sudden high frequency oscillation – cause g q y

under investigation (an inverter?)

See Information That SCADA MissesSee Information That SCADA Misses

Wind Farm

Dip

Oscillations

Vol

tage

DP

erce

nt

System Reconnection and RestorationSystem Reconnection and RestorationE lE lExampleExample

Frequencies of three islanded systemsq y

Analyze Analyze thethethe the

CommunicaCommunica--tionstions NetworkNetwork

Model VerificationModel VerificationE lE lExampleExample

Observe inertia (H) from generation rejection event( ) g j

System SettingsSystem SettingsE lE l J 18 2012J 18 2012Example Example –– January 18, 2012January 18, 2012

Modal analysis shows marginal damping prior to unusual system events led to power systemprior to unusual system events, led to power system

stabilizer settings changes

AntiAnti--Islanding ExampleIslanding Example

• Implemented with Florida Power & Light

• Required for IEEE 1547

• Must disconnect DG within 2 seconds• Must disconnect DG within 2 seconds

• Options include♦ Transfer trip

♦ Vector shift♦ Vector shift

♦ Injected signal

Architecture With SynchrophasorsArchitecture With Synchrophasors

WideWide--Area IDS Uses Slip and Area IDS Uses Slip and A l ti f I l di D t tiA l ti f I l di D t tiAcceleration for Islanding DetectionAcceleration for Islanding Detection

WideWide--Area IDS Detects Islanding During Area IDS Detects Islanding During Mi i l P E h C ditiMi i l P E h C ditiMinimal Power Exchange ConditionsMinimal Power Exchange Conditions

1(DG)

1(RMT-SRC)

• Synchrophasor data are sent from the DG y pand remote source

• SVP time aligns and makes coherent data• SVP time-aligns and makes coherent data available to logic

Responsive to All ConditionsResponsive to All Conditions

Responsive to All ConditionsResponsive to All Conditionson

dsS

eco

Big Creek Controls Rector Big Creek Controls Rector Static VAR CompensatorStatic VAR CompensatorStatic VAR CompensatorStatic VAR Compensator

ClosedClosed--Loop Control at Big Creek Loop Control at Big Creek St bili V ltSt bili V ltStabilizes VoltageStabilizes Voltage

244 50

242

240

0

238

236

–50BC 3 MaxBC 3 MinR t A

BC 3 Average

236

234

Rector AverageRector MaxRector MinSVC Average Output

–100

232

230

–150

228 –200

Transient Stability Transient Stability C t lC t lControlControl

CFE Mexico

CFE AGSS SystemCFE AGSS System

Angostura (ANG)

Sabino (SSB)

Chicoasen (MMT)

Simple Synchrophasor Simple Synchrophasor G tG t Sh ddi L iSh ddi L iGeneratorGenerator--Shedding LogicShedding Logic

WideWide--Area Protection at Relay SpeedsArea Protection at Relay Speeds

230 kV Backbone Connects Countries 230 kV Backbone Connects Countries F G t l t PF G t l t PFrom Guatemala to PanamaFrom Guatemala to Panama

Guatemala Wheels Power From Guatemala Wheels Power From M i t C t l A iM i t C t l A iMexico to Central AmericaMexico to Central America

Supplementary Control Scheme (SCS) Supplementary Control Scheme (SCS) T i I t ti t El S l dT i I t ti t El S l dTrips Interconnection to El SalvadorTrips Interconnection to El Salvador

Oscillation Stability Assessment With Oscillation Stability Assessment With R lti C t l A ti Di l dR lti C t l A ti Di l dResulting Control Action Displayed Resulting Control Action Displayed

Traditional Control to Isolate LineTraditional Control to Isolate Line

Master Control

230 kV 206 kVBus 1 Bus 2

224 kV 224 kV230 kV 206 kV224 kV 224 kV

Tap Down Close Capacitor

Traditional Method Causes DisturbancesTraditional Method Causes Disturbances

TimeTime--Synchronized Control in ActionSynchronized Control in Action

Synchronous Control to Isolate LineSynchronous Control to Isolate Line

Master ControlLocal Local

Control Control

Execute Recipe Simultaneously

Bus 1 Bus 2224 kV 224 kV

T DTap Down Close Capacitor

Recipe Method Minimizes ImpactRecipe Method Minimizes Impact

Directly Measure the State Directly Measure the State

MasterSynchronized Data

V1

V2Master Data . . .

Vn

• Detect bad dataSub 1 Sub n• Average

measurements

D t i t l

RTAC . . . SVP

V • Determine topology

• Calculate V at adjacent stations:

V1

Vn adjacent stations:V’ = V + ZI

SRP Operator Closes Tie Based on Synchrophasors, 2004

What if I Have Multiple Tie Points?What if I Have Multiple Tie Points?

Refinery Selects From 5 Breakers

Utilities Are Operating Closer to the EdgeUtilities Are Operating Closer to the EdgeMargin

Margin

1 01.0

Operating

Point

BifurcationPoint

0.5

1.0 1.3 1.5 1.7PU Nominal Load0.0

Long Island: Monitor Angles Between Long Island: Monitor Angles Between Transmission Distribution Buses toTransmission Distribution Buses toTransmission Distribution Buses to Transmission Distribution Buses to Detect & Prevent Voltage CollapseDetect & Prevent Voltage Collapse

Vr

ZL = R + jXVS 0Vr

S = P + jQ

V)θsin(1S2

s X)θcos(2V)θsin(1S 2

smax

MAKING ELECTRIC POWER SAFE POWER SAFE, RELIABLE, AND ECONOMICAL FOR ECONOMICAL FOR 30 YEARS