Cascade-Based Planning Analysis

Robert W. Cummings

NERC Senior Director of Engineering and Reliability Initiatives

1

WHAT SHOULD BE STUDIED

2

3

Tenets of Cascading Analysis – Why do it?

Learn from History

• What is a credible combinations of events?

• What should I think might get involved?

• Be imaginative – Murphy is!

Be a Student of the System!!

• Constantly observe how your system behaves normally and when under stress

• How does your system interact with the rest of the Interconnection?

• Practice simulating actual events – Good for model validation

– Helps avoid complacency

4

Some Trends in Event Analysis

1. Protection system misoperations (39)

2. Unexpected Gen. Turbine Control Action (33)

3. Transmission equipment failures (18) (most initiating)

4. Voltage sensitivity of gen. aux. power systems (13)

5. Human Error (12)

6. Near-term load forecasting error (6)

7. Wiring errors (incidental) (5)

8. Relay loadability (5)

9. Inter-area oscillations (5)

10. SPS & RAS Misoperations (5)

5

Two Top Disturbance Elements • 33 – Unexpected generator turbine control actions

– 29 in 8 events

– 2 units CAUSAL in a system separation event

• 13 – Voltage and/or frequency sensitivity of generation auxiliary power systems (not included above) or plant / Unit Digital Control Systems

• Problem: THESE BEHAVIORS ARE NOT MODELED OR STUDIED – Boiler and Turbine controls are not modeled

– Understand roll of controls action – Power-Load Unbalance (PLU)

– Typical dynamic analysis – only analyzing t = 0 to t = 20 seconds – control actions can go well beyond that

– What about fuel system controls?

6

Planning a Cascading Study Be imaginative!

• You may have to design study methodology to mimic past disturbances to see if the system is still susceptible

• Think in bus-breaker mode – Bus-line model thinking will not get you there

– Break-to-breaker study needed to analyze protection system misoperations

• Know your protection system configurations – Directional distance, reach, differentials, breaker failure – local or

remote, Remedial Action Schemes, transfer tripping, etc.

• Expect the analysis to piece-wise – slices of time to mimic the developing overloads or potential voltage collapses

7

Planning a Cascading Study (cont.) • What parameters do you have to monitor within the study to

tell you what would be the next step of the cascade? – Look for overloaded lines – How overloaded are they?

– Voltages below 0.9 per unit – How low?

– Results of one run flow into another

– Reactive flow into both ends of a line?

Think about controls!!

• Inverter controls play a part too!

• Behavior during protracted faults – breaker failure (to operate) emulation

– SLG faults transition to multi-phase in 20-30 cycles

– Longer faults evolve into 3-phase faults

8

Planning a Cascading Study (cont.) • Protracted faults mean protracted low voltages

– UVLS schemes?

• Inverter behavior during low voltages – Inverter blocking – current injection?

– Inverter tripping?

• Know your loads!

– Know your station load composition – residential, commercial, industrial, etc. • Helps characterize load behavior under abnormal conditions

– Loads with high-quality power requirements may “leave the system”

– Voltage-sensitive or frequency sensitive loads must be monitored during the study

9

Study Practices Cardinal Rule of forensic event analysis – Everything happens for a reason!!

• There is no “sympathy in the power system”

• Think like a protection system or plant control system – What the heck did it see?? Why did it react the way it did?

– Take advantage of all available disturbance monitoring data (PRC-002)

• Leave no questions unanswered

Don’t be myopic in the scope of your analysis!

• Examine the Interconnection to ensure there were no wider area impacts

Timing is everything!

• Creating a detailed sequence of events is crucial!

10

Study Practices (Continued) Cascading typically starts slow and progresses as multiple things go wrong

Know the cascade players

• Triggering event – lightning arrestor failure

• Causal events – Breaker fails to operate – protection system failure

• Contributory events – Protection system miscoordination widens impact

• Coincidental events – It was a dark and stormy night

• Resultant events – fault progressed from single-phase to multi-phase

EXAMPLE 1 NORTHEAST BLACKOUT

AUGUST 14, 2003

11

2003 Blackout Signature

12

-2000

-1000

0

1000

2000

3000

4000

16:10:38 16:10:40 16:10:42 16:10:44 16:10:46 16:10:48

Time - EDT

MW/MVAr

0

50

100

150

200

250

300

kV

MW

MVAr

kV

Major Path to Cleveland Blocked

13

ONTARIO

4:08:59 - 4:09:07

PM

ONTARIO

Generation Trips

8

Cascade Moves into Michigan

4:10:36 PM

9

Power Transfers Shift 4:10:38.6 PM

10

North of Lake

Superior

4:10:43 – 4:10:45

PM

Northeast Island Separates from EI

11

End of the Cascade

12

Area affected by blackout

Service maintained in isolated pockets

57

58

59

60

61

62

63

64

16:10:30 16:10:40 16:10:50 16:11:00Time

Fre

qu

en

cy

(H

z)

Lambton ONT-MI

NY-West

NY-East

-3,000

-2,000

-1,000

0

1,000

2,000

3,000

4,000

5,000

16:10:30 16:10:40 16:10:50 16:11:00Time

Pow

er Flo

ws (M

W)

New York into Ontario

PJM into New York

Ontario into Michigan

New York into New England

Sta

rt o

f split

betw

ee

n

East

and W

est

MI

16:10:36

Detr

oit, C

leve

land s

ep

ara

ted f

rom

W.

MI.

16:10:38

Cle

ve

lan

d c

ut

off

fro

m P

A

16:10:38.6

NY

separa

tes f

rom

PA

16:10:39.5

Cle

ve

lan

d s

ep

ara

tes

from

Tole

do,

isla

nds

16:10:41.9

NY

separa

tes f

rom

NJ

16:10:45.2

NY

and N

ew

Engla

nd s

ep

ara

te

16:10:48

Split

com

ple

te b

etw

ee

n

East

and W

est

NY

16:10:49

16:10:30

to

16:11:00

1630 M

W D

etr

oit g

enera

tion t

rips

16:10:42

On

tario s

plits

fro

m W

est

NY

16:10:50

On

tario r

econ

nects

with W

est

NY

16:10:56

2003 Blackout Analysis

13

0

20

40

60

80

100

120

140

Outages

% o

f N

orm

al R

ati

ng

s (

Am

ps

)

Sammis-Star

345 kV

CantC-Tidd

345 kV

Star-S.Cant

345 kV

Hanna-Jun

345 kV

Hard-Chamb

345 kV

Can

tC X

fmr

Bab

b-W

.Ak

138 k

V

Hard

-Ch

am

b

345 k

V

Han

na-J

un

345 k

V

Sta

r-S.C

an

t

345 k

V

Clo

v-T

orre

y

138 k

V

E.L

ima-N

.Lib

138 k

V

W.A

k-P

V Q

21

138 k

V

E.L

ima-N

.Fin

138 k

V

Ch

am

-W.A

k

138 k

V

W.A

k 1

38 k

V

Bkr F

ailu

re

Dale

-W.C

an

138 k

V

2003 Blackout Simulations 3 Months to Build

14

Some Key Elements of the Cascade • Vegetation management

• Relay loadability

• Miscoordination of generator controls and system protection

• Generator underfrequency protection and under / over speed controls not coordinated

• Pole Slipping

• UFLS failures – Constrained by extreme undervoltage

– Time delays too long

21

What to look for in the study

• Slow cascading – not all cascades happen at dynamic speeds

• Slow, progressive Voltage collapse

• Increasing overloads of key elements over time

• Loading above Surge Impedance Limits (SIL)

Reactive power entering the line from both ends

• Excessive bus angles across transmission lines

Potential for parts of the system separating by out-of-step conditions

If it trips can you reclose it?? Doubtful for angles above 45 degrees

22

23

Angular Separation Analysis

-170

-160

-150

-140

-130

-120

-110

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

15:05:00 15:32:00 15:44:00 15:51:00 16:05:00 16:06:01 16:09:05 16:10:38

Time (EDT)

Rela

tive P

hase A

ng

le

Cleveland West MI

Normal Angle ~ -25º

Reference:

Browns Ferry

16

24

24

Detroit

Cleveland

Western MI

NJ

NY Ontario

15

Degrees

Muskingum – Ohio Central - Galion

East Lima – Fostoria Central

Sammis - Star

16:05:50 to 16:08:52

25

16:08:50 to 16:10:50 Western MI

Detroit

out of

Synch

NJ NY

Ontario Cleveland

Cleveland

Separation

NY-PJM Separation

40

Degrees

18

EXAMPLE2 FLORIDA SYSTEM DISTURBANCE

FEBRUARY 26, 2008

26

Event Overview

• Delayed clearing of 138 kV SLG fault progressed to a 1.7 second 3-Ø fault

• Loss of 1,350 MW load near fault

• Loss of 2,500 MW of generation near fault

• Loss of 2,300 MW more load by UFLS program

• Loss of 1,800 MW more generation across the Region

• Oscillatory effects across interconnection

27

Key Elements of Cascade • Protection system turned off

• Single Ø fault progressed to three Ø fault – 1.7 seconds

• Nuclear units tripped on undervoltage as designed

• Eight turbines unexpected trips; – auxiliary bus voltage protection

– rate of frequency change -- burner lean blowout phenomenon

• Proper UFLS action – prevented system separation

• Inter-area oscillations – 1,000 MW swings in TVA 500 kV system

– 600 MW swings in New England 345 kV system

– 12 kV swings on Ontario 230 kV system

30

29

Florida UFLS Activation

• Event Summary – Impact contained within

FRCC

– Majority of load restored within 2 hrs.

– No major equipment damage reported

– No thermal O/L

– 2 Nuclear units • (tripped as designed)

FRCC

RC Footprint

FRCC RC Visibility

Actuation of

UFLS

Location of

138 kV – 3θ fault

Generation Trips

Turkey Point (FPL)

Dorsey (MH)

Calloway

/ Rush

Island

TVA

Interconnection-wide Impact 29

What to look for in the study

• Localized voltage collapse

• Low voltages on generators

Motor controllers often drop out for voltages below 0.87 per unit

• Potential inter-area oscillations

• Excessive bus angles across transmission lines

Potential for parts of the system separating by out-of-step conditions

If it trips can you reclose it?? Doubtful for angles above 45 degrees

31

EXAMPLE 3 PACIFIC SOUTHWEST DISTURBANCE

SEPTEMBER 8, 2011

32

CAISO Freq/ACE 1635-1705 MST

CISO Freq/ACE 1635-1705 MST

From RA Tool

1-Min. CAISO Freq. 1500-0530 MST

From RA Tool

System Separation & SONGS Trip

Loss of San Onofre Gen.

Loss of SDGE

Load

From RA Tool

Initial FNet FDR Angular Plot

AZ-NM-CO

California

FNET FDR Locations

Event A

Event A Detail

Event B

Event C

Event D

What we learned with FNET • The frequency shows four main events

A. The initial separation around 22:27:39 (UTC), resulting in a ‘slow’ frequency dip of about -30 mHz over about 25 seconds

B. A frequency ramp beginning around 22:32:10 increasing frequency +30 mHz over about 15 seconds

C. A frequency drop around 22:37:55 of over -40 mHz (B-A) over about 12 seconds

D. A frequency jump around 22:38:21 of over +150 mHz (C-A) in less than 5 seconds, settling at around +80 mHz (B-A) in about 20 seconds

Components of the Outage

• Over 30 ‘major’ element operations over the course of 11 minutes

Line and transformer trips

Generator trips and runback

Load shedding

Over 50 additional ‘minor’ operations such as capacitor and reactor switching

• Over 6 GB of data of different qualities and resolution

Operator logs, PI historian, SCADA, PMU, DFR, relay records

What to look for in the study • Think like a phasor measurement unit (PMU)

But don’t be fooled by phase-jumps at inception of a fault, clearing of a fault, or significant reactive switching

• Think Multi- dimensionally! Don’t fixate on single reading

Frequency, voltage, current, time – multi-dimensional plots

• Potential for inter-area oscillations

• Recognize tripping of various system elements

Line trip, Transformer trip, Load trip

• Excessive bus angles across transmission lines

If it trips can you reclose it?? Doubtful for angles above 45 degrees

45

Phase 1 – Pre-Disturbance • Hot, shoulder season day; some

generation and transmission

outages

• High loading on some key

facilities: H-NG at 78% of normal

rating; CV transformers at 83%

• 44 minutes before loss of H-NG,

IID’s RTCA results showed loss

of CV-1 transformer would load

CV-2 transformer above its relay

trip point

• 15:27:39: APS technician

skipped a critical step in isolating

the series capacitor bank at

North Gila substation; H-NG trips

Phase 2 – Trip of H-NG 500 kV

• H-NG 500 kV trips at

15:27:39

• APS tells WECC RC line

expected to be restored

quickly

• H-NG flow redistributes: 77%

to SCE-SDGE (Path 44);

remainder to IID, and WALC

• CV transformers immediately

overloaded above relay

settings

• Path 44 at 5,900 amps; 8,000

amp limit on SONGS

separation scheme

15:27:39 – 15:28:16

Initiating Event – Voltage Divergence Hassayampa – North Gila 500 kV Trip

Series Capacitor Bypass Switch Arcs

Over

Hass. – N. Gila 500 kV Line Trip

Hassyampa –N. Gila 500 kV

line trip

CCM Unit 1 generator trip

South of SONGS Current

Phase 3 – Trip of CV Transformers • 15:28:16 – CV-2 and CV-1

230/92kV transformers trip on

overload relays

• Severe low voltage in WALC

161 kV system

• Loading on Path 44

increases to 6,700 amps

15:28:16 – 15:32:10

Coachella Valley Transf. Trip

Coachella Valley 230/92 kV

transformers trip

South of SONGS Current

Phase 4 – Ramon Transformer Trip • 15:32:10 Ramon 230/92kV

transformer trips on overload

relay

• 15:32:13 Blythe-Niland

161kV line trips

• 15:32:15 Niland – CV 161kV

line trips

• IID undervoltage load

shedding; loss of generation

and 92 kV transmission lines

• Severe low voltage in WALC

161 kV system

• Loading on Path 44

increases to 7,800 amps;

settles at 7,200 amps

15:32:10 – 15:35:40

Ramon Transformer Trip

Ramon 230/92 kV transformer trip

Multiple line, generator and load trips

Voltage collapse in pocket, followed by load tripping

South of SONGS Current

Voltage in Northern IID 92 kV System

Ramon 230/92 kV Transformer Trip

Trip of Over 400 MW in Northern IID 92 kV Load

Pocket

Over-Voltage Trip of 92 kV System

Capacitors

Blythe 161 kV Voltage

Trip of Coachella Valley 230/92 kV

Transformers

Trip of Hassayampa – North Gila 500 kV

Line

Ramon 230/92 kV Transformer Trip

Trip of Over 400 MW in Northern IID 92 kV Load

Pocket

Yucca 161/69 kV Transformers 1 and

2 Trip

El Centro – Pilot Knob 161 kV Line

Trip

Phase 5 – Yuma Separates

15:35:40 – 15:37:55

• Yuma AZ Separates from IID

and WALC when Gila and

Yucca transformers trip

• Yuma load pocket isolated on

single tie to SDG&E

• Loading on Path 44

increases to 7,400 amps after

Gila transformer trip; to 7,800

amps after Yucca

transformers and generator

trip

Yuma Separation

Gila 161/69 kV transformers trip

YCA generating units trip

Yucca 161/69 kV transformers trip

Pilot Knob 161/92 kV

transformers trip

South of SONGS Current

Phase 6 – High-Speed Cascade • El Centro – Pilot Knob 161kV

line trips; all IID 92 kV system

radial from SDG&E via S-Line

• WALC 161 kV system voltage

returns to normal

• Path 44 exceeds 8,000 amp

setting and timer starts

15:37:55

El Centro – Pilot Knob 161 kV line

trip

Imperial Valley – El Centro 230 kV

“S” line tripCLR generating units trip SONGS

separation

Phase 6 – High-Speed Cascade South of SONGS Current

SONGS Separation Frequency Impacts

59.85

59.95

60.05

60.15

60.25

60.35

60.45

60.55

60.65

60.75

15:38:13.920 15:38:18.240 15:38:22.560 15:38:26.880 15:38:31.200 15:38:35.520 15:38:39.840 15:38:44.160 15:38:48.480

Ault (Denver) Mead (Las Vegas) Tesla (Sacramento) Palo Verde

Grand Coulee SONGS Recalculated Frequency Devers Recalculated Frequency

60.730 Hz Phase

Jump Transient

~60.210 Hz System

Zenith (Point C)

~60.050 Hz System

Response (Value B)

59.977 Hz Pre-Event (Value A)

SONGS generators ring down

SONGS separation

SONGS Unit 3 GSU trip and transfer of

aux load

SONGS Unit 2 GSU trip and transfer of

aux load

SONGS aux load on startup

transformers

UFLS Operations in the Island

Phase 4 Example

Two GTs

& UVLS

Blythe-Niland

Colmac GT

CV-Niland

UVLS

Motor stalling

Phase 5 Example

transformers

trip

generators

trip

Phase 6 Example

‘A’ line trip

Blythe RAS

‘S’ line RAS

generators

‘S’ line RAS

generators

Devers SVC Output

2 1 3 4 5 6 7

Capacitor Switching

capacitor switching candidate signatures

location confirmed by comparing voltage

Questions?

68