Power System Stability

Tutorial
Power System Stability in the New Industry

Environment: Challenges and Solutions
presented by:
Dr. Prabha S. Kundur

Kundur Power Systems Solutions, Inc.
Toronto, Ontario
Canada
Copyright © P. Kundur
This material should not be used without the author's consent 1529pk - 1
Power System Stability and Control
Tutorial Outline
1. Brief Introduction to Power System Stability

 Basic concepts
 Classification
2. Examples of Major System Blackouts Caused by Different Forms of Instability
3. Challenges to Secure Operation of today's Power Systems
4. Major System Blackouts in 2003 and 2004
5. Comprehensive Approach to Enhancing Power System Stability
1529pk - 2
Power System Stability
 Refers to continuance of intact operation of power system, following a

disturbance
 Recognized as an important problem for secure system operation since
the 1920s
 Major concern since the infamous November 9, 1965 blackout of
Northeast US and Ontario
 criteria and analytical tools used till now largely based on the developments
that followed
 Presents many new challenges for today's power systems
1529pk - 3
Power System Stability: Basic Concepts
 Power System Stability denotes the ability of an electric power system,

for a given initial operating condition, to regain a state of operating
equilibrium after being subjected to a physical disturbance, with all
system variables bounded so that the system integrity is preserved
 integrity of the system is preserved when practically the entire power
system remains intact with no undue tripping of generators or loads
 Stability is a condition of equilibrium between opposing forces:

 instability results when a disturbance leads to a sustained imbalance
between the opposing forces
Ref: IEEE/CIGRE TF Report, "Definition and Classification of Power System Stability",

IEEE Trans. on Power Systems, Vol. 19, pp. 1387-1401, August 2004
1529pk - 4
Basic Concepts (cont'd)
 Following a transient disturbance, if the power system is stable it will reach a new
equilibrium state with practically the entire system intact:
 faulted element and any connected load are disconnected
 actions of automatic controls and possibly operator action will eventually restore
system to normal state
 On the other hand, if the system is unstable, it will result in a run-away or

run-down situation; for example:
 a progressive increase in angular separation of generator rotors, or
 a progressive decrease in bus voltages
 An unstable system condition could lead to cascading outages, and a shut-down of a

major portion of the power system
1529pk - 5
Classification of Power System Stability
 Classification into various categories greatly facilitates:

 analysis of stability problems
 identification of essential factors which contribute to instability
 devising methods of improving stable operation
 Classification is based on the following considerations:

 physical nature of the resulting instability
 size of the disturbance considered
 devices, processes, and the time span involved
 We should always keep in mind the overall stability !

 solutions to problems of one category should not be at the expense of another
1529pk - 6
Power
PowerSystem
SystemStability
Stability Consideration
for
Classification
Physical
Rotor Angle Frequency Voltage Nature/ Main
Rotor Angle Frequency Voltage
Stability Stability Stability System
Stability Stability Stability
Parameter
Small-Signal Transient
Transient Large-Disturbance
Large-Disturbance Small-Disturbance
Small-Disturbance Size of
Small-Signal
Stability Stability
Stability Voltage
VoltageStability
Stability Voltage
VoltageStability
Stability Disturbance
Stability
Short Term Short

Short Term ShortTerm
Term Long
LongTerm
Term
Time Span
Short Term Long Term
Short Term Long Term
1529pk - 7
Rotor Angle Stability
 Ability of interconnected synchronous machines to remain in synchronism

after being subjected to a disturbance
 Depends on the ability to restore equilibrium between electromagnetic
torque and mechanical torque of each synchronous machine
 If the generators become unstable when perturbed, it is as a result of
 a run-away situation due to torque imbalance
 A fundamental factor is the manner in which power outputs of synchronous

machines vary as their rotor angles swing
 Instability that may result occurs in the form of increasing angular swings of
some generators leading to loss of synchronism with other generators
1529pk - 8
Transient Stability
 Term traditionally used to denote large-disturbance angle stability

 Ability of a power system to maintain synchronism when subjected to a
severe transient disturbance:
 influenced by the nonlinear power-angle relationship
 stability depends on the initial operating condition and severity of the
disturbance
 A wide variety of disturbances can occur on the system:
 The system is, however, designed and operated so as to be stable for a
selected set of contingencies
 usually, transmission faults: L-G, L-L-G, three phase
1529pk - 9
Small-Signal (Angle) Stability
 Small-Signal (or Small-Disturbance) Stability is the ability of a power system

to maintain synchronism under small disturbances
 disturbance considered sufficiently small if linearization of system equations
is permissible for analysis
 Instability that may result can be of two forms:
 aperidic increase in rotor angle due to lack of sufficient synchronizing torque
 rotor oscillations of increasing amplitude due to lack of sufficient damping
torque
 In today's practical power systems, SSS problems are usually associated with
oscillatory modes
 local plant mode oscillations: 0.8 to 2.0 Hz
 interarea oscillations: 0.1 to 0.8 Hz
1529pk - 10
Voltage Stability
 Ability of power system to maintain steady voltages at all buses in the system
after being subjected to a disturbance
 A system experiences voltage instability when a disturbance, increase in load

demand, or change in system condition causes:
 a progressive and uncontrollable fall or rise in voltage of buses
in a small area or a relatively large area
 Main factor causing voltage instability is the inability of power system to maintain
a proper balance of reactive power and voltage control actions
 The driving force for voltage instability is usually the load characteristics
1529pk - 11
Short-Term and Long-Term Voltage Stability
 Short-term voltage stability involves dynamics of fast acting load

components such as induction motors, electronically controlled loads and
HVDC converters
 study period of interest is in the order of several seconds
 dynamic modeling of loads often essential; analysis requires solution of
differential equations using time-domain simulations
 faults/short-circuits near loads could be important
 Long-term voltage stability involves slower acting equipment such as tap-

changing transformers, thermostatically controlled loads, and generator
field current limiters
 study period may extend to several minutes
1529pk - 12
Frequency Stability
 Ability to maintain steady frequency within a nominal range following a

disturbance resulting in a significant imbalance between generation and
load
 Instability that may result occurs in the form of sustained frequency

swings leading to tripping of generating units and/or loads
 In a small "island" system, frequency stability could be of concern for

any disturbance causing a significant loss of load or generation
1529pk - 13
Frequency Stability (cont'd)
 In a large interconnected system, frequency stability could be of

concern only following a severe system upset resulting in the system
splitting into islands
 Depends on the ability to restore balance between generation and load

of island systems with minimum loss of load and generation
 Generally, frequency stability problems are associated with

inadequacies in equipment responses, poor coordination of control and
protection systems
1529pk - 14
Examples of Major System Blackouts Caused by
Different Forms of Instability
1. November 9, 1965 blackout of Northeast U.S. and Ontario
2. April 19, 1972, blackout of Eastern Ontario
3. July 2, 1996 disturbance of WSCC (Western North American

Interconnected) System
4. August 10, 1996 disturbance of WSCC system
5. March 11, 1999 Brazil blackout
1529pk - 15
November 9, 1965 Blackout of
Northeast U.S. and Ontario
1529pk - 16
November 9, 1965 Blackout of NE U.S. and Ontario
 Clear day with mild weather; load levels in the region normal
 Problem began at 5:16 p.m.
 Within a few minutes, there was a complete shut down of electric service to:
 virtually all of the states of New York, Connecticut, Rhode Island,
Massachusetts, Vermont
 parts of New Hampshire, New Jersey and Pennsylvania
 most of Ontario
 Nearly 30 million people were without power for about 13 hours
1529pk - 17
North American Eastern Interconnected System
1529pk - 18
Events that Caused the 1965 Blackout
 The initial event was the operation of a backup relay (zone 3) at

Beck GS in Ontario near Niagara Falls
 opened circuit Q29BD, one of five 230 kV circuits connecting Beck
GS to load centers in Toronto and Hamilton
 Prior to opening of Q29BD, the five circuits were carrying

 1200 MW of Beck generation, and
 500 MW import from Western NY State on Niagara ties
 Loading on Q29BD was 361 MW at 248 kV;

The relay setting corresponded to 375 MW
1529pk - 19
Events that Caused the 1965 Blackout (cont'd)
1529pk - 20
Events that Caused the 1965 Blackout (cont’d)
 Opening of circuit Q29BD resulted in sequential tripping of the remaining four

parallel circuits
 Power flow reversed to New York: total change of 1700 MW
 Generators in Western New York and Beck GS lost synchronism, followed by

cascading outages: Transient (Angle) Instability !
 After about 7 seconds from the initial disturbance

 system split into several separate islands
 Eventually most generation and load lost due to the inability of islanded systems
to stabilize: Frequency Instability !
1529pk - 21
Formation of Reliability Councils
 Northeast Power Coordinating Council (NPCC) formed in January 1966

 to improve coordination in planning and operation among utilities
 first Regional Reliability Council (RRC) in North America
 Other eight RRCs formed in the following months
 National/North American Electric Reliability Council (NERC) established in 1968
 Detailed reliability criteria were developed;

Procedures for exchange of data and conducting stability studies were established
 Many of these developments have had an influence on utility practices worldwide;

still largely used !
1529pk - 22
Special Protections Implemented after the 1965 Blackout
 P Relays on Niagara Ties
 trip Niagara ties to NY when P exceeds set value;

cross-trip St. Lawrence ties to NY
 in place until mid 1980s
 Underfrequency load shedding (UFLS) throughout the

interconnected system
 beginning of the use of UFLS by the industry
1529pk - 23
April 19, 1972 Blackout of
Eastern Ontario
1529pk - 24
April 19, 1972 Disturbance: Eastern Ontario
 Incident:
 230 kV lines east of Toronto tripped due to communication malfunction; ties to New York at St.
Lawrence tripped
 generation rich island formed in eastern Ontario (G=3900 MW, L=3000 MW)
 frequency rose to 62.5 Hz and then dropped to 59.0 Hz due to speed governor
 underfrequency load shedding !
 frequency rose to 62.6 Hz and dropped to 58.7 Hz
 stabilized at 60.8 Hz with 1875 MW generation
 Frequency Instability !
 Source of problem: overspeed controls associated with prime-mover governors of Pickering “A:”
NGS
1529pk - 25
MHC Turbine Governing System with Auxiliary
Governor
1529pk - 26
Transient Response of Nuclear Units with Auxiliary
Governor
1529pk - 27
Transient Response of Nuclear Units with Auxiliary
Governor Out-of-Service
1529pk - 28
July 2, 1996 WSCC / WECC
(Western North American
Interconnected System)
Disturbance
1529pk - 29
WSCC July 2, 1996 Disturbance
 Started in Wyoming and Idaho area at 14:24:37
 Loads were high in Southern Idaho and Utah;

High temperature around 38°C
 Heavy power transfers from Pacific NW to California

 Pacific AC interties - 4300 MW (4800 rating)
 Pacific HVDC intertie - 2800 MW (3100 capacity)
1529pk - 30
WSCC July 2, 1996 Disturbance (cont'd)
1529pk - 31
 LG fault on 345 kV line from Jim Bridger 2000 MW plant in Wyoming to

Idaho due to flashover to a tree
 tripping of parallel line due to relay misoperation
 Tripping of two (of four) Jim Bridger units as stability control; this should
have stabilized the system
 Faulty relay tripped 230 kV line in Eastern Oregon
 Voltage decay in southern Idaho and slow decay in central Oregon
1529pk - 32
WSCC July 2, 1996 Disturbance (cont’d)
 About 24 seconds later, a long 230 kV line (Amps line) from western Montana to Southern
Idaho tripped, due to zone 3 relay operation
 parallel 161 kV line subsequently tripped
 Rapid voltage decay in Idaho and Oregon
 Three seconds later, four 230 kV lines from Hells Canyon generation to Boise tripped
 Two seconds later, Pacific intertie lines separated
 Cascading to five islands 35 seconds after initial fault
 2.2 million customers experienced outages; total load lost 11,900 MW
 Voltage Instability!!!
1529pk - 33
1529pk - 34
Loss of voltage control following

the tripping of the Amps line
Time in Seconds
1529pk - 35
TSAT was Used to Replicate Disturbance
in Time Domain
MEASURED RESPONSE
SIMULATED RESPONSE
1529pk - 36
August 10, 1996 WSCC
(WECC) Disturbance
1529pk - 37
WSCC August 10, 1996 Disturbance
 High ambient temperatures in Northwest;

high power transfer from Canada to California
 Prior to main outage, three 500 kV line sections from lower Columbia River
to load centres in Oregon were out of service due to tree faults
 California-Oregon Interties loaded to 4330 MW north to south
 Pacific DC Intertie loaded at 2680 MW north to south
 2300 MW flow from British Columbia
 Main outage: Ross-Lexington 230 kV line at 15:47:36
 Growing 0.23 Hz interarea oscillations caused tripping of lines resulting in
formation of four islands
 Small-Signal Oscillatory Instability !
1529pk - 38
August 10th, 1996 WSCC Event
1529pk - 39
WSCC August 10, 1996 Disturbance (cont'd)
3000
2900 Malin - Round Mountain MW Flow

2800
2700
2600
2500
2400
2300 0 3 6 9 12 16 19 22 25 28 31 34 37 40 43 47 50 53 56 59 62 65 68 71 74
Time in Seconds
1529pk - 40
WSCC August 10, 1996 Disturbance (cont'd)
As a result of the undamped

oscillations, the system split
into four large islands
Over 7.5 million customers

experienced outages ranging
from a few minutes to nine
hours! Total load loss 30,500
MW
1529pk - 41
TSAT was Used to Replicate Disturbance in
Time Domain
MEASURED RESPONSE
SIMULATED RESPONSE
1529pk - 42
1529pk - 43
Sites Selected for PSS Modifications
San Onofre
(Addition) Palo Verde
(Tune existing)
1529pk - 44
Power System Stabilizers
With existing controls

Eigenvalue = 0.0597 + j 1.771
Frequency = 0.2818 Hz
Damping = -0.0337
With PSS modifications

Eigenvalue = -0.0717 + j 1.673
Frequency = 0.2664
Damping = 0.0429
1529pk - 45
Design of HVDC Modulation
 HVDC intertie shown (as expected) to have low participation in the mode of
interest (0.23 Hz interarea oscillations)
 Often however, HVDC can be modulated to improve damping, provided
adequate input signal is found and proper compensator is designed
 SSAT used to examine frequency response for several potential input
signals
 Frequency response magnitude identified local bus frequency as having
good operability/controllability of the mode of interest
 Frequency response phase used to design compensator which provides
proper modulation signal to HVDC controls
 TSAT and SSAT used to verify modulation performance
1529pk - 46
TSAT Verification of Effectiveness of HVDC
Modulation
Without HVDC Modulation

Eigenvalue = 0.0597 + j 1.771
Frequency = 0.2818 Hz
Damping = -0.0337
With HVDC Modulation

Eigenvalue = -0.108 + j 1.797
Frequency = 0.2859
Damping = 0.0602
1529pk - 47
March 11, 1999 Brazil
Blackout
1529pk - 48
March 11, 1999 Brazil Blackout
 Time: 22:16:00h, System Load: 34,200 MW
 Description of the event:

 L-G fault at Bauru Substation as a result of lightning causing a bus insulator
flashover
 the bus arrangement at Bauru such that the fault is cleared by opening five 440
kV lines
 the power system survived the initial event, but resulted in instability when a
short heavily loaded 440 kV line was tripped by zone 3 relay
 cascading outages of several power plants in Sao Paulo area, followed by loss
of HVDC and 750 kV AC links from Itaipu
 complete system break up: 24,700 MW load loss; several islands remained in
operation with a total load of about 10,000 MW
 Transient instability followed by voltage problems
1529pk - 49
March 11, 1999 Brazil Blackout (cont'd)
 Measures to improve system security:

 Joint Working Group comprising ELECTROBRAS, CEPEL and ONS staff
formed
 organized activities into 8 Task Forces
 Four international experts as advisors
 Remedial Actions:
 power system divided into 5 security zones: regions with major generation and
transmission system;
emergency controls added for enhancing stability
 improved layout and protection of major EHV substations
 improved maintenance of substation equipment and protection/control
equipment
 improved restoration plans
1529pk - 50
Challenges to Secure Operation of
Today's Power Systems
1529pk - 51
Limitations of Traditional Approach to Power
System Stability
 Focus largely on one aspect of stability: "transient stability"

 Deterministic approach for system security assessment
 System designed and operated to withstand
 loss of any single element preceded by a fault
 referred to as N-1 criterion
 Analysis by time-domain simulation of selected operating conditions
 scenarios based on judgment/experience
 Operating limits based on off-line studies
 system operated conservatively within pre-established limits
 "Adhoc Approach" to application of power system stability controls
1529pk - 52
Challenges to Secure Operation of Today's
Power Systems
 Power Systems are large complex systems covering vast geographic areas
 national/continental grids
 highly nonlinear higher order system
 Many processes whose operations need to be coordinated

 millions of devices requiring harmonious interplay
 Increasing use of Wind Power for generation of electricity
 requires careful consideration in integration with power grids
cont'd
1529pk - 53
Challenges to Secure Operation of Today's Power
Systems (cont'd)
 Complex modes of instability

 global problems
 different forms of instability: rotor angle, voltage, frequency
 "Deregulated" market environment

 many entities with diverse business interests
 system expansion and operation driven largely by economic drivers
 lack of coordinated planning
1529pk - 54
Example of a Complex Mode of Instability
A transmission line fault causes transient instability of a remote

area:
 Sensitive to conditions in the faulted area
 Nature of the stability problem is not readily apparent
1529pk - 55
North American Western Interconnected System
1529pk - 56
Case A
 4-cycle fault on Palo Verde - Devers line (Arizona-California)
 Alberta to B.C. transfer 500 MW
 East of River interface flow 7300 MW
Note: power flow conditions considered for this study as unusual, and
do not represent present operating conditions
1529pk - 57
1529pk - 58
BPA Bus Voltage
1529pk - 59
BC Hydro Bus Voltage Near Alberta Tie
1529pk - 60
Rotor Angles in B.C. and Alberta
1529pk - 61
Case B
 East of River interface flow reduced to 7000 MW

(from 7300 MW for Case A)
1529pk - 62
BC Hydro and Alberta Bus Voltages
1529pk - 63
Rotor Angles in B.C. and Alberta
1529pk - 64
North American Western Interconnected System
1529pk - 65
Major Power System Blackouts in 2003
and 2004
1529pk - 66
Blackouts in 2003 and 2004
We had several wake up calls since 2003:

 August 14, 2003 blackout of North East USA and Ontario
 63,000 MW load loss affecting 50 million people
 September 23, 2003 blackout of South Sweden and East Denmark
 6,500 MW load loss affecting 4 million people
 September 28, 2003 blackout of Italy
 50,000 MW load unsupplied affecting 60 million people
 August 12, 2004 blackout of three Australian States: Queensland,
NSW and Victoria
 load loss 1,000 MW
1529pk - 67
August 14, 2003 Blackout of Northeast US
and Canada
1529pk - 68
14 August 2003 Blackout of Northeast US - Canada
 Approximately 50 million people in 8 states in the US and

2 Canadian provinces affected
 63 GW of load interrupted (11% of total load supplied by Eastern North

American Interconnected System)
 During this disturbance, over 400 transmission lines and 531

generating units at 261 power plants tripped
 For details refer to: "Final Report of Aug 14, 2003 Blackout in the US
and Canada: Causes and Recommendations", US-Canada Power
System Outage Task Force, April 5, 2004.
www.NERC.com
1529pk - 69
NERC Regions Affected: MAAC, ECAR, NPCC
1529pk - 70
Conditions Prior to Blackout
 Electricity demand high but not unusually high

 Power transfer levels high, but within established limits and previous
operating conditions
 Planned outages of generating units in the affected area: Cook 2,
Davis Bess plant, East Lake 4, Sammis 3 and Monroe 1
 Reactive power supply problems in the regions of Indiana and Ohio
prior to noon
 Operators took actions to boost voltages
 voltages within limits
 System operating in compliance with NERC operating policies prior
to 15:05 Eastern Daylight Time
1529pk - 71
Blackout Started in Midwest
1529pk - 72
Sequence of Events
 The Midwest ISO (MISO) state estimator and real-time contingency analysis
(RTCA) software not functioning properly from 12:15 to 16:04
 prevented MISO from performing proper "early warning" assessments as the
events were unfolding
 At the First Energy (FE) Control Center, a number of computer software
problems occurred on the Energy Management System (EMS) starting at
14:14
 contributed to inadequate situation awareness at FE until 15:45
 The first significant event was the outage of East Lake generating unit #5 in
the FE system at 13:31:34
 producing high reactive power output
 voltage regulator tripped to manual on overexcitation
 unit tripped when operator tried to restore AVR
1529pk - 73
East Lake 5 Trip: 1:31:34 pm
ONTARIO
1529pk - 74
Sequence of Events cont'd
 Initial line trips in Ohio, all due to tree contact:

 Chamberlin-Harding 345 kV line at 15:05:41
 Hanna-Juniper 345 kV line at 15:32:03
 Star-South Canton 345 kV line at 15:41:35
 Due to EMS failures at FE and MISO control centers, no proper actions
(such as load shedding) taken
 Critical event leading to widespread cascading outages in Ohio and
beyond was tripping of Sammis-Star 345 kV line at 16:05:57
 Zone 3 relay operation due to low voltage and high power flow
 Load shedding in northeast Ohio at this stage could have prevented
cascading outages that followed
1529pk - 75
1529pk - 76
Sequence of Events
 Tripping of many additional 345 kV lines in Ohio and Michigan by Zone 3 (or
Zone 2 set similar to Zone 3) relays
 Tripping of several generators in Ohio and Michigan
 At 16:10:38, due to cascading loss of major lines in Ohio and Michigan, power
transfer from Canada (Ontario) to the US on the Michigan border shifted
 power started flowing counter clockwise from Pennsylvania through New York
and Ontario into Michigan
 3700 MW of reverse power flow to serve loads in Michigan and Ohio, which were
severed from rest of interconnected system except Ontario
 Voltage collapsed due to extremely heavy loadings on transmission lines
 Cascading outages of several hundred lines and generators leading to
blackout of the region
1529pk - 77
1529pk - 78
End of Cascade
Some Local Load

Interrupted
Areas Affected by the Blackout

Service maintained
in some area
1529pk - 79
Primary Causes of Blackout
(as identified by the US-Canada Outage Task Force)
1. Inadequate understanding of the power system requirements:
 First Energy (FE) failed to conduct rigorous long-term planning studies and
sufficient voltage stability analyses of Ohio control area
 FE used operational criteria that did not reflect actual system behaviour
and needs
 ECAR (East Central Area Reliability Council) did not conduct an
independent review or analysis of FE's voltage criteria and operating
needs
 Some NERC planning standards were sufficiently ambiguous that FE
could interpret them in a way that resulted in inadequate reliability for
system operation
cont'd
1529pk - 80
Causes of Blackout cont'd
2. Inadequate level of situation awareness:

 FE failed to ensure security of its system after significant unforeseen
contingencies
 FE lacked procedures to ensure that its operators were continually
aware of the functional state of their critical monitoring tools
 FE did not have adequate backup tools for system monitoring
3. Inadequate level of vegetation management (tree trimming)

 FE failed to adequately manage tree growth into transmission rights-
of-way
 resulted in the outage of three 345 kV lines and one 138 kV line
cont'd
1529pk - 81
Causes of Blackout cont'd
4. Inadequate level of support from the Reliability Coordinator

 due to failure of state estimator, MISO did not become aware of FE's
system problems early enough
 did not provide assistance to FE
 MISO and PJM (Regional Transmission operator) did not have in

place an adequate level of procedures and guidelines for dealing with
security limit violations due to a contingency near their common
boundary
1529pk - 82
September 23, 2003 Blackout of Southern
Sweden and Eastern Denmark
1529pk - 83
The Transmission Grid in the Nordic Countries
1529pk - 84
Blackout of 23 September 2003 in Southern Sweden
and Eastern Denmark
 Pre-disturbance conditions:
 system moderately loaded
 facilities out of services for maintenance:
 400 kV lines in South Sweden
 4 nuclear units in South Sweden
 3 HVDC links to Germany and Poland
 The first contingency was loss of a 1200 MW nuclear unit in South
Sweden at 12:30 due to problems with steam valves
 increase of power transfer from the north
 system security still acceptable
cont'd
1529pk - 85
and Eastern Denmark (cont'd)
 Five minutes later (at 12:35) a disconnector damage caused a double

busbar fault at a location 300 km away from the first contingency
 resulted in loss of a number of lines in the southwestern grid and two 900
MW nuclear units
 At 12:37, voltage collapse in the eastern grid section south of
Stockholm area
 isolated southern Sweden and eastern Denmark system from northern
and central grid
1529pk - 86
The Blackout in Southern Sweden and Eastern
Denmark, September 23, 2003
Hallsbe rg
Voltage Collapse
Glan
Kimstad
Kolstad
Strö mma Tenhult

Simpeva rp
Ho rre d
B reare d Alvesta
Nybro Isolated Subsystem
Hemsjö
Söderåse n
Barsebäck Line outages due to:

The fault in Horred
Maintenance work
The voltage collapse
1529pk - 87
The Blackout in Southern Sweden and Eastern
Denmark, September 23, 2003
The blacked-out area after the grid separation at 12.37
1529pk - 88
and Eastern Denmark cont'd
 The isolated system had enough generation to cover only about 30% of its
demand
 voltage and frequency collapsed within a few seconds, blacking out the area
 Impact of the blackout:
 loss of 4700 MW load in south Sweden
 1.6 million people affected
 City of Malmo and regional airports and rail transportation without power
 loss of 1850 MW in eastern Denmark
 2.4 million people affected
 City of Copenhagen, airport and rail transportation without power
 Result of an (n-3) contingency, well beyond "design contingencies"
1529pk - 89
September 28, 2003 Blackout of Italy
1529pk - 90
Italian System Blackout of 28 September 2003
 Predisturbance conditions (Sunday, 3:00 am):
 total load in Italy was 27,700 MW, with 3638 MW pump load
 total import from rest of Europe was 6651 MW
 Sequence of events:
 a tree flashover caused tripping of a major tie-line between Italy and Switzerland (Mettlen-
Lavorgo 380 kV line) at 03:01:22
 Sychro-check relay prevented automatic and manual reclosure of line due to the large
angle (42°) across the breaker
 resulted in an overload on a parallel path
 attempts to reduce the overload by Swiss transmission operators by network change was
not successful
 at 03:21 import by Italy was reduced by 300 MW but was not sufficient to mitigate the
overload of a second 380 kV line (Sils-Soazza), which tripped at 03:25:22 due to sag and
tree contact
1529pk - 91
Italian System Blackout of 28 September 2003 cont'd
 the cascading trend continued and the power deficit in Italy was such that the
ties to France, Austria and Slovania were tripped
 the outages left the Italian system with a power shortage of 6400 MW
 the frequency decay could not be controlled adequately by under-frequency
load shedding
 over the course of several minutes, the entire Italian System collapsed at
3:28:00
 The blackout affected about 60 million people

 total energy not delivered 180 GWh
 worst blackout in the history of Italy
 power was restored after 3 hours in the northern area and during the same day
for most of Italy
1529pk - 92
What Can We Do To Prevent
Blackouts?
1529pk - 93
Factors Impacting on System Security
Physical System
Integrated Generation,
Transmission, Distribution
System
Business Structure
Owning and operating entities; Financial
and contractual arrangements
Regulatory Framework
Governments, Reliability Councils
1529pk - 94
Comprehensive Approach to Enhancing System
Stability
 Impractical to achieve 100% reliability of power systems

 Good design and operating practices could significantly minimize the
occurrence and impact of widespread outages
 Reliability criteria: risk-based security criteria
 Improved protective relaying
 Robust stability controls
 Coordinated emergency controls
 Comprehensive stability assessment: analytical tools and models
 Real-time system system monitoring and control
 Wide-spread use of distributed generation
 Reliability Management System
 Good vegetation management
1529pk - 95
Reliability Criteria
 At present, systems designed and operated to withstand loss of any single

element preceded by single-, double-, or three-phase fault
 referred to as "N-1 criterion"
 formulated nearly 40 years ago after the 1965 blackout
 Need for using risk-based security assessment criteria
 consider multiple outages
 account for probability and consequences of instability
 Built-in overall strength or robustness best defense against
catastrophic failures !
1529pk - 96
Improved Protective Relaying
 State-of-the-art protective relaying for generating units and transmission

lines
 adaptive relaying
 Replacement of zone 3 and other backup relaying on important lines with
improved relaying
 Improved protection and control at power plants to minimize unit tripping
for voltage and frequency excursions
 Protective relay improvements to prevent tripping of critical elements on
overload
 control actions to relieve overload
1529pk - 97
Robust Stability Controls
 Greater use of stability controls

 excitation control (PSS), FACTS, HVDC, secondary voltage control
 multi-purpose controls
 multiple controllers
 Coordination, integration and robustness present challenges

 good control design procedures and tools have evolved
 Hardware design should provide

 high degree of functional reliability
 flexibility for maintenance and testing
 Industry should make better use of controls !
1529pk - 98
Emergency Controls for Extreme Contingencies
 Contingencies more severe than normal design contingencies

 multiple contingencies
 can occur anywhere on the system in any form
 Currently, emergency controls used to protect against some
 generator tripping, load shedding, dynamic breaking, controlled system
separation, transfer tap-changer blocking
 Need for a systematic approach to cover against all likely extreme
contingencies
1529pk - 99
"Defense Plan" Against Extreme Contingencies:
Coordinated Emergency Controls
 Judicious choice of emergency controls

 protection against different scenarios
 identification of scenarios based on past experience, knowledge of unique
characteristics of system, probabilistic approach
 Coordination of different emergency control schemes

 complement each other
 act properly in complex situations
 Response-based emergency controls should generally be preferred

 "self-healing" power systems
 Need for advancing this technology !
1529pk - 100
Examples of Response-Based Emergency Control
Schemes
1. Scheme for prevention of voltage collapse in Eastern Ontario

 fully automated and coordinated emergency controls for voltage
stability
2. Transient Excitation Boosting

 for enhancing transient (angle) stability of systems with dominant
interarea swing
1529pk - 101
Example 1: Prevention of Voltage Collapse in
Eastern Ontario
 Implemented in early 1980s to cope with delays in building 500 kV line
 Under high load conditions, loss of a major 230 kV line leads to voltage
collapse of Ottawa area
 A coordinated scheme consisting of fast line reclosure, load rejection,

shunt capacitor switching, and transformer ULTC blocking
1529pk - 102
Example 1: (cont'd)
The coordinated scheme:

a) Fast reclosure of major lines (1.3s)
 first line of defense
b) Load rejection (1.5s)

 9 blocks, 750 MW; armed by operator
 voltage/time dependent
c) Shunt capacitors switching (1.8 to 8.0s)

 36 banks in 17 TSs
d) Transformer ULTC blocking at 14 TSs

 unblocked when voltage recovers
1529pk - 103
Example 1: (cont'd)
 Coordination provided by appropriate selection of voltage and time

settings
 triggered by voltage drop magnitude and duration
 Following a contingency, depending on the severity (power flow, line

outage), only the required level of control action provided
1529pk - 104
1300 MW
1374 MW
1529pk - 105
1529pk - 106
Response-Based Emergency Controls
Example 2: Transient Excitation Boosting
 In situations with dominant interarea swing, PSS reduces excitation

after the first local mode swing
 Improvements in TS achieved by keeping excitation at ceiling until
highest composite swing
 increase in internal voltage
 increase in voltage also increases power consumed by area load
1529pk - 107
Block Diagram of TSEC Scheme
1529pk - 108
Effect of TSEC on Transient Stability
1529pk - 109
Example 2: (cont'd)
 Transient Excitation Boosting, TSEC, applied to four major plants in

Ontario:
 Nanticoke (4000 MW), Bruce A and B (6000 MW), Lennox (2000 MW)
 signal proportional to angle swing
 integrated with PSS and coordinated with terminal voltage limiter
 In effect, a nonlinear adaptive closed loop control

 may use local or remote signals
 imposes little duty on equipment
1529pk - 110
Comprehensive Stability Analysis Tools
 Powerful analytical tools have been developed capable of

comprehensive analysis for system design and operation:
 all forms of stability
 large systems with detailed models
 complementary use of time-domain and modal analysis
 automated procedures for considering large number of scenarios
 Industry gradually shifting to the use of new tools
 Lack of widespread understanding and appreciation for the use of

eigenvalue based modal analyses techniques
1529pk - 111
State-of-the-Art On-Line Dynamic Security
Assessment (DSA)
 Practical tools have been developed with the required accuracy, speed and
robustness
 a variety of analytical techniques integrated
 distributed hardware architecture using low cost PCs
 integrated with energy management system
 Capable of assessing rotor angle stability and voltage stability

 determine critical contingencies automatically
 security limits/margins for all desired energy transactions
 identify remedial measures
 The industry has yet to take full advantage of these developments !
1529pk - 112
Dynamic Security Assessment Tools Developed and
Used by Powertech for System Design and Operation
 Powerful set of complementary programs:

 flexible and detailed models
 alternative and efficient solution techniques
 Transient (Angle) Stability Assessment: TSAT
 Small-Signal (Angle) Stability Assessment: SSAT
 Voltage Stability Assessment: VSAT

 Frequency Stability Analysis: LTSP *
cont'd
* LTSP currently not maintained/supported

1529pk - 113
Powertech DSA Tools (cont'd)
 Automated procedures for:

 contingency screening and ranking
 consideration of a large number of scenarios
 stability limit search
 power flow dispatch
 determination of stability margins
 identifying remedial measures for maintaining stability and
increasing stability margins
 Significant savings in computation and engineering times
1529pk - 114
On-Line Voltage Stability Assessment Tool
(VSAT)
1529pk - 115
Key Elements of VSAT
 Interface with EMS; Model Initialization
 Contingency screening and selection
 Determination of secure operating region

 using static analysis
 Determination of remedial actions
 Fast time-domain simulation

 validation and checking
1529pk - 116
Contingency Selection Module
 Impractical to consider every conceivable contingency

 A limited number (typically 20) critical contingencies determined for detailed studies
 Performance Indices based on a few power flow solutions and reactive reserve not reliable
 A fast screening method used:
 based on exact margin to voltage collapse and full power flow solutions
 number of power flow solutions 1.2 to 2.0 times number of contingencies
 Supplemented with user-specified contingencies
1529pk - 117
Fig. 4 Automatic Critical Contingency Selection
1529pk - 118
Security Computation Module
 Engine for voltage stability analysis

 static analysis with detailed models
 Secure region is defined by a number of Coordinates (SRCs)
 key system parameters: MW generation, area load, interface transfers, etc.
 Voltage security determined by
 voltage stability margin
 MVAr reserves of key reactive sources
 post-contingency voltage decline
 Modal analysis of powerflow Jacobian matrix identifies areas prone to
instability
 Specialized powerflow dispatcher and solver to quickly search for
stability limit
1529pk - 119
Modelling:
 generator capability curves
 governor response, economic dispatch, AGC
 nonlinear loads
 control of ULTCs, switched shunts, etc.
Inputs and Outputs:

 Inputs
 list of contingencies produced by screening and ranking (+user defined)
 base case powerflow from state estimator
 definition of SCRs
 voltage security criteria and definition of parameter of stress
 Output
 secure region in secure region space
1529pk - 120
Secure Operating Region
1529pk - 121
Secure Operating Region
1529pk - 122
Remedial Measures Module
 Determines necessary remedial measures to

 ensure sufficient stability margins
 expand the secure region
 Preventative control actions:
 taken prior to a contingency
 caps/reactor switching, generation redispatch, voltage rescheduling
 Corrective (emergency) control actions:
 applied following a contingency
 load shedding, generator runback, transformer tap changer blocking
 Ranking of each remedial measure using:
 sensitivity analysis
 user-defined priorities
1529pk - 123
Ranking and Applying Remedial Measures
 Objective is to identify the most effective remedial measures to give the

desired stability margin
 Obtain solved power flow case for the most severe contingency
 gradually introduce the effect of the contingency
 bus injection compensation technique
 Compute the sensitivities of reactive power (or bus voltage) to different
control measures
 rank the remedial measures
 Apply controls one at a time in order of ranking until power flow solves for
the most severe contingency
1529pk - 124
Expanding the Secure Region: Remedial Measures
1529pk - 125
Fast Time-Domain Simulation Module
 Determines the essential dynamic phenomena without step-by-step

numerical integration
 when chronology of events significant
 for validating the effect of remedial measures
 Focuses on the evolution of system dynamic response driven by slow

dynamics
 transformer tap changers, field current limiters, switched caps
 Captures the effects of fast dynamics by solving associated steady state

equations
1529pk - 126
Mathematical Formulation
 The complete set of differential/algebraic equations of a power system has
the following general form:
  f  X, V , Z 
X
YV  I  X , V , Z 
Where:
X = state vector
V = bus voltage vector
I = current injector vector
Y = network admittance matrix
Z = variables associated with the slow
control devices including ULTCs, loads, switchable reactors
and capacitors, and field current limiters
1529pk - 127
Mathematical Formulation
 At each equilibrium point, Z=Zi and the system operating condition is

obtained by solving the following set of nonlinear algebraic equations:
0  f  X, V, Zi 
YV  I  X, V, Zi 
 As time progresses, the slow control devices operate and the values of Z
change. The above set of nonlinear algebraic equations is solved every
time the values of Z change.
1529pk - 128
1529pk - 129
VSAT Structure
Powerflow
Solver
1529pk - 130
Transient Stability Assessment Tool (TSAT)
1529pk - 131
Transient Stability Assessment (TSA)
 Time-domain simulations essential

 modeling detail and accuracy
 Sole dependence on time-domain simulations has severe limitations
 high computational burden
 no stability margin/sensitivity information
 requires considerable human interaction
 Supplementary techniques for speeding up and automating overall
process
 Methods available for deriving useful indices
 Transient Energy Function (TEF)
 Signal Energy Analysis
 Extended Equal Area Criterion (EEAC)
1529pk - 132
A Practical Tool for TSA
 Overall architecture similar to that of VSA
 Time-domain program, with detailed models and efficient solution
techniques, forms simulation engine
 EEAC used for screening contingencies, computing stability margin,
stability limit search, and early termination of simulation
 “Prony analysis” for calculation of damping of critical modes of oscillation
 A powerflow dispatcher and solver for finding the stability limit
 a fully automated process
 No modeling compromises;
can handle multi-swing instability
1529pk - 133
EEAC
 Integrates the dynamic response in the multimachine space, and maps the
resultant trajectory into a set of one-machine-infinite-bus planes
 By applying complementary cluster center of inertia (CCCI) transformations
 Keeps all dynamic information in the multimachine space
 Stability analysis can be quantitatively performed for the image OMIB
systems
 Has the same accuracy and modeling flexibility
 Fast, quantitative
1529pk - 134
EEAC
 Loss of transient stability in a power system always starts in a binary splitting of generators:
 Critical cluster of generators
 Rest of the system
 At any given point in the

time-domain trajectory of
the system, the system
can be visualized as a
one-machine-infinite-bus
(OMIB) system
1529pk - 135
EEAC
 The classical equal area criterion can be extended to the visual OMIB system
Stability margin of the system is defined as
 Ad  Aa
 100 x if the system is stable  Ad  Aa 
Ad

A  Aa
100 x d if the system is unstable  Aa  Ad 
 Aa
Thus, -100   , and
if the system is stable

 if the system is unstable
 can be used as a stability index
1529pk - 136
Use of EEAC Theory
 Contingency screening
 stability margin gives an indication of the relative severity
 Corrective measures for maintaining secure system operation
 critical cluster of generators (CCG) provides valuable information
 Power transfer limit search
 stability limit can be determined in four iterations using stability margin
 each iteration involves a detailed simulation and computation of stability index
1529pk - 137
Results - Test System
System description
 BC Hydro system
 1430 buses
 186 generators
 4 HVDC links
Interface
 GMS and PCN output

 Base case transfer = 3158 MW
Contingency
 Three phase fault at GMS 500 kV bus

 Tripping of one of two 500 kV lines from
GMS to WSN
1529pk - 138
Limit Search Results
1529pk - 139
Speed Enhancement: Parallel Processing
 Code parallelization
 differential equations easily parallelized, but not network equations
 speed-ups limited by serial slowdown effect
 up to 7 times speed-up can be achieved with 20-30 processors
 not an effective way
 Conventional serial computers offer much faster computational per-CPU
 For multiple contingencies
 perform initialization only once
 run contingencies on multiple processors - one processor per contingency
1529pk - 140
TSAT Structure
Solved Powerflow
+
Dynamic Data
Contingency Screening & Full

Transaction Ranking (EEAC) Contingency List
Definitions
Must Run
Contingencies
Powerflow
Dispatcher
Time-Domain
Simulation
Increase
Transfer
Stability
Indices
No
Security Limit?
Yes
Sufficient No Remedial
Margin? Measures
Yes
STOP
1529pk - 141
Computational Performance of DSA
 Target cycle time from capture of state estimation to completion of

security assessment for all specified transactions:
 20 minutes
 TSA and VSA functions performed in parallel

 distributed processing on separate CPUs
 This can be readily achieved with low cost PCs
1529pk - 142
Computational Speed of DSA (cont'd)
Power System model with 4655 buses, 156 generators, using 1.7 GHz,
Pentium 4 PC with 256 MB memory
Voltage Stability Assessment:
- screening 300 contingencies 20.0 secs
- detailed security analysis 1.2 secs
with 20 critical contingencies
- one transfer limit search 12.0 secs
Transient Stability Assessment:

- screening 100 contingencies 75.0 secs
- 10 second simulations with 75.0 secs
10 critical contingencies
- one transfer limit search 120.0 secs
- total time for complete assessment < 5 mins
1529pk - 143
Future Trends in DSA: Intelligent Systems
 Knowledge base created using simulation of a large number cases and system
measurements
 Automatic learning, data mining, and decision trees to build intelligent systems
 Fast analysis using a broad knowledge base and automatic decision making
 Provides new insight into factors and system parameters affecting stability
 More effective in dealing with uncertainties and large dimensioned problems
 We just completed a PRECARN project: "POSSIT"
1529pk - 144
DSA Using Intelligent Systems
1529pk - 145
Real-Time Monitoring and Control:
An Emerging Technology
 Advances in communications technology have made it possible to

 monitor power systems over a wide area
 remotely control many functions
 Research on use of multisensor data fusion technology
 process data from different monitors, integrate and process information
 identify phenomenon associated with impending emergency
 make intelligent control decisions
 A fast and effective way to predict onset of emergency conditions and take
remedial actions
The ultimate "self-healing" power system !
1529pk - 146
Distributed Generation (DG)
 Offer significant economic, environmental and security benefits
 Microturbines
 small, high speed power plants
 operate on natural gas or gas from landfills
 Fuel Cells
 combines hydrogen with oxygen from air to generate electricity
 hydrogen may be supplied from an external source or generated inside fuel cell by
reforming a hydrocarbon fuel
 Not vulnerable to power grid failure due to system instability or natural calamities
 protection and controls should be designed so that units continue to operate when
isolated from the grid
1529pk - 147
Reliability Management System
 Roles and responsibilities of individual entities

 well chosen, clearly defined and properly enforced
 Coordination of reliability management
 Need for a single entity with overall responsibility for security of entire
interconnected system
 real-time decisions
 System operators with high level of expertise in system stability

 phenomena, tools
1529pk - 148
Summary
1. The new electricity supply industry presents increasing challenges for stable and
secure operation of power systems
2. State-of-the-art methods have advanced our capabilities significantly
 comprehensive stability analysis tools
 automated tools for system planning/design
 on-line Dynamic Security Assessment (DSA)
 coordinated design of robust stability controls
3. Industry is yet to take full advantage of these developments
cont'd
1529pk - 149
Summary (cont'd)
4. Future directions will be to explore new techniques which can better

deal with growing uncertainties and increasing complexities of the
problem
 risk-based security assessment
 intelligent systems for DSA
 "self-healing" power systems
 real-time monitoring and control
5. Wide-spread use of distributed generation could be a cost effective

means of minimizing the impact of power grid failures
1529pk - 150

Power System Stability

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Power System Stability

Uploaded by

Copyright:

Available Formats

Tutorial

Power System Stability in the New Industry

Dr. Prabha S. Kundur

1. Brief Introduction to Power System Stability

2. Examples of Major System Blackouts Caused by Different Forms of Instability

3. Challenges to Secure Operation of today's Power Systems

4. Major System Blackouts in 2003 and 2004

5. Comprehensive Approach to Enhancing Power System Stability

 Refers to continuance of intact operation of power system, following a

 Power System Stability denotes the ability of an electric power system,

 Stability is a condition of equilibrium between opposing forces:

Ref: IEEE/CIGRE TF Report, "Definition and Classification of Power System Stability",

 On the other hand, if the system is unstable, it will result in a run-away or

 An unstable system condition could lead to cascading outages, and a shut-down of a

 Classification into various categories greatly facilitates:

 Classification is based on the following considerations:

 We should always keep in mind the overall stability !

Short Term Short

 Ability of interconnected synchronous machines to remain in synchronism

 A fundamental factor is the manner in which power outputs of synchronous

 Term traditionally used to denote large-disturbance angle stability

 Small-Signal (or Small-Disturbance) Stability is the ability of a power system

 A system experiences voltage instability when a disturbance, increase in load

 Short-term voltage stability involves dynamics of fast acting load

 Long-term voltage stability involves slower acting equipment such as tap-

 Ability to maintain steady frequency within a nominal range following a

 Instability that may result occurs in the form of sustained frequency

 In a small "island" system, frequency stability could be of concern for

 In a large interconnected system, frequency stability could be of

 Depends on the ability to restore balance between generation and load

 Generally, frequency stability problems are associated with

1. November 9, 1965 blackout of Northeast U.S. and Ontario

2. April 19, 1972, blackout of Eastern Ontario

3. July 2, 1996 disturbance of WSCC (Western North American

4. August 10, 1996 disturbance of WSCC system

5. March 11, 1999 Brazil blackout

 Problem began at 5:16 p.m.

 Nearly 30 million people were without power for about 13 hours

 The initial event was the operation of a backup relay (zone 3) at

 Prior to opening of Q29BD, the five circuits were carrying

 Loading on Q29BD was 361 MW at 248 kV;

 Opening of circuit Q29BD resulted in sequential tripping of the remaining four

 Power flow reversed to New York: total change of 1700 MW

 Generators in Western New York and Beck GS lost synchronism, followed by

 After about 7 seconds from the initial disturbance

 Northeast Power Coordinating Council (NPCC) formed in January 1966

 Other eight RRCs formed in the following months

 National/North American Electric Reliability Council (NERC) established in 1968

 Detailed reliability criteria were developed;

 Many of these developments have had an influence on utility practices worldwide;

 P Relays on Niagara Ties

 trip Niagara ties to NY when P exceeds set value;

 Underfrequency load shedding (UFLS) throughout the

 beginning of the use of UFLS by the industry

 Started in Wyoming and Idaho area at 14:24:37

 Loads were high in Southern Idaho and Utah;

 Heavy power transfers from Pacific NW to California

 LG fault on 345 kV line from Jim Bridger 2000 MW plant in Wyoming to

 Faulty relay tripped 230 kV line in Eastern Oregon

 Voltage decay in southern Idaho and slow decay in central Oregon

 Rapid voltage decay in Idaho and Oregon

 Two seconds later, Pacific intertie lines separated

 Cascading to five islands 35 seconds after initial fault

 2.2 million customers experienced outages; total load lost 11,900 MW