1. 1 “Cascading Events and How to Prevent Them” The International Meeting of VLPGO WG#1 October 25, 2005

2. 2 Background Recurrent Cascading Outage Worldwide after Market Liberalization 1 st International Meeting of Very Large Power Grid Operators(VLPGO) - October 25-26, 2004, - Philadelphia, USA - Objectives : To find and share common concerns in “Maintaining Reliability” To exchange good/bad experiences and the best existing practices To cooperate on developing necessary measures Set up three working groups WG#1 :Cascading Events and How to Prevent Them(Lead: TEPCO) WG#2 :EMS Architectures for 21th Century (Lead : PJM/MISO) WG#3 :Advanced Decision Support Tools (Lead : RTE)

3. 3 Objectives and Means WG-1 surveyed the ways to prevent cascading events through a questionnaire to identify: Common causes and mechanisms Existing measures to be recommended New technologies to be developed in the future

4. 4 Contributors to WG#1 Mr.Hideaki TANAKA TEPCO JAPAN (Coordinator) Mr.Masanobu KAMINAGA TEPCO JAPAN Mr.Michel KORMOS PJM USA Dr.Yuri MAKAROV CAISO USA Mr.Temistocle BAFFA SCIROCCO GRTN ITALY Mr.Jean Michel TESSERON RTE FRANCE Mr.Ian WELCH National Grid UK

5. 5 Contents 1. Scope of Survey 2. Survey Results 3. Conclusion

6. 6 1. Scope of Survey

7. 7 Power System States and Transition

8. 8 Scope of each WG De-Centralized System Centralized System MonitoringWG-3 *New technologies used by operators for decision making WG-2 *Standardization of architecture platform for EMS Control *Three WGs focused on technical issues excluding institutional issues, such as power market design. WG-1

9. 9 Issues associated with Cascading Events

10. 10 2. Survey Results

11. 11 Features of Very Large Power Grid NG PJM ★ ★ ★ ★ ★ RTE ★ TEPCO CAISO GRTN OrganizationSystem Capacity [GW] Highest Voltage [kV] TEPCO 64 500 GRTN 54 380 National Grid 54 400 PJM 133 765 CAISO 48 500 RTE 83 400

12. 12 System Configuration and Interconnection with Neighboring Systems Organization System Configuration Interconnections EHVHV,MV TEPCOMesh [ 63 kA] RadialA 500kV AC link Two BTBs GRTNMesh [ 50 kA] RadialEight 380kV AC links A 500kV DC link A 200kV DC link National GridMesh [ 63 kA] RadialAC Link DC link PJMMesh [ 63 kA] Mesh248 AC links CAISOMesh [ 63 kA] MeshFive AC links RTEMesh [ 63 kA] Mesh40 AC links A 270kV DC link [Maximum Short Circuit Current]

13. 13 Power Flow Level and Critical Stability Issues Maximum Power Flow/ SIL Critical Issues Small- disturbance TransientFrequencyVoltageThermal Overloading TEPCO 2.0 ○ (1) ○ (2) ○ (3) GRTN 1.5 ○ (3) ○ (1) ○ (2) National Grid ○○○ PJM ○○○ CAISO○○○ RTE○ (2) ○ (3) ○ (1) (Priority)

14. 14 Overview of Past Cascading Events Date/ Area MW Lost, Duration Causes (Trigger) Critical Phenomenon Nov. 1978/ France 30,000MW 7 hr -Major 400kV lines trips due to overloading -Voltage collapse -System separation Aug. 1981/ UK 1,900MW 2.5 hr -Loss of 3 400kV circuits and lower voltage interconnection -System separation Jan. 1987/ France 8,000MW 3 hr -Multiple generator trips-Voltage collapse Jul. 1987/ Tokyo 8,000MW 4 hr -Insufficient VAR supply for high rate of load pickup -Voltage collapse Jul./Aug. 1996/ Western US 11,850MW ? hr 28,000MW 9 hr -Loss of multiple lines and loss of critical generation -Voltage collapse -System separation -Power oscillation

15. 15 Overview of Past Cascading Events Date/ Area MW Lost, Duration Causes (Trigger) Critical phenomenon Aug. 2003 Northern US 61,800MW 42 hr -Multiple line trips-Voltage collapse -System separation Aug. 2003 London 724MW 0.7 hr -Incorrect operation of a backup relay -Overloading Sep. 2003 Italy 20,000MW 20 hr -Multiple EHV line trips-Voltage collapse (before system separation) - Frequency collapse (after system separation) -Power oscillation -System separation Sep. 2003 Scandia 6,550MW 6.5 hr -Scrum of a Nuke plant -Double-bus fault -Voltage collapse -Power oscillation May. 2005 Moscow 2,500MW 32 hr -CT explosion -Multiple Tr explosion&fire -Overloading

16. 16 Common Causes of Recent Cascading Outages Deregulation of Electricity Market: Monopoly to Competition Priority to Market Mechanism More Players Less Maintenance Delay in Network Enhancement Change of System Operation Rules Increase of Inter-regional (National) Power Exchange Complicated and Enlarged Power Grid Information to be Handled Uncertainty in Operating Condition Fault Frequency Difficulty in Responding to Abnormal Situation Heavier Duty in Accommodating Electricity Transaction More.. (To Operators) (To Interconnected Power System)

17. 17 Mechanism of Cascading Outages-1 -Unexpected heavy loading -Unscheduled generation outage -Single fault (tree touching etc.) -Combination of above events -Delay in grasping the situation -Delay in communication with neighboring operators -Delay in taking mitigation action Trigger Delay in Initial Action  Impact of Market Liberalization Cost Reduction Wide-area Heavy Power Transaction  Inappropriate On-line Monitoring Need for Sophisticated On-line Monitoring System Need for On-line Contingency Analysis Tool Need for Automatic Preventive Control

18. 18 Mechanism of Cascading Outages-2 -Transmission Lines -Transformers -Generators Protective Action Need for Time Coordination with Safety Nets, such as UFLS and UVLS -Overloading -Voltage Collapse -Power Oscillation -Loss of Synchronism -Frequency Declining Cascading Events: Alert to Emergency Need for Enhancement of Emergency Control, so called ‘Safety Net,’ including ‘Islanding’

19. 19 1) On-line state estimator(SE) is commonly used *Maximum Capacity: 7400 nodes, 2500 generators, every one minute 2)On-line contingency analysis is also commonly implemented, both for ‘voltage instability’ and ‘thermal overloading’ 3) On-line direct monitoring of power system oscillation has started in some countries. In the US and EU, GPS-based PMU (Phasor Measurement Unit) is applied to monitor the phase. 1) To improve the accuracy of SE by using the PMU in combination with the conventional SE 2) To extend the scope of contingency evaluation into phase angle stability, in particular transient stability 3) To estimate the frequency/power regulation performance of the system Existing Countermeasures (Security Monitoring)

20. 20 1) Power System Stabilizer is a common tool to prevent small-signal instability. 2) High-speed re-closing including multi-pole re-closing is adopted in order to improve transient stability, while aiming to put the network back to its initial state. 3) Automatic Generator Control is commonly used to maintain system frequency. 4) A wide variety of de-centralized automatic control systems are used to prevent the transition to the alert state, as well as to lighten operators’ burden. 5) In France, a centralized automatic control system called the Secondary Voltage Control System, has been in operation at the regional level, the purpose of which is to get better control performance. 1) To select a centralized system or de-centralized system appropriately in accordance with system features 2) To develop a sophisticated algorithm that can provide the operators with information on how to prevent the transition from “Alert” to “Emergency”. Existing Countermeasures (Preventive Control)

21. 21 Existing Countermeasures(Emergency Control) 1) Several types of ‘emergency controls’ have been developed as a ‘safety net’ and are in operation. They are categorized into the following three categories: OFLS, SPS (Generators, Pumped-Storage Units etc.) UFLS, UVLS, SPS (Loads), [Blocking of Tap Changers] Islanding 1) To keep the interconnection as long as possible, even when an emergency occurs, ‘Time coordination’ between the equipment protection system and emergency control systems must be examined. 2) To consider quicker restoration when designing the safety net.

22. 22 1) The following systems, mainly for “Monitoring “, are now under development.  On-line transient stability assessment in the US and Japan  PMU-based small-signal stability monitoring system in the US.  A wide area monitoring system, which covers all issues regarding cascading events and includes the centralized voltage control system in Italy.  Monitoring of Generator performance and of f/P power system performance in France. 2)Currently, no emergency control system is being developed. Future Countermeasures (Under Development)

23. 23 Other Issues 1) Institutional Measures ◆ Give TSO operators more authority ◆ Establish a reliability standard applied to all relevant stakeholders ◆ Contract among stakeholders who comply with the reliability standard ◆ Establish strong coordination between TSOs in different time frames 2)Other Measures ◆ Aim Operator Training Simulator at: Improving the knowledge and skills of the individuals Developing the operator’s tolerance for psychological stress and capability of coping with abnormal conditions ◆ Risk indices used to set adequate reserve margins or to allow operators know the necessity of load shedding

24. 24 3. Conclusion

25. 25 Summary (Existing Measures) 1)There were several types of automatic centralized/de-centralized systems for “Monitoring”, “Preventive Control”and “Emergency Control”, respectively. Monitoring : ◎ Preventive Control : ○ ○ Emergency Control : ◎ 2) Each VLPGO will be able to select and employ the most appropriate systems from the “Seasoned Best Practice Menu”.

26. 26 1)Since the use of “Emergency control ” is a last resort measure, we should focus future R&D mainly on “Monitoring ” and “Preventive Control ”, which are tools used “upstream” of cascading events. 2)Upon reviewing survey results, we have identified the following R&D topics to be further addressed. Summary (Future R&D Topics)

27. 27 On-line Dynamic Assessment (PJM,TEPCO) - On-line High Speed Screening On-line System stability(Steady state/Dynamic) Monitoring (CAISO) - PMU application - Eigenvalue calculation Wide Area Measurement (GRTN) Generator and System f/P Performance Monitoring(RTE) On-line Corrective Action(switching or re-dispatching) Indicating System (RTE)

28. 28 3) In the process of R&D, we should take into account : Up-to-date IT technologies Parallel computing techniques Advanced algorithms for on-line analyses Standardization of software and architecture.

29. 29 Recommendation on 2006 action plan Develop a “Comprehensive Survey Paper” on the ways to prevent cascading blackouts by the next VLPGO meeting for presentation at an International Conference such as CIGRE 1. Recruit new members and dispatch questionnaire. 2. As a new and last aspect, add a “Restoration”. 3. Extend the survey area to the papers published by the PGOs with a capacity of less than 50GW.

30. 30 Thank you for your attention