ML12334A544

Official Exhibit - NYS000088-00-BD01 - the Effects of Integrating Wind Power on Transmission System Planning, Reliability, and Operations, Report on Phase 2, System Performance Evaluation (March 2005) Excerpted: Pp. v-x, 1.1-2.16

ML12334A544
Person / Time
Site:	Indian Point
Issue date:	12/14/2011
From:	Bai X, Clark K, Jordan G, Miller N, Zimberlin J GE Power Systems Energy Consulting
To:	Saintcross J NRC/SECY, State of NY, Energy Research & Development Authority
SECY RAS
References
RAS 21536, 50-247-LR, 50-286-LR, ASLBP 07-858-03-LR-BD01
Download: ML12334A544 (31)

Text

United States Nuclear Regulatory Commission Official Hearing Exhibit NYS000088 In the Matter of:

Entergy Nuclear Operations, Inc.

(Indian Point Nuclear Generating Units 2 and 3)

Submitted: December 14, 2011 ASLBP #: 07-858-03-LR-BD01 Docket #: 05000247 l 05000286 EXCERPT Exhibit #: NYS000088-00-BD01 Identified: 10/15/2012 Admitted: 10/15/2012 Withdrawn:

Rejected: Stricken:

Other:

THE EFFECTS OF INTEGRATING WIND POWER ON TRANSMISSION SYSTEM PLANNING, RELIABILITY, AND OPERATIONS Report on Phase 2:

System Perfonnance Evaluation Prepared for:

THE NEW YORK STATE ENERGY RESEARCH AND DEVELOPMENT AUTHORITY AlbanY,NY John Saintcross Senior Project Manager Prepared by:

GE ENERGY ENERGY CONSULTING Richard Piwko, Project Manager Xinggang Bai Kara Clark Gary Jordan Nicholas Miller Joy Zimberlin March 4, 2005 OAGI0000191_001

Table of Contents INTRODUCTION ................................................................................................................. 1.1

1.1 BACKGROUND

............................................................................................................... 1.1 1.2 WIND GENERATION SCENARIO ...................................................................................... 1.2 1.3 TIMESCALES FOR POWER SYSTEM PLANNING AND OPERATIONS ..................................... 1.3 1.4 TECHNICAL ApPROACH .................................................................................................. 1.7 1.4. 1 Forecast Accuracy ................................................................................................. 1.7 1.4.2 Wind and Load Variability ...................................................................................... 1. 7 1.4. 3 Operational 1.7 1.4.4 Effective 1. 8 1.5 DATA ........................................................................................................................... 1.8 2 EXECUTIVE SU MMARY ..................................................................................................... 2.1 2.1 STUDY SCENARIO FOR WIND GENERATION ..................................................................... 2.1 2.2 IMPACT ON SYSTEM PLANNING ...................................................................................... 2.2 2.2.1 NYISO System Reliability Impact Study (SRIS) ..................................................... 2.2 2.2.2 NYSRC Reliability Rules for Planning and Operation ............................................ 2.2 2.2.3 Generation Interconnection Requirements ............................................................ 2.2 2.2.4 Future Interconnection Options .............................................................................. 2.4 2.3 IMPACT ON SYSTEM OPERATIONS .................................................................................. 2.5 2.3. 1 Forecasting and Market Operations ...................................................................... 2.7 2.3.2 Hourly Variability .................................................................................................... 2.9 2.3.3 Load-Following ...................................................... ............................................... 2.10 2.3.4 Regulation ............................................................................................................ 2.11 2.3.5 Spinning Reserves ............................................................................................... 2.12 2.3.6 System Operating Costs ...................................................................................... 2.12 2.3.7 Energy Displacement and Emission Reductions ................................................. 2.13 2.3.8 Transmission Congestion .................................................................................... 2.14 2.4 IMPACT ON SYSTEM RELIABILITY .................................................................................. 2.15 2.4. 1 Effective Capacity of Wind Generators ................................................................ 2.15 2.4. 2 System Stability ................................................................................................... 2. 15

2.5 CONCLUSION

S ............................................................................................................ 2.16 3 FORECAST ACCURACy .................................................................................................... 3.1 3.1 VARIABILITY AND PREDICTABILITY .................................................................................. 3.1 3.2 DAY-AHEAD FORECASTING ............................................................................................ 3.2 3.2.1 Day-Ahead Load Forecasting ................................................................................ 3.2 3.2.2 Day-Ahead Wind Forecasting ................................................................................ 3.3 3.2.3 Discussion of Timing .............................................................................................. 3.4 3.3 DAY-AHEAD FORECASTING ERROR ANALYSIS ................................................................ 3.5 3.3. 1 Day-Ahead Forecasting Error Analysis for January 2001 ...................................... 3.6 3.3.2 Day-Ahead Forecasting Error Analysis for Multiple Months ................................ 3.10 3.4 HOUR-AHEAD FORECASTING ....................................................................................... 3.16 3.5 CENTRALIZED VERSUS DECENTRALIZED FORECASTING ................................................ 3.21

3.6 CONCLUSION

S AND RECOMMENDATIONS ...................................................................... 3.22 3.6. 1 Conclusions .......................................................................................................... 3.22 3.6.2 Recommendations ............................................................................................... 3.23 4 HOURLY PRODUCTION SIMULATION ANALySiS .......................................................... 4.1

4.1 INTRODUCTION

.............................................................................................................. 4.1 4.1.1 Description of Cases .............................................................................................. 4.1 4.2 ANALYSIS OF RESULTS .................................................................................................. 4.2 4.2.1 Energy Displacement ............................................................................................. 4.3 GE Energy 3/04/05 OAGI0000191 005

4.2.2 Emission Reductions ............................................................................................. 4.6 4.2.3 Transmission Congestion ...................................................................................... 4.8 4.2.4 Economic Impact ................................................................................................. 4.10 4.3

SUMMARY

.................................................................................................................. 4.14 5 WIND AND LOAD VARIABILITY ........................................................................................ 5.1 5.1 ANNUALANDSEASONAL VARIABILITY ............................................................................. 5.1 5.2 HOURLY VARIABILITY .................................................................................................... 5.2 52.1 Daily Load 5.2 5.2.1.1 Diurnal Characteristics.

5.2.1.2 Geographic Characteristics ......................................................................................... 5.3 52.2 Statistical Analysis of Hourly Load Variability ........................................................ 55 52.3 Temporal Nature of Wind Penetration ................................................................... 56 52.4 Hourly Variability of Wind ....................................................................................... 58 52.5 Combined Load and Wind Variability ..................................................................... 59 5.2.5.1 Trends in Hourly Variability ....................................................................................... 5.11 52.6 Time of Day Trends ............................................................................................. 512 5.3 FIVE-M INUTE VARIABILITy ........................................................................................... 5.14 5.4 SiX-SECOND VARIABILITY ............................................................................................ 5.16 5.4 1 AGC Performance ................................................................................................ 5 19 5.42 One-Second Wind Variability ............................................................................... 522 5.4.2.1 One-Second \Mnd Variability Of One \Mnd Farm ..................................................... 5.23 5.43 Coincidence of Load and Wind Variability ........................................................... 527

5.5 CONCLUSION

S AND RECOMMENDATIONS ...................................................................... 5.29 55 1 Conclusions .......................................................................................................... 5.29 552 Recommendations ............................................................................................... 5.29 6 OPERATIONAL IMPACTS .................................................................................................. 6.1 6.1 QSS ANALYSIS ............................................................................................................ 6.1

6. 1. 1 Approach ................................................................................................................ 6. 1 6.1.1.1 Data ............................................................................................................................ 6.3 6.1.1.2 Study Scenarios .......................................................................................................... 6.6 6.1.2 Results ................................................................................................................. 6.12 6.1.2.1 Large-Scale Wind and Load Changes ...................................................................... 6.12 6.1.2.2 Wind Generation Variability ...................................................................................... 6.19 6.1.2.3 Active Power Control ................................................................................................ 6.20 6.2 STABILITY ANALYSIS ................................................................................................... 6.23 6.2. 1 Approach .............................................................................................................. 6.23 6.2.1.1 Data ..........................................................................................................................6.24 6.2.1.2 Study Scenarios ........................................................................................................6.27 6.2.2 Results ................................................................................................................. 6.28 6.2.2.1 Wind Farm Performance ........................................................................................... 6.28 6.2.2.2 System Performance ................................................................................................ 6.40

6.3 CONCLUSION

S ............................................................................................................ 6.41 7 EFFECTIVE CAPACiTY ...................................................................................................... 7.1

7.1 INTRODUCTION

.............................................................................................................. 7.1 7.2 WINDANDLoADSHAPES .............................................................................................. 7.1 7.3 LOLP ANALYSIS ........................................................................................................... 7.7

73. 1 2001 and 2002 Analysis ......................................................................................... 78 73.2 UCAP calculations ................................................................................................. 7. 9 7.3.2.1 Comparison to Phase 1 Results .............................................................................. .7.1 0

7. 3. 3 Approximate Techniques ..................................................................................... 7. 11 7.4

SUMMARY

.................................................................................................................. 7.16 8 SUGGESTED CHANGES TO PLANNING AND OPERATING PRACTiCES .................... 8.1 8.1 NYISO PLANNING PRACTICES AND CRITERIA ................................................................ 8.1 GE Energy vi 3/04/05 OAGI0000191 006

8.11 Impact of Wind Generation on Steady-State Analysis ........................................... 8.2 8.1.2 Impact of Wind Generation on Stability Analysis ................................................... 8.2 8.2 NERC, NPCC, AND NYSRC RELIABILITY CRITERIA ...................................................... 8.3 8.3 NYISO TRANSMISSION RELIABILITY AND CAPACITY REQUiREMENTS ............................... 8.5 84 ANCILLARY SERViCES ................................................................................................... 8.6 8.5 NYISO MARKET DESIGN ............................................................................................... 8.6 9 REFERENCES .................................................................................................................... 9.1 Appendices APPENDIX A DATA

SUMMARY

.............................................................................................. A-1 APPENDIX B DAY AHEAD FORECAST ANALYSIS FOR 11 MONTHS................................ B-1 APPENDIX C HOURLY VARIABILITY STATISTICS BY MONTH .......................................... C-1 APPENDIX D WIND TURBINE-GENERATOR (WTG) MODELS ............................................ D-1 APPENDIX E OTHER DYNAMIC MODELS ............................................................................. E-1 APPENDIX F MAPS' PROGRAM DESCRiPTION .................................................................. F-1 APPENDIX G MARS PROGRAM DESCRIPTION ................................................................... G-1 Table of Figures Figure 1.1 New York Control Area Load Zones .......................................................................... 1.3 Figure 1.2 Time Scales for System Planning and Operation Processes .................................... 1.5 Figure 1.3 Wind Variability and Impact on System Operation Processes .................................. 1.6 Figure 2.1 Standard Deviation of Day-Ahead Forecast Errors ................................................... 2. 7 Figure 2.2 Summer Morning Load Rise - Hourly Variability ....................................................... 2.10 Figure 2.3 Five-Minute Statewide Variability ............................................................................ 2.11 Figure 24 Duration Curve of Hourly Flows on UPNY-SENY Interface .................................... 2.14 Figure 3.1 Variability and Predictability of Non-dispatchable Generating Resources ................. 3.2 Figure 3.2 Wind Forecasting Accuracy for an individual Wind Farm ........................................... 34 Figure 3.3. Timeline for Day-Ahead Forecasting .......................................................................... 3.5 Figure 34 Day-Ahead Forecasts vs. Actual Hourly for January 2001 ......................................... 3.6 Figure 3.5 Day-Ahead Forecast Errors for January 2001 ............................................................ 3.7 Figure 3.6 Day-Ahead Error Duration Curve for January 2001 .................................................... 3.8 Figure 3.7 Mean Absolute Error (MAE) for Individual Wind Farms Forecasts - January 2001 .. 3.10 Figure 3.8 Standard Deviation of Day Ahead Forecast Errors ................................................... 3.11 Figure 3.9 Day-Ahead Positive Forecast Error Frequency ........................................................ 3.12 Figure 3.10 Day-Ahead Negative Forecast Error Frequency .................................................... 3.12 Figure 3.11 Positive Energy Error for Day-Ahead Forecasts .................................................... 3.13 Figure 3.12 Negative Energy Error for Day-Ahead Forecasts ................................................... 3.13 Figure 3.13 Distribution of Forecast Energy Errors ................................................................... 3.16 Figure 3.14. Day-Ahead and Hour-Ahead Wind Forecast and Actual Wind for January 2001 .. 3.17 Figure 3.15 Day-Ahead and Hour-Ahead Wind Forecast Error for January 2001 ...................... 3.18 Figure 3.16 Day-Ahead and Hour-Ahead Wind Forecast Error Duration for January 2001 ..... 3.18 Figure 3.17 Day-Ahead vs Hour-Ahead Wind Forecast Error Sigma ........................................ 3.19 Figure 4.1 2001 Energy Displacement by Technology ................................................................ 43 Figure 4.2 2002 Energy Displacement by Technology ................................................................ 43 Figure 4.3 2001 Energy Displacement by Fuel ............................................................................ 44 Figure 442002 Energy Displacement by Fuel ............................................................................ 44 Figure 4.5 2001 Zonal Wind Generation and Displaced Thermal Generation ............................ 45 Figure 4.6 2002 Zonal Wind Generation and Displaced Thermal Generation ............................ 45 Figure 4.7 2001 Regional Energy Displacement ......................................................................... 46 Figure 4.8 2002 Regional Energy Displacement ......................................................................... 46 GE Energy vii 3/04/05 OAGI0000191 007

Figure 4.9 2001 Emission Reductions .......................................................................................... 4.7 Figure 4.10 2002 Emission Reductions ....................................................................................... .4.7 Figure 4.11 Hours Limiting on UPNY-SENY Interface ................................................................ .4.8 Figure 4.12 Duration Curve of Hourly Flows on UPNY-SENY Interface ..................................... .4.8 Figure 4.13 Duration Curve of Hourly Flows on Total East Interface .......................................... .4.9 Figure 4.14 2001 Spot Price Duration Curve - Genesee Area .................................................. .4.1 0 Figure 4.15 2001 Economic Impact ............................................................................................ 4.11 Figure 4.16 2002 Economic Impact ............................................................................................ 4.12 Figure 4.17 2001 Zonal Load Weighted Spot Price Reduction ................................................. .4.13 Figure 4.18 2002 Zonal Load Weighted Spot Price Reduction ................................................. .4.13 Figure 5.1 Annual Wind Production - Duration Curve ................................................................. 5.2 Figure 5.2 State-wide Daily Load Profile for January 8, 2003 and August 1, 2003 ..................... 5.3 Figure 5.3 Daily Load Cycle for Superzone A-E for January 8, 2003 and August 1, 2003 ......... 5.4 Figure 5.4 Daily Load Profiles for Zone K for January 8, 2003 and August 1, 2003 ................... 5.4 Figure 5.5 January 2001 Hourly Load Change ............................................................................ 5.5 Figure 5.6 Hourly Load Variability in Superzone A-E for January 2001 ........................................ 5.6 Figure 5.7 Range of Penetration based on actual wind for January 2001 ................................... 5.7 Figure 5.8 Hours Greater than 10% Penetration for Representative (Forecast) Months ............ 5.8 Figure 5.9 Hourly Variability of Statewide Wind Alone for January 2001 ..................................... 5.8 Figure 5.10 Statewide Hourly Variability for January 2001 ......................................................... 5.1 0 Figure 5.11 Superzone A-E Hourly Variability for January 2001 ............................................... 5.10 Figure 5.12 Zone K Hourly Variability for January 2001 ............................................................ 5.11 Figure 5.13 Standard Deviation of Hourly Load Variance (by month for 11 sample months) ... 5.12 Figure 5.14 Summer Morning Load Rise - Hourly Variability .................................................... 5.13 Figure 5.15 Winter Evening Load Rise - Hourly Variability ........................................................ 5.14 Figure 5.16 Five-minute Variability Statewide ........................................................................... 5.15 Figure 5.17 Five-minute Variability for Superzone A-E .............................................................. 5.15 Figure 5.18 Five-minute Variability for Zone K ........................................................................... 5.16 Figure 5.19 State-wide Load Variation Around Five-minute Running Average for January 8, 20035.17 Figure 5.20 Superzone A-E Load Variation Around Five-minute Running Average for 2003 Figure 5.21 Zone K Load Variation Around Five-minute Running Average for January 8,20035.18 Figure 5.22 Six-second Variation by Zone, for Various Sample Days ....................................... 5.19 Figure 5.23 NYISO ACE for January 8, 2003 ............................................................................. 5.20 Figure 5.24 Histogram of ACE values for January 8, 2003 ........................................................ 5.20 Figure 5.25 Distribution of ACE Standard Deviation .................................................................. 5.21 Figure 5.26 Variability Statistics (One-second) for Samples ..................................................... 5.23 Figure 5.27 One Month of One-second Resolution Data from Operating Wind Farm .............. 5.24 Figure 5.28 Second-to-second Change for One Month ............................................................. 5.25 Figure 5.29 Rolling 10-minute Standard Deviation on One-second Change ............................ 5.26 Figure 5.30 Histogram of Standard Deviation of One-second Change for Sliding 10-minute Window ...............................................................................................................................5.26 Figure 5.31 Histogram of Statewide 6-second Variance ........................................................... 5.27 Figure 5.32 Histogram of Superzone A-E 6-second Variance .................................................... 5.28 Figure 5.33 Histogram of Zone K 6-second Variance ................................................................ 5.28 Figure 6.1. Total New York State Wind Generation (MW) over Selected 3-Hour Intervals ......... 6.4 Figure 6.2. Example Load Profile from August 2003 ................................................................... 6.5 Figure 6.3. Wind Generation Study Scenarios ............................................................................. 6.8 Figure 6.4. Wind and Load Study Scenarios ............................................................................... 6.9 Figure 6.5. Distribution of Hourly Wind and Load Variations ..................................................... 6.10 Figure 6.6. Distribution of 5-Minute Wind and Load Variations ................................................. 6.10 Figure 6.7. QSS Results for August Morning Load Rise, No Wind Generation ......................... 6.14 Figure 6.8. QSS Results for August Morning Load Rise, August Wind Generation Decrease .. 6.14 Figure 6.9. QSS Results for August Morning Load Rise, September Wind Generation Decrease.6.15 Figure 6.10. AGC Proxy Unit Output for August/September Study Scenarios .......................... 6.15 Figure 6.11 Example Unit Output for August/September Study Scenarios .............................. 6.16 GE Energy viii 3/04/05 OAGI0000191 008

Figure 6.12. AGC Proxy Unit Output for May/October Study Scenarios ................................... 6.18 Figure 6.13. Rate Limit Impact for May/October Study Scenarios ............................................. 6.18 Figure 6.14. AGC Proxy Unit Output and Rate Limit Impact for April Study Scenarios ............. 6.20 Figure 6.15. Individual Wind Farm Power Output with and without Active Power Control ........ 6.22 Figure 6.16. Total New York State Wind Generation with and without Active Power Control. .. 6.22 Figure 6.17. AGC Proxy Unit Output for October Study Scenarios with and without APe. ....... 6.23 Figure 6.18. Wind Farm Power Output Comparison .................................................................. 6.26 Figure 6.19. Wind and Load Profiles for 10-minute Stability Simulations .................................. 6.27 Figure 6.20. Impact of Wind Generation on System Performance ............................................ 6.29 Figure 6.21 Impact of LVRT on System Performance .............................................................. 6.31 Figure 6.22. Minimum Terminal Voltages for All Wind Farms in LVRT Example ...................... 6.32 Figure 6.23. Local Performance with and without Voltage Regulation ...................................... 6.34 Figure 6.24. System Performance with and without Voltage Regulation ................................... 6.35 Figure 6.25. System Performance with Different Types of WTGs ............................................. 6.37 Figure 6.26. Local Performance with Different Types of WTGs ................................................ 6.38 Figure 6.27. System Response to Frequency Swings with and without Wind Generation ........ 6.39 Figure 6.28. AGC & Frequency Response to August Load and Wind Profiles .......................... 6.41 Figure 7.1 Monthly Wind Capacity Factors .................................................................................. .7.2 Figure 7.2 Hourly Wind Capacity Factors ..................................................................................... 7.2 Figure 7.3 Average Seasonal Wind Shape, NYISO 2002 ........................................................... .7.3 Figure 7.4 2001 Average Load versus Average Wind ................................................................. .7.3 Figure 7.5 2002 Average Load and Average Wind ..................................................................... .7.4 Figure 7.6 2001 Annual Load versus Wind Scatter Plot .............................................................. .7.4 Figure 7.7 2002 Annual Load versus Wind Scatter Plot .............................................................. .7.5 Figure 7.8 July, 2002 Load versus Wind Scatter Plot .................................................................. .7.5 Figure 7.9 August, 2002 Load versus Wind Scatter Plot .............................................................. 7.6 Figure 7.10 July 2001 Wind and Load versus Time-of-Day ........................................................ .7.6 Figure 7.11 July 2002 Peak Week Wind and Load ..................................................................... .7.7 Figure 7.12 Annual Reliability Impact of Wind Sites .................................................................... .7.8 Figure 7.13 NYISO LOLP from 2001 Shapes ............................................................................... 7.9 Figure 7.14 NYISO LOLP from 2002 Shapes ............................................................................. 7.10 Figure 7.15 Phase 1 adjusted results ......................................................................................... 7.11 Figure 7.16 NYISO 2003 Daily Peak Load Profile ...................................................................... 7.12 Figure 7.17 NYISO June 2003 Loads ......................................................................................... 7.13 Figure 7.18 Peak Hour of the Day - Summer ............................................................................ .7.13 Figure 7.19 Wind Capacity Factors in Peak Load Hours ............................................................ 7.14 Figure 7.20 Annual and Peak Capacity Factors by Site for Year 2002 Shapes........................ 7.14 Figure 7.21 Average Hourly Wind Shapes for 2002 ................................................................... 7.15 Figure 7.22 NYISO Wind Capacity Factors by Zone .................................................................. 7.15 Figure 7.23 NYISO Wind Capacity Factors ................................................................................ 7.16 Figure 8.1. Hourly Wind Deltas, 2001 through 2003 .................................................................. 8.4 Figure 8.2. Hourly Wind Outputs, 2001 through 2003 ................................................................ 8.4 Table of Tables Table 1.1 Study Scenario - Wind and Load MW by Zone .......................................................... 1.2 Table 2.1 Wind Generation Included In Study Scenario ............................................................. 2.1 Table 2.2 Summary of Key Analytical Results for Study Scenario ............................................. 2.6 Table 2.3 Hourly Variability With and Without Wind Generation ................................................ 2.9 Table 2.4 Annual Operating Cost Impacts for 2001 Wind and Load Profiles (Unit commitment based on wind generation forecast) ................................................................................... 2.13 Table 2.5 Annual Operating Cost Impacts for 2001 Wind and Load Profiles (Wind generation not included for unit commitment) ............................................................................................. 2.13 Table 3.1 Forecast Error Statistics for January 2001 .................................................................. 3.9 GE Energy ix 3/04/05 OAGI0000191 009

Table 3.2 2001 Day-Ahead Forecast Error Statistics (4 months) .............................................. 3.14 Table 3.3 2002 Day-Ahead Forecast Error Statistics (4 months) ............................................... 3.14 Table 3.4 2003 Day-Ahead Forecast Error Statistics (3 Months) ............................................... 3.15 Table 3.5 Day-Ahead vs Hour-Ahead Wind Forecast Error Statistics for January 2001 ........... 3.19 Table 3.6 Statistics on Wind forecast Error for 2001 ................................................................... 3.20 Table 3.7 Statistics on Wind Forecast Error 2002 ...................................................................... 3.20 Table 3.8 Statistics on Wind Forecast Error 2003 ...................................................................... 3.21 Table 4.1 Description of Cases .................................................................................................... .4.1 Table 5.1 ACE Statistics from Eight Representative Days ........................................................ 5.22 Table 5.2 Statistics on six-second variability ............................................................................. 5.29 Table 6.1 Summary of QSS Power Flow System Conditions with No Wind Generation ............ 6.3 Table 6.2. Summary of QSS Study Scenarios ........................................................................... 6.11 Table 6.3. Summary of Stability Power Flow System Conditions with No Wind Generation .... 6.24 GE Energy 3/04/05 OAGI0000191 010

Introduction 1 Introduction 1.1 8ackg rou nd In response to emerging market conditions, and in recognition of the unique operating characteristics of wind generation, the New York Independent System Operator (NYISO) and New York State Energy Research and Development Authority (NY SERDA) commissioned a joint study to produce empirical information that will assist the NYISO in evaluating the reliability implications of increased wind generation. The work was divided into two phases.

Phase 1, Preliminary Overall Reliability Assessment, was completed in early 2004. This initial phase provided a preliminary, overall, screening assessment of the impact of large-scale wind generation on the reliability of the New York State Bulk Power System (NYSBPS). This assessment included:

Review of world experience with wind generation, focusing on regions that have integrated significant penetration of wind resources into their power grids Fatal flaw power flow analysis to determine the maximum power output at prospective wind generation sites that can be accommodated by the existing transmission system infrastructure, considering thermal ratings oftransmission lines Reliability analysis to determine the contribution of prospective wind generation towards meeting New York State requirements for Loss Of Load Expectation (LOLE)

Review of current planning and operating practices to identifY New York State Reliability Council (NYSRC), Northeast Power Coordinating Council (NPCC), North-American Electric Reliability Council (NERC), and NYISO rules, policies, and criteria that may require modification to be compatible with high penetration of wind generation Phase 2 builds on what was learned in Phase 1. A base case wind scenario with 3,300 MW of wind generation (10% of NY State peak load) was selected for analysis. Operation of the NYSBPS with 3,300 MW of wind was evaluated in numerous ways, considering impacts on the following aspects of grid performance:

Reliability and generation capacity Forecast accuracy Operation of day-ahead and hour-ahead markets Economic dispatch and load following Regulation Stability performance following major disturbances to the grid.

Detailed analysis of economic impacts and evaluation of possible generator retirements were not included in the scope ofthis study.

Results ofthese Phase 2 analyses are presented in this report.

GE Energy 1.1 3/04/05 OAGI0000191 011

Introduction 1.2 Wind Generation Scenario Starting from the original 10,026 MW of wind generation at 101 sites evaluated in Phase I, two alternate scenarios with 3,300 MW of wind generation were considered. The project team selected a scenario with 3,300 MW of wind generation in 33 locations across New York State.

Table 1.1 shows the location (by zone) of the wind farms included in the study. The lower portion of Table 1.1 lists the "Superzones" used by NY State Department of Public Service (DPS) for the RPS study. Load zones within the New York Control Area are illustrated in Figure 1.1.

The wind generation in Zone K, Long Island, is located offshore. The rest ofthe sites are land-based wind farms. The 600 MW site in Zone K was divided into 5 separate wind farms for interconnection into the Long Island transmission grid. Thus, the 33 wind sites are modeled in loadflow and stability simulations as 37 individual wind farms.

The majority of the interconnections were at the 115kV voltage level and above. Four of the Long Island interconnections were at the 69kV voltage level. No interconnections were below 69kV.

As a point of reference, the NYISO queue of proposed new generation presently has a total of 1939 MW in wind projects.

Table 1.1 Study Scenario - Wind and Load MW by Zone Total Potential 2008 Noncoincident Wind MWin Wind as % of Wind Generation Peak Load Study Scenario Peak Load Zone A 3.070 2.910 684.2 24%

Zone B 1.197 2.016 358.5 18%

Zone C 1.306 2.922 569.7 19%

Zone D 483 902 322.6 36%

Zone E 2.832 1.592 399.8 25%

Zone F 434 2.260 260.6 12%

Zone G 105 2.260 104.6 5%

Zone H 0 972 0.0 0%

Zone I 0 1.608 0.0 0%

Zone J 0 11.988 0.0 0%

Zone K 600 5.275 600.0 11%

sum 10.026 34,704 3300.0 10%

DPS Zn 1 8.887 10.342 2334.8 23%

DPS Zn 2 538 7.099 365.2 5%

DPS Zn 3 600 17.263 600.0 3%

sum 10.026 34,704 3300.0 10%

Notes: DPS Zn 1 = Zones A + B + C + D + E DPS Zn 2 = Zones F + G + H OPSZn 3 = Zones I + K GE Energy 1.2 3/04/05 OAGI0000191 012

Introduction Figure 1 1 New York Control Area Load Zones The majority of the wind generation in the study scenario is located in upstate NY, Zones A tlrrough E. In those zones, penetration of wind generation is 23% of peak zonal load. The 600 MW of offshore wind generation in Zone K represents 11 % of peak load in that zone.

The model of the New York State Bulk Power System (NYSBPS) used in this study was derived from NYISO's 2008 transmission and generation modeled Zonal load profiles were derived from measured data from years 2001-2003, scaled upward to be consistent with pr~ected load levels in 2008. Selection of year 2008 for the study scenario is conservative, since 3,300 J\1\V of operational wind generation is more than would be expected by that time.

Wind turbine-generators were assmned to have characteristics consistent with present state-of-the-art technology, and included continuously controllable reactive power capability (0.95 power factor at point of intercOlUlection), voltage regulation, and low-voltage ride-tlrrough (L VR1).

1_3 Timescales for Power System Planning and Operations The power system is a dynamic system, subject to continuously changing conditions, some of wbich can be anticipated and some of wbich calUlOt The primary fimction of the power system is GEEnergy 13 OAGI0000191 013

Introduction to serve a continuously varying customer load. From a control perspective, the load is the primary independent variable - the driver to which all the controllable elements in the power system must be positioned and respond. There are annual, seasonal, daily, minute-to-minute and second-to-second changes in the amount (and character) of load served by the system. The reliability of the system then becomes dependent on the ability of the system to accommodate expected and unexpected changes and disturbances while maintaining quality and continuity of senrice to the customers.

As illustrated in Figure 1.2, there are several time frames of variability, and each time frame has corresponding planning requirements, operating practices, information requirements, economic implications and technical challenges. Much of the analysis presented in this report is aimed at quantitatively evaluating the impact of significant wind variability in each of the time frames on the reliability and performance ofthe NYSBPS.

Figure 1.2 shows four timeframes covering progressively shorter periods of time. In the longest timeframe, planners must look several years into the future to determine the infrastructure requirements of the system. This timeframe includes the time required to permit and build new physical infrastructure. In the next faster timeframe, day-to-day planning and operations must prepare the system for the upcoming diurnal load cycles. In this time frame, decisions on unit commitment and dispatch of resources must be made. Operating practices must ensure reliable operation with the available resources. During the actual day of operation, the generation must change on an hour-to-hour and minute-to-minute basis. This is the fastest time frame in which economics and human decision-making playa substantial role. Unit commitment and scheduling decisions made the day ahead are implemented and refined to meet the changing load. In NY State, the economic dispatch process issues load following commands to individual generators at 5-minute intervals. In the fastest time frame (at the bottom of the figure), cycle-to-cyle and second-to-second variations in the system are handled primarily by automated controls. The system automatic controls are hierarchical, with all individual generating facilities exhibiting specific behaviors in response to changes in the system that are locally observable (i.e. are detected at the generating plant or substation). In addition, a subset of generators provide regulation by following commands from the centralized automatic generation control (AGC), to meet overall system control objectives including scheduled interchange and system frequency.

GE Energy 1.4 3/04/05 OAGI0000191 014

Introduction Planning and Technology Operation Process Issues Capacity Valuation Resource and (UCAP,ICAP)

Capacity Planning and Long-Term Load (Reliability) Growth Forecasting

......-- 10 Minutes -----.

Figure 1.2 Time Scales for System Planning and Operation Processes GE Energy 1.5 3/04/05 OAGI0000191 015

Introduction Seasonal \/\And Patterns Regional Wind Fluctuations Wind Farm Power Fluctuations Individual WTG Power Fluctuations Resource Planning Lnit Conrnitment and Scheduling Load Follow ing c:::::::::::J Regulation

~c ~

Q) ~0 iii

I "0 '"

0 ~  :;;

Q)

?

1:>  % ~ ~ ~ ~ ~ ~

0 00 0 0 s, '&

0  %

Time Scale (seconds)

Figure 1.3 Wind Variability and Impact on System Operation Processes Wind, as a variable and largely undispatchable generating resource, will impact all of these planning and operation processes. Wind variability has its own characteristics and time frames.

As with system load, there are seasonal, diurnal, hour-to-hour , minute-to-minute and second-to-second variations. In the case of wind generation, as the time frame decreases the correlation between wind generating resources dropsi This is shown in the upper portion of Figure 1.3, where the spatial aspect of wind variation is correlated to the time-scale of temporal variations.

Individual wind turbine-generators (WTGs) commonly experience power output variations in the one-second to several-minute timeframe. When many WTGs are grouped together in a wind fann, the short-term variations of individual WTGs are attenuated as a percentage of the aggregate, and the dominant power output variations for the entire wind fann occur in the minute-to-hour time frame. Similarly, the minute-to-minute power output of individual wind fanns are attenuated in systems with multiple wind fanns, leaving regional wind fluctuations in the hour-to-day time frame as the dominant system-wide effect. Seasonal wind patterns, of course, fall into the several-month timeframe.

The lower portion of Figure 1.3 shows how these wind variations relate to the four groups of planning and operation processes identified in Figure 1.2.

GE Energy 1.6 3/04/05 OAGI0000191 016

Introduction 1.4 Technical Approach The technical approach for this project addresses the range of processes involved in the planning and operation ofthe NYSBPS, over the range oftimescales from seconds to years. The analysis focuses on the overall performance of the NYSBPS with a high penetration of wind generation, and does not address all localized effects related to each individual wind farm.

The bulk ofthe technical analysis was grouped into four major areas as described below.

1.4.1 Forecast Accuracy The accuracy of the wind forecast affects unit commitment and operating reserve policies.

Accuracy of wind generation forecasting was evaluated, and related to the historical accuracy of load forecasts used in the day-ahead market.

1.4.2 Wind and Load Variability The NYSBPS already deals with significant variability in system load. Wind generation, as a variable power source, adds to the total variability that the NYSBPS must accommodate. The analysis of variability addressed the both major contributors to variability over several time frames:

Variability: Variability due to load alone Variability due to wind alone Combined variability due to load and wind, synchronized over the same calendar periods.

Time Frames: Hourly 5-minutes (load-following; economic dispatch)

Seconds (regulation, AGC)

This analysis used consistent sets of historical wind data and historical load data, for the same time periods.

1.4.3 Operational Impact Operational impacts cover a range oftime scales, from seconds to multiple hours. Operation of the NYSBPS was simulated with and without wind generation (per the study scenario) as follows:

Simulation of statewide operations for an entire year using MAPS, focusing on dispatch and unit commitment issues as a function of wind forecast accuracy.

GE Energy 1.7 3/04/05 OAGI0000191 017

Introduction Quasi-steady-state simulation of selected 3-hour periods for wind and load variability, focusing on issues that affect load following.

Stability simulation of selected 10-minute periods, focusing on regulation and other short-term control and protection issues (voltage regulation, low-voltage ride-through, AGC, etc.)

1.4.4 Effective Capacity Using the Multi-Area Reliability Simulation (MARS) program, the effective capacity of wind generation, was quantified by comparing it with a typical fossil-fired power plant. This analysis includes consideration ofthe seasonal and diurnal variability in wind generation output relative to periods of peak system load, when generating resources have the greatest impact of overall system reliability as measured by loss-of-Ioad probability (LOLP).

In addition to quantifYing the likely range of unforced capacity (UCAP) for wind generation in NY State, approximate techniques for calculating the UCAP of individual wind farms were developed.

1.5 Data Technical information and data for this study were obtained from the following sources:

NYISO provided power flow and stability datasets, historical operating data for years 1999-2003, and contingency lists for the NYSBPS and NYSRC reliability datasets.

A WS TrueWind provided data on potential wind generation sites in NY State, wind MW generation at those sites based on historical weather data, and technical information related to wind generation and wind forecasting.

NYSDPS provided generation fuel cost and heat rate data from the preliminary RPS analyses.

Appendix A contains detailed descriptions of data provided by NYISO and A WS TrueWind.

GE Energy 1.8 3/04/05 OAGI0000191 018

Executive Summary 2 Executive Summary This study evaluated the impact of wind generation on the New York State Bulk Power System (NYSBPS) over a broad range of subject areas, including planning, operation, economics, and reliability. Key results and conclusions are summarized here. Details of the analysis, and the reasoning behind the conclusions, are further explained in Chapters 3-8.

2.1 Study Scenario for Wind Generation The technical analysis for this study focused on a wind generation scenario that included a total of 3,300 MW of wind generation in 33 locations throughout New York State (see Table 2.1). Most of the wind sites are located upstate, but there is one large offshore facility near Long Island (Zone K). The total amount of wind generation (nameplate rating) in this scenario corresponds to approximately 10% of New York State's 2008 projected peak load. The majority of the wind farm interconnections were at the 115kV voltage level and above. Some interconnections for the Long Island site were at the 69kV voltage level. No interconnections were below 69kV.

Table 2.1 Wind Generation Included In Study Scenario Wind Generation Wind Generation as Total for NY 3300.0 10%

PowertIow and operational models for the study scenario were derived from NYISO's 2008 system model. Hourly and shorter-term load profiles were based on actual historical data from years 2001-2003, but were scaled to match the projected load for 2008. Profiles of wind generation at the 33 locations were derived from historical weather records for years 2001-2003, so wind generation in the study scenario was treated as though the wind generators were actually in operation during those years.

Observations and conclusions presented in this report are based on analysis ofthis study scenario.

GE Energy 2.1 3/04/05 OAGI0000191 019

Executive Summary 2.2 Impact on System Planning A wide variety of standards, policies and criteria were reviewed to assess their impact on wind generation, and to determine if changes were needed to accommodate wind generation. In general, it was found that the existing rules and criteria could be applied to wind generation. A few specific items are discussed below.

2.2.1 NYISO System Reliability Impact Study (SRIS)

NYISO's SRIS is intended to confirm that a new facility complies with applicable reliability standards, to assess the impact of the new facility on the reliability of the pre-existing power system, to evaluate alternatives for eliminating adverse impacts (if any), and assess the impact of the new facility on transmission transfer limits. The SRIS policy is directly applicable to wind generation in its present form.

2.2.2 NYSRC Reliability Rules for Planning and Operation NYSRC reliability rules are outlined in the document NYSRC Reliability Rules for Planning and Operating the New York State Power System, which addresses both resource adequacy and system security. A few minor changes related to planning studies are recommended:

The rules for steady-state analysis require evaluation of single-element (N-l) and extreme contingencies. Normally, loss of one generator in a multi-generator power plant would be a single-element contingency. Wind farms are comprised of many wind turbine-generators connected to a common interconnection bus. It is recommended that the loss ofthe entire wind farm be considered a single-element contingency for the purpose of NYSRC reliability criteria.

However, simultaneous loss of multiple wind farms due to loss of wind in not a credible event.

No changes to NYSRC rules for extreme contingencies, or multiple-element outages, are recommended.

NYSRC rules for stability analysis require evaluation of both design criteria and extreme faults.

No changes to these rules or their interpretation are required for wind generation.

2.2.3 Generation Interconnection Requirements In the Phase 1 report, it was recommended that New York State adopt some of the interconnection requirements that have emerged from the experiences of other systems.

Specifically, New York State should require all new wind farms to have the following features:

GE Energy 2.2 3/04/05 OAGI0000191 020

Executive Summary Voltage regulation at the Point-Of-Interconnection, with a guaranteed power factor range Low-voltage ride-through (LVRT)

A specified level of monitoring, metering, and event recording Power curtailment capability (enables system operator to impose a limit on wind farm power output)

The above features are implemented in wind farms around the world, and are proven technology.

During Phase 2, technical analysis was performed to evaluate some ofthese features with respect to performance of the NYSBPS. Specifically, the impact of voltage regulation and low voltage ride through (LVR T) on system performance was demonstrated. The results showed that voltage regulation with a +/-O.95 power factor range improves system response to disturbances, ensuring a faster voltage recovery and reduced post-fault voltage dips. In addition, L VR T ensures that wind farms remain connected to the NYSBPS under low voltage conditions due to faults or other system disturbances, and mitigates concerns about loss of multiple wind farms due to system events. Good performance was demonstrated with LVRT parameters that are less aggressive than the emerging industry consensus. It is recommended that New York adopt the emerging L VR T specification.

No operating conditions were found to justifY the need for wind power curtailment at a statewide level (i.e., backing down all wind generators at the same time). However, for system reliability reasons, NYISO should require a power curtailment feature on new wind farms as a mechanism to posture the power system to handle temporary local transmission limitations (e.g., line out of service) or in anticipation of severe weather (e.g., intentionally curtail wind generation in advance of a severe storm affecting a large portion of the state). Such curtailment could be done by NYISO sending maximum power orders to wind farm operators (similar to the existing process for re-dispatching a thermal generator via the plant operator) or via SCADA for the case of unmanned generation facilities. This type of curtailment is envisioned as a farm-level function, not necessarily a turbine-level function. For example, if NYISO needed to limit power output of a specific wind farm due to a temporary transmission line outage, the wind farm operator could temporarily curtail generation by limiting output or shutting down a portion of the wind turbines in the wind farm. This is the same as would be done at any other dispatchable generating facility in New York State under the same circumstances.

Interconnection requirements are different for each transmission owner in New York State. In general, standards for interconnection of wind turbines are the same as for other generation.

GE Energy 2.3 3/04/05 OAGI0000191 021

Executive Summary Thus, frequency and voltage ranges, power factor ranges and other protection requirements remain largely unchanged. However, some features, such as governor control and power system stabilizer (PSS), are either technically impractical now or inappropriate for wind generators.

Presently, New York State has varied requirements for generator power factor. NERC Planning Standards require the following, "At continuous rated power output, new synchronous generators should have an overexcited power factor capability, measured at the generator terminals, of 0.9 or less and an underexcited power factor capability of 0.95 or less."ii Niagara Mohawk's requirements are consistent with those of NERC, but LIP A requires generators to have a power factor capability of 0.90 leading to 0.90 lagging at the point of delivery.

FERC NOPR RM05-4-000 (dated January 24,2005) proposes that "a wind plant shall maintain a power factor within the range of 0.95 leading to 0.95 lagging, measured at the high voltage side of the wind plant substation transformer(s)."iii The requirement for measurement at the high side voltage recognizes the distributed nature of wind plants. The FERC NOPR power factor range measured at the high side bus is consistent with NERC requirements at generator terminals.

It is recommended that wind generation facilities meet power factor requirements consistent with other generation facilities in New York State and with existing local interconnection criteria, but translated to the high side voltage of the wind plant substation transformer. NYISO and New York State transmission owners may wish to re-evaluate the power factor requirements after FERC enacts a rule.

2.2.4 Future Interconnection Options In the Phase I report, the following features were identified as emerging in response to system needs, and should be considered by New York State in the future as they become available:

Ability to set power ramp rates Governor functions Reserve functions Zero-power voltage regulation During Phase 2, technical analysis was performed to evaluate one ofthese features with respect to performance ofthe NYSBPS. Specifically, the ability to set power ramp rates for wind farms was demonstrated. The example ramp rate limit function resulted in a decrease in statewide GE Energy 2.4 3/04/05 OAGI0000191 022

Executive Summary regulation requirements at the expense of wind energy production. Therefore, such a function should only be used in specific applications to ensure system reliability.

2.3 Impact on System Operations Table 2.2 provides a condensed summary of many key study results, arranged according to time scale. The following sections discuss each item in detail.

GE Energy 2.5 3/04/05 OAGI0000191 023

Executive Summary Table 2.2 Summary of Key Analytical Results for Study Scenario Time Without Wind With Wind Technical Issue Comments Scale Generation Generation Years UCAP of Wind UCAP land-based ~ 10%

• UCAP is site-specific Generation UCAP offshore ~ 36% (one site in L.I.)

• Simple calculation method proposed

• Incremental increase can be accommodated by existing processes and resources in NY State Days Day-Ahead Forecasting and Unit Commitment Forecasting error:

c; ~ 700-800 MW Forecasting error:

c; ~ 850-950 MW

• Even without forecasts, wind energy displaces conventional generation, reduces system operating costs, and reduces emissions.

• Accurate wind forecasts can improve results by another 30%

Hours Hourly Variability c;=858MW c;=910MW

• Incremental increase can be accommodated by existing processes and resources in NY State Largest Hourly Load Rise 2575 MW 2756 MW

• Incremental increase can be accommodated by existing processes and resources in NY State Minutes Load Following (5-min Variability) c; = 54.4 MW c; = 56.2 MW

• Incremental increase can be accommodated by existing processes and resources in NY State 36 MW increase required to

• NYISO presently exceeds NERC criteria Regulation 225 to 275 MW maintain same performance

• May still meet minimum NERC criteria with existing regulating capability Seconds Spinning Reserve 1200 MW 1200 MW

• No change to spinning reserve requirement Stability 8% post-fault voltage dip (typical) 5% post-fault voltage dip (typical)

• State-of-the-art wind generators do not participate in power swings, and improve post-fault response of the interconnected power grid.

Note: c; ~ standard deviation GE Energy 2.6 3/04/05 OAGI0000191 024

Executive Summary 2.3.1 Forecasting and Market Operations NYISO's day-ahead market presently uses day-ahead load forecasts as part of the generation commitment and scheduling process. The error between forecast load and actual load introduces a level of uncertainty that must be accommodated by NYISO's operating practices. Wind generation introduces another element of uncertainty. Analysis of wind forecast performance for the study scenario shows that errors in day-ahead wind generation forecasts have standard deviations of approximately 400 MW, or 12% ofthe aggregate rating of all the wind generators (3,300 MW).

Figure 2.1 shows the standard deviations ofload forecast error, wind forecast error, and combined "Load minus Wind" forecast error for 11 selected months of years 2001-2003. The figure shows that total forecasting error (Load-Wind) is somewhat higher than the forecasting error due to load alone. For example, in the peak load months (points on the right-hand side), the total forecast error increases from 700-800 MW without wind generation (Load alone) to 850-950 MW with 3,300 MW of wind generation (Load-Wind). NYISO operational processes to deal with uncertainty in load forecasting already exist. The same processes can be used to handle the increase in forecast uncertainty due to wind generation.

1000 800

~ ",,,,

6- 600

" ~~

'E" 400 **..'"

Ol U5 200 15000 20000 25000 30000 35000 Peak Load for Corresponding Month (MW)

I. Load _ Wind '" Load - windl OJanuary 2001 Figure 2.1 Standard Deviation of Day-Ahead Forecast Errors Accuracy of wind forecasts improves as the lead-time decreases. For the study scenario, errors in hour-ahead wind generation forecasts are expected to have standard deviations of approximately 145 MW, or 4.2%.

GE Energy 2.7 3/04/05 OAGI0000191 025

Executive Summary Wind forecast uncertainties are of sufficient magnitude at the levels of penetration examined in this study to warrant the use of state-of-the-art forecasting. Data collection from existing and new wind farms should proceed immediately, in order to provide input to, and increase the fidelity of, wind forecasts for when the system achieves higher levels of penetration. New York should also consider meteorological data collection and analysis from proposed and promising wind generation locations in order to aid and accelerate the integration of high fidelity wind forecasting into NYISO's operating practices.

The existing day-ahead and hour-ahead energy markets in New York have sufficient flexibility to accommodate wind generation without any significant changes. It may also be advantageous for the forecasting to be performed from a central location to ensure a consistency of methodologies and so that changing weather patterns can be noted quickly. With these factors in place wind generation can be held accountable to similar standards as conventional generation in terms of meeting their day-ahead forecast, with one exception; imbalance penalties should not be imposed on wind generation Wind projects would need to settle discrepancies between their forecast and actual outputs in the energy balancing market. However, because wind is largely non-dispatchable, any additional penalties for imbalance should be eliminated. The FERC Order 888 allows imbalance penalties to be applied to generators that operate outside oftheir schedule. As applied in New York, any "overgeneration" can be accepted without payment and any "undergeneration" is priced at the greater of 150% of the spot price or $IOO/MWh. Strict application of these policies in the MAPS analysis performed would result in the loss of roughly 90% ofthe wind generation revenue, which would be disastrous to their future development. The intent of the penalties is to prevent generators from "gaming" the market but their application to intermittent resources such as wind and solar would result in negative and unintended consequences. If a wind generator forecasted 100 MW for a particular hour but can only produce 80 MW due to a lack of wind then no amount of penalties can get them to produce the remaining 20 MW. Their only option would be to bid less, or zero, in the day ahead market and possibly even bid low in the hour ahead market. However, the MAPS analysis showed that as much as 25% of the value of the wind energy to the system could be lost if it is not properly accounted in the day ahead commitment process. Any imbalance penalties for under-generation would tend to encourage underbidding the day ahead forecast, to the detriment ofthe entire system.

In order to take advantage ofthe spatial diversity of multiple plants, it may also be appropriate to aggregate wind generation on a zonal or regional basis rather than treating them as individual plants.

GE Energy 2.8 3/04/05 OAGI0000191 026

Executive Summary Wind forecasting may be performed in either a centralized or decentralized manner. With either approach, forecasts would be generated for each individual wind farm. However, centralized wind forecasting has several advantages that the NYISO may wish to consider:

Application of a consistent methodology, which should achieve more consistent results across projects More effective identification of approaching weather systems affecting all wind plants, to warn the ISO of impending large shifts in wind generation Use of data from each plant to improve the forecasts at other plants. For example, a change in output of one plant might signal a similar change in other plants downstream of the first. Individual forecasters would not have access to the data from other projects to make this possible.

Care should be taken in the structuring of any financial incentives that may be offered to encourage the development of wind generation. The market for wind generation (including incentives) should be structured to:

Reward the accuracy of wind generation forecasts, and encourage wind generators to reduce production during periods oflight load and excessive generation.

The second item above is particularly critical to overall system reliability. If excessive wind generation causes the NYISO to shut down critical base-load generators with long shutdown/restart cycle times, the system could be placed in a position of reduced reliability. The market for wind power should be structured so that wind generators have clear financial incentives to reduce output when energy spot prices are very low (or negative).

2.3.2 Hourly Variability Load and wind production vary from day-to-day and hour-to-hour, exhibiting characteristic diurnal patterns. The wind variability increases the inherent variability that already exists due to loads. Table 2.3 shows the changes in hourly variability due to the addition of wind generation, expressed as standard deviations (a).

Table 2.3 Hourly Variability With and Without Wind Generation Without Wind With Wind Increase Statewide 858 MW 910 MW 6%

Superzone A-E 268 MW 313 MW 17%

Zone K 149 MW 171 MW 15%

GE Energy 2.9 3/04/05 OAGI0000191 027

Executive Summary System operators give special attention to periods of peak demand and rapid rise in load. The summer moming load rise, especially during periods of sustained hot weather, presents one of the more severe tests to the system. Figure 2.2 shows the hour-to-hour variability for the load rise period for momings during June through September. The natural diurnal tendency for wind generation to fall off during this period causes higher rates of rise. In this sample, 31% of the hours have rise rates greater than 2,000 MW/hr without wind, with the worst single hour rising 2,575 MW. With the addition of wind generators, this increases to 34% of hours with rise rates greater than 2,000 MW/hr, and the worst single hourly rise is 2,756 MW. Existing NYISO operating practices are expected to accommodate this increase.

200 150 G'

c

~ 100

! Load&Wind 1'""""--1 50

-3000 -2000 -1000 1000 2000 3000 (MW)

IDLoad _Load-Wind I Figure 2.2 Summer Morning Load Rise - Hourly Variability 2.3.3 Load-Following The impact of 3,300 MW of wind generation imposed on existing load-following performance was evaluated by both statistical analysis and time-response simulations.

NYISO sends economic dispatch commands to generators at 5-minute intervals. Statistical results are summarized as a histogram in Figure 2.3, showing the distribution of 5-minute changes in load with and without wind. These results indicate that wind generation would introduce only a small increase in the load-following duty for generators on economic dispatch. The standard GE Energy 2.10 3/04/05 OAGI0000191 028

Executive Summary deviation of the statewide samples increases by 1.8 MW (3%), from 54.4 MW without wind generation to 56.2 MW with wind generation.

600 500 400 200 100 Load & Wind

,..A-,

-250 -150 -50 50 150 250 (MW)

I OLoad _Load-Wind I Figure 2.3 Five-Minute Statewide Variability Quasi-steady-state (QSS) time simulations were performed to evaluate load-following performance during selected periods when both load and wind experienced large changes (e.g.,

rising load while wind generation declines, and vice-versa). The simulations were for load and wind profiles near the upper extremes of both Figure 2.2 and Figure 2.3, as indicated by the annotations on the figures. The results show that the existing economically dispatched generators would accommodate the increase in load-following duty.

2.3.4 Regulation NYISO's automatic generation control (AGe) system maintains intertie flows and system frequency by issuing power commands to the regulating units at 6-second intervals. Existing operating practices require 225 MW to 275 MW of regulating units on-line, depending on the season. The impact of 3,300 MW of wind generation imposed on the existing regulating scheme was evaluated by both statistical analysis and stability simulations.

The statistical analysis of the study scenario shows that the standard deviation (a) of 6-second variability due to load alone is 71 MW. As a check of existing regulation practice, this result suggests that 3a, or 213 MW, of regulation would cover 99.7% ofthe time. With the addition of 3,300 MW of wind generation, the standard deviation increases from 71 MW to 83 MW. This implies that a 36 MW (3a) increase in regulating capability will maintain the existing level of GE Energy 2.11 3/04/05 OAGI0000191 029

Executive Summary regulation performance with the addition of3,300 MW of wind generation. Stability simulations covering selected 10-minute periods produced similar results.

This conclusion is further reinforced by the results of the 5-minute variability analysis.

Variations in periods less than five minutes are addressed by regulation, while longer-term variations are addressed by economic dispatch (load-following). The analysis shows the standard deviation of combined load and wind variability for 5-minute periods is 56.2 MW (up from 54.4 MW due to load alone).

NYISO regulation performance (CPS1 and CPS2) presently exceeds NERC criteria. It is possible that the NYISO grid could accommodate 3,300 MW of wind generation with no increase in NYISO's regulation capability, and still meet minimum NERC criteria.

2.3.5 Spinning Reserves Spinning reserves are required to cover the largest single contingency that results in a loss of generation. The present requirement is 1,200 MW. Analysis of historical statewide wind data indicates that loss of all wind generation due to abrupt loss of wind in not a credible contingency, and hence, the spinning reserve requirement would not be affected. Short-term changes in wind are stochastic (as are short-term changes in load). A review of the wind plant data revealed no sudden change in wind output in three years that would be sufficiently rapid to qualifY as a loss-of-generation contingency.

2.3.6 System Operating Costs GE's Multi-Area Production Simulation (MAPS) program was used to simulate the hourly operation of the NYSBPS for several years, with and without wind generation per the study scenario. Several different techniques for integrating wind generation into NYISO's unit commitment and day-ahead market were considered. The most likely approach involves using day-ahead wind generation forecasts for the unit commitment process, and scheduling wind generation before hydro. The process essentially shifts hydro generation within a several day period to make the best use of wind resources when they are available. Operating cost impacts for this approach are summarized in Table 2.4, based on the 2001 historical hourly load and wind profiles. (Note: System-wide impacts include NYISO, ISO-NE, and PJM.) The MAPS simulation results also indicate a $1.80/MWh average reduction in spot price in New York State.

GE Energy 2.12 3/04/05 OAGI0000191 030

Executive Summary Table 2.4 Annual Operating Cost Impacts for 2001 Wind and Load Profiles (Unit commitment based on wind generation forecast)

System-Wide NYISO Total variable cost reduction (includes fuel cost, variable

$ 430M $ 350M O&M, start-up costs, and emission payments)

Total variable cost reduction per MW-hour of wind

$48/ MWh $39/ MWh generation Wind revenue $ 315M $ 315M Non-wind generator revenue reductions $ 795M $ 515M Load payment reductions (calculated as product of hourly

$ 515M $ 305M load and the corresponding locational spot price)

The operating costs depend on how the wind resources are treated in the day-ahead unit commitment process. If wind generation forecasts are not used for unit commitment, then too many units are committed and efficiency of operation suffers. The operating costs for this situation are summarized in Table 2.5. In this case, unit commitment is performed as if no wind generation is expected, and wind energy just "shows up" in the real time market. The results indicate that energy consumers benefit from greater load payment reductions, but non-wind generators suffer due to inefficient operation of committed units. Comparing the system-wide variable cost reductions for these two cases, there is a $430M-$335M ~ $95M annual benefit to be gained from using wind energy forecasts for day-ahead unit commitment.

Table 2.5 Annual Operating Cost Impacts for 2001 Wind and Load Profiles (Wind generation not included for unit commitment)

System-Wide NYISO Total variable cost reduction (includes fuel cost, variable

$ 335M $ 225M O&M, start-up costs, and emission payments)

Total variable cost reduction per MW-hour of wind

$38/ MWh $25/ MWh generation Wind revenue $ 305M $ 305M Non-wind generator revenue reductions $ 960M $ 600M Load payment reductions (calculated as product of hourly

$ 720M $ 455M load and the corresponding locational spot price)

Any economic incentives that may be offered to wind generators should be designed to encourage use of state-of-the-art forecasting and active participation in the day-ahead power market.

2.3.7 Energy Displacement and Emission Reductions Energy produced by wind generators will displace energy that would have been provided by other generators. Considering wind and load profiles for years 2001 and 2002, 65% of the energy GE Energy 2.13 3/04/05 OAGI0000191 031

Executive Summary displaced by wind generation would corne from natural gas, 15% from coal, 10% from oil, and 10% from imports. As with the economic impacts discussed above, the unit commitment process affects the relative proportions of energy displaced, but the general trend is the same regardless of how wind generation is treated in the unit commitment process.

By displacing energy from fossil-fired generators, wind generation causes reductions in emissions from those generators. Based on wind and load profiles for years 2001 and 2002, annual NOx emissions would be reduced by 6,400 tons and SOx emissions would be reduced by 12,000 tons.

2.3.8 Transmission Congestion Because most of the wind generation is located in upstate New York, transmission flows increase from upstate to downstate with the addition of wind generation. Figure 2.4 shows a time-duration curve of the UPNY-SENY (upstate New York to Southeast New York) interface flow for year 2008, with and without wind generation per the study scenario. Without wind generation, interface flow is at its limit for approximately 1100 hours0.0127 days <br />0.306 hours <br />0.00182 weeks <br />4.1855e-4 months <br />. Wind generation increases the number of hours at limit to 1300. Most ofthe time, the interface is not limited and increased flows due to wind generation are accommodated.

UPNY-SENY Interface Flows 6000 5000

~

~ 4000 2:

~

3000 ~

~ 2000 I-nowind I ~

LL I-actual wind I 1000 1001 2001 3001 4001 5001 6001 7001 8001 Hours Figure 2.4 Duration Curve of Hourly Flows on UPNY-SENY Interface GE Energy 2.14 3/04/05 OAGI0000191 032

Executive Summary 2.4 Impact on System Reliability 2.4.1 Effective Capacity of Wind Generators The effective capacity of wind generation in the study scenario was quantified using rigorous loss-of-Ioad probability (LOLP) calculations with the Multi-Area Reliability Simulation (MARS) program. The results show that the effective capacities, UCAP, ofthe inland wind sites in New York are about 10% oftheir rated capacities, even though their energy capacity factors are on the order of 30%. This is due to both the seasonal and daily patterns of the wind generation being largely "out-of-phase" with NYISO load patterns. The offshore wind generation site near Long Island exhibits both annual and peak period effective capacities on the order of 40% - nearly equal to their energy capacity factors. The higher effective capacity is due to the daily wind patterns peaking several hours earlier in the day than the rest of the inland wind sites and therefore being much more in line with the load demand.

An approximate methodology for calculating effective capacity, UCAP, of wind generation was demonstrated. A wind generator's effective capacity can be estimated from its energy capacity factor during a four-hour peak load period (1:00 pm to 5:00 pm) in the summer months. This method produces results in close agreement with the full LOLP analytical methodology.

2.4.2 System Stability The transient stability behavior of wind generation, particularly vector controlled WTGs, is significantly different from that of conventional synchronous generation. The net result of this behavior difference is that wind farms generally exhibit better stability behavior than equivalent (same size and location) conventional synchronous generation. In fact, simulation results demonstrate that overall stability performance of the NYSBPS is better with 3,300 MW of wind generation than it is without wind generation. Both post-fault voltage dips and oscillations in interface flows are improved with the addition of vector controlled wind turbine-generators.

It is recommended that New York State require all new wind farms to include voltage regulation and low voltage ride through (LVRT) features. Voltage regulation improves system response to disturbances, ensuring a faster voltage recovery and reduced post-fault voltage dips. LVRT ensures that wind farms remain connected to the NYSBPS under low voltage conditions due to faults or other system disturbances. Good performance was demonstrated with LVRT parameters that are less aggressive than the emerging industry consensus. However, it is recommended that NYS adopt the emerging LVRT specification (15% voltage at the point of interconnection for 625 GE Energy 2.15 3/04/05 OAGI0000191 033

Executive Summary milliseconds), consistent with the recent FERC NOPR on wind generation interconnection requirements.

2.5 Conclusions Based on the results of this study, it is expected that the NYSBPS can reliably accommodate at least 10% penetration, 3,300 MW, of wind generation with only minor adjustments to its existing planning, operation, and reliability practices. This conclusion is subject to several assumptions incorporated in the development ofthe study scenario:

Individual wind farms installed in NY State would require approval per the existing NYISO procedures, including SRIS.

Ratings of wind farms would need to be within the capacity oflocal transmission facilities, or subject to local constraints.

Wind farms would include state-of-the-art technology, with reactive power, voltage regulation, and LVRT capabilities consistent with the recommendations in this report.

GE Energy 2.16 3/04/05 OAGI0000191 034