ML12090A820

From kanterella
Jump to navigation Jump to search
Entergy Pre-Filed Evidentiary Hearing Exhibit ENT000503, Growing Wind - Final Report of the NYISO 2010 Wind Generation Study
ML12090A820
Person / Time
Site: Indian Point  Entergy icon.png
Issue date: 09/30/2010
From:
New York Independent System Operator
To:
Atomic Safety and Licensing Board Panel
SECY RAS
Shared Package
ML12090A819 List:
References
RAS 22161, 50-247-LR, 50-286-LR, ASLBP 07-858-03-LR-BD01
Download: ML12090A820 (120)
v • d • e


Text

ENT000503 Submitted: March 30, 2012 Growing Wind Final Report of the NYISO 2010 Wind Generation Study September 2010

Caution and Disclaimer The contents of these materials are for information purposes and are provided as is without representation or warranty of any kind, including without limitation, accuracy, completeness or fitness for any particular purposes. The New York Independent System Operator assumes no responsibility to the reader or any other party for the consequences of any errors or omissions. The NYISO may revise these materials at any time in its sole discretion without notice to the reader.

NYISO Wind Generation Study September 2010

Contents Executive Summary ........................................................................................................................i

1. Introduction .............................................................................................................................i
2. Technical Approach...............................................................................................................ii
3. Study Findings ......................................................................................................................iv 3.1 Reliability Finding: .......................................................................................................... iv 3.2 Operation and Dispatch Simulation Findings: .............................................................. v 3.3 Resource Adequacy Findings:....................................................................................... vi 3.4 Production Cost Simulation Findings:.......................................................................... vi 3.5 Environmental Findings: .............................................................................................. viii 3.6 Transmission Planning Findings:................................................................................ viii
4.

Conclusions:

..........................................................................................................................x NYISO Wind Generation Study......................................................................................................1

1. Purpose...................................................................................................................................1
2. Background ............................................................................................................................2
3. Wind Plant Integration - Issues ...........................................................................................3
4. Study Tasks and Process .....................................................................................................4
5. Wind Study Results ...............................................................................................................6 5.1. Results for Task 1 - Study Assumptions: ...................................................................... 6 5.2. Results for Task 2 - Wind Plant Performance Monitoring:........................................... 6 5.3. Results for Task 3 - European, US and Canada Experience with Wind Plants:......... 8 5.4. Results for Task 4 - Assessing the Impact of Wind Plants on System Operations: 10 5.4.1. Introduction.......................................................................................................... 10 5.4.2. The Critical Importance of Net Load.................................................................... 14 5.4.3. Net Load Variability Characterization .................................................................. 15 5.4.4. Measuring Variability ........................................................................................... 15 5.4.5. Impact of Wind on System Net-Load Variability .................................................. 19 5.4.6. Impact of Increasing Wind Penetration on System Regulation ........................... 24 5.4.7. Impact of Increasing Wind Penetration on Load Following and Ramping .......... 31 5.4.8. Impact of Increasing Wind Penetration on Operating Reserves ......................... 43 5.4.9. Impact of Increasing Wind Penetration on Resource Adequacy Requirements . 43 5.4.10. Summary of findings for Task 4........................................................................... 45 NYISO Wind Generation Study September 2010

5.5. Results for Task 5 - Impacts of Wind Generation on Transmission Infrastructure: 46 5.5.1.

Introduction:

......................................................................................................... 46 5.5.2. Initial Assessment of Transmission Constraints.................................................. 46 5.5.3. Evaluation of Multiple Zone Sources................................................................... 49 5.5.4. Transient Stability Analysis.................................................................................. 50 5.6. Results for Task 6 - Production Simulation Analysis:................................................ 51 5.6.1. Introduction.......................................................................................................... 51 5.6.2. Locational Based Marginal Prices (LBMP).......................................................... 52 5.6.3. Fuel Types Displaced by Wind Generation ......................................................... 58 5.6.4. Wind Generation Impact on System Production Costs ....................................... 62 5.6.5. Wind Generation Impact on Emissions ............................................................... 64 5.6.6. Changes in Imports and Exports ......................................................................... 67 5.6.7. Changes in Congestion Payments and Uplift...................................................... 71 5.6.8. Changes in Thermal Plant Capacity Factors....................................................... 75 5.7. Results for Task 7 - Identify Transmission System Upgrades: ................................. 76 5.7.1. Identification of Bottled Wind Resources............................................................. 76 5.7.2. Overview of Transmission Upgrades .................................................................. 77 5.7.3. Transmission Upgrades for 6,000 MW Buildout.................................................. 78 5.7.4. Step 1: Tower contingency mitigations................................................................ 79 5.7.5. Step 2 Upgrades:................................................................................................. 80 5.7.6. Step 3 Upgrades:................................................................................................. 81 5.7.7. Step 4 Upgrades:................................................................................................. 82 5.7.8. Step 5 Upgrades:................................................................................................. 83 5.7.9. Steps 6 - 7 Upgrades:.......................................................................................... 84 5.7.10. Alternative Solution.............................................................................................. 85 5.7.11. Transmission Upgrades for the 8,000 MW Buildout............................................ 87 5.7.12. Assessment of the Relative Cost/Value of the Benefits of the Transmission Upgrades Studied ................................................................................................................. 88 5.7.13. Summary of Transmission Upgrades Analysis.................................................... 90

6. Wind Study Conclusions.....................................................................................................91 6.1. Overall Study Findings .................................................................................................. 91 6.2. Summary of Study Results............................................................................................ 92 6.2.1. Reliability Finding: ............................................................................................... 92 6.2.2. Operation and Dispatch Simulation Findings: ..................................................... 93 6.2.3. Resource Adequacy Findings: ............................................................................ 94 6.2.4. Production Cost Simulation Findings: ................................................................. 94 6.2.5. Environmental Findings: ...................................................................................... 95 6.2.6. Transmission Planning Findings: ........................................................................ 96 NYISO Wind Generation Study l August 2010

Figures Figure 5.1: High Speed Cutout Event approx. 12 noon on 6/10/08 ................................................ 7 Figure 5.2: Power System Time Scales ........................................................................................ 11 Figure 5.3: Hourly Wind Output for 8,000 MW of Wind ................................................................. 14 Figure 5.4: Normal Distribution...................................................................................................... 16 Figure 5.5: Load, Net Load and Wind for the Peak Day of July 2013 with 6,000 MW of Wind..... 17 Figure 5.6: Load and Net Load for the Peak Day of July 2013 with 6,000 MW of Wind ............... 17 Figure 5.7: Load and Net Load for the Simulated 2018 Summer Peak Week .............................. 19 Figure 5.8: Load and Net Load for the Simulated 2018 Winter Peak Week ................................. 19 Figure 5.9: Distribution of 60-minute Deltas for Load and Net Load with 8 GW of Wind .............. 20 Figure 5.10: Net-load  (adjusted for load growth) VS Installed Wind in MWs............................. 21 Figure 5.11: Normalized  VS Installed Wind MWs ...................................................................... 22 Figure 5.12: Monthly  of the 10-min. net-load  for 2018 Based on 2006 Wind Data ................ 22 Figure 5.13: Sigma by the Hour of the Day for 5-min. Net-Load  for 2008 ................................. 23 Figure 5.14: Sigma by the Hour of the Day for 5-min. Net-Load  for 2018 ................................. 23 Figure 5.15: Current and Proposed Regulation Requirements ..................................................... 30 Figure 5.16: Current and Proposed Regulation Requirements ..................................................... 30 Figure 5.17: Current and Proposed Regulation Requirements ..................................................... 31 Figure 5.18: 2018 Hourly Loads, Wind Generation and Ramps for the Week of Peak Wind Generation ............................................................................................................................ 32 Figure 5.19: 2018 Hourly Net-Load Ramps VS Load Ramps for the Week of Peak Wind Generation ............................................................................................................................ 32 Figure 5.20: 2018 Hourly Loads, Wind Generation and Ramps for the Week of the Peak Load . 33 Figure 5.21: 2018 Hourly Net-Load Ramps VS Load Ramps for the Week of the Peak Load ..... 33 Figure 5.22: Annual Duration Curve for 5-minute Ramp Events................................................... 34 Figure 5.23: Top 50 Hours of 5-Minute Up Ramp Events ............................................................. 35 Figure 5.24: Top 50 Hours of 5-Minute Down Ramp Events ........................................................ 35 Figure 5.25: Annual Duration Curve for 1 Hour Ramp Events ...................................................... 36 Figure 5.26: Top 50 Hours of 1 Hour Up Ramp Events ................................................................ 36 Figure 5.27: Top 50 Hours of 1 Hour Down Ramp Events............................................................ 37 Figure 5.28: Annual Duration Curve for 4 Hour Ramp Events ...................................................... 37 Figure 5.29: Commitment of Thermal Fossil Generation VS Installed Wind 2013 Loads............. 39 Figure 5.30: Commitment of Thermal Fossil Generation VS Installed Wind 2018 Loads............. 39 Figure 5.31: Ratio of the Committed Fossil Fired Generation MW to the Energy Produced ........ 40 Figure 5.32: Ratio of the Committed Fossil Fired Generation MW to the Energy Produced ........ 41 Figure 5.33: Fossil Fuel Ramping Capability VS Net-Load Ramp ................................................ 42 Figure 5.34: New York Transmission Map Displaying (circles) Where Local Transmission Facilities Limit Wind Plant Output ......................................................................................... 47 Figure 5.35: Transmission Facilities Zones A, B and C ................................................................ 48 Figure 5.36: Transmission Facilities Zones D and E..................................................................... 49 Figure 5.37: LBMP for 2013 VS Wind Penetration for NYCA ....................................................... 52 Figure 5.38: 2013 LBMP VS Wind Penetration for Superzones A-E ............................................ 53 Figure 5.39: 2013 LBMP VS Wind Penetration for Superzones F-I .............................................. 53 Figure 5.40: 2013 LBMP VS Wind Penetration for Superzones J-K ............................................. 54 Figure 5.41: 2018 LBMP VS Wind Penetration for the NYCA....................................................... 55 Figure 5.42: 2018 LBMP VS Wind Penetration for Superzones A-E ............................................ 55 Figure 5.43: 2018 LBMP VS Wind Penetration for Superzones F-I .............................................. 56 Figure 5.44: 2018 LBMP VS Wind Penetration for Superzones J-K ............................................. 56 Figure 5.45: 2018 LBMP VS Wind Penetration for NYCA for the no Unit Commitment and 10%

MAPE Commitment Sensitivity ............................................................................................. 57 Figure 5.46: Fuel Types Displaced for 2013 for the NYCA ........................................................... 58 Figure 5.47: Fuel Types Displaced for 2013 for Superzone A-E................................................... 59 Figure 5.48: Fuel Types Displaced for 2013 for the Superzone F-I .............................................. 59 Figure 5.49: Fuel Types Displaced for 2013 for the Superzone J-K ............................................. 60 NYISO Wind Generation Study l August 2010

Figure 5.50: Fuel Types Displaced for 2018 for the NYCA ........................................................... 60 Figure 5.51: Fuel Types Displaced for 2018 for the Superzone A-E............................................. 61 Figure 5.52: Fuel Types Displaced for 2018 for the Superzone F-I .............................................. 61 Figure 5.53: Fuel Types Displaced for 2018 for the Superzone J-K ............................................. 62 Figure 5.54: Change in Production Costs for 2013 as the Level of Installed Wind Generation Increases .............................................................................................................................. 63 Figure 5.55: Change in Production Costs for 2018 as the Level of Installed Wind Generation Increases .............................................................................................................................. 63 Figure 5.56: Reductions in CO2 (short tons) as Wind Generation Increases for 2013 .................. 64 Figure 5.57: Reductions in NOx (short tons) as Wind Generation Increases for 2013.................. 65 Figure 5.58: Reductions in SO2 (short tons) as Wind Generation Increases for 2013 .................. 65 Figure 5.59: Reduction in CO2 (short tons) as Wind Generation Increases for 2018 .................... 66 Figure 5.60: Reduction in NOx (short tons) as Wind Generation Increases for 2018 ................... 66 Figure 5.61: Reduction in SO2 (short tons) as Wind Generation Increases for 2018 .................... 67 Figure 5.62: Imports for 2013 as Wind Plant Penetration Increases............................................. 68 Figure 5.63: Exports for 2013 as Wind Plant Penetration Increases ............................................ 68 Figure 5.64: Net Import/Exports for 2013 as Wind Plant Penetration Increases .......................... 69 Figure 5.65: Import for 2018 as Wind Plant Penetration Increases .............................................. 69 Figure 5.66: Export for 2018 as Wind Plant Penetration Increases .............................................. 70 Figure 5.67: Net Import/Exports for 2018 as Wind Plant Penetration Increases .......................... 70 Figure 5.68: Congestion for 2013 by the Level of Installed Wind for Superzones F-I................... 72 Figure 5.69: Congestion for 2013 by the Level of Installed Wind for Superzones J-K.................. 72 Figure 5.70: Congestion for 2018 by the Level of Installed Wind for Superzones F-I................... 73 Figure 5.71: Congestion for 2018 by the Level of Installed Wind for Superzones J-K.................. 73 Figure 5.72: Uplift Costs for 2013 as the Level of Installed Wind Increases................................. 74 Figure 5.73: Uplift Costs for 2018 as the Level of Installed Wind Increases................................. 74 Figure 5.74: Transmission Map with 230kV Upgrade (dotted line) ............................................... 86 Figure 5-75: Summary Upgrade Cost/MWH ................................................................................. 89 NYISO Wind Generation Study l August 2010

Tables Table 5-1: List of Wind Plant Units ................................................................................................ 12 Table 5-2: Simulated Minimum Loads and Minimum Net-Loads .................................................. 18 Table 5-3: Trough to Peak Maximum Increases for Summer and Winter 2018............................ 18 Table 5-4: Annual Load/Net-Load   by Timeframes, Load Levels and Wind Penetration ........ 20 Table 5-5: Projected Peak Loads and Wind Plant Penetration Levels.......................................... 25 Table 5-6: Regulation Requirements: Historical (Pre-Study) Sunday & Weekday Requirements 26 Table 5-7: Legend ......................................................................................................................... 28 Table 5-8: Regulation Requirements: Study Year 2011 (34,768MW Peak Load) ........................ 28 Table 5-9: Regulation Requirements: Study Year 2013 (35,475MW Peak Load) ........................ 29 Table 5-10: Regulation Requirements: Study Year 2018 (37,130MW Peak Load) ...................... 29 Table 5-11: Expected Capacity Factors for Wind Plants............................................................... 44 Table 5-12: Powerflow Case Summaries ...................................................................................... 50 Table 5-13: Description of Contingencies Selected for Testing .................................................... 51 Table 5-14: Simulation Results...................................................................................................... 51 Table 5-15: Summary of Average LBMP for 2013 ........................................................................ 54 Table 5-16: Summary of Average LBMP for 2018 ........................................................................ 57 Table 5-17: Thermal Plant Capacity Factors for the NYISO ......................................................... 75 Table 5-18: Thermal Plant Capacity Factors for Superzone A-E .................................................. 75 Table 5-19: Thermal Plant Capacity Factors for Superzone F-I.................................................... 75 Table 5-20: Thermal Plant Capacity Factors for Superzone J-K................................................... 75 Table 5-21: Summary of Base Case Wind Resource Bottling ...................................................... 78 Table 5-22: Summary of Wind Resource Bottling - 6,000 MW Base Case Upgrades ................. 85 Table 5-23: Comparison of Watertown Alternatives - 6,000 MW Case........................................ 86 Table 5-24: Bottled Energy (MWh) Summary - 6,000 MW Case ................................................. 87 Table 5-25: Summary of Wind Resource Bottling - 8,000 MW Base Case Upgrades ................. 87 Table 5-26: Cumulative Transmission Upgrade Costs ($1000) .................................................... 88 Table 5-27: Cumulative Unbottled Wind Plant Energy Production (MWH)................................. 88 Table 5-28: 15 Year Annualized Transmission Upgrade Cost/MWH ............................................ 89 NYISO Wind Generation Study l August 2010

Executive Summary

1. Introduction In 2004, the New York State Public Service Commission (PSC) adopted a Renewable Portfolio Standard (RPS) that requires 25% of New York States electricity needs to be supplied by renewable resources by 2013. The development of the RPS prompted the New York Independent System Operator (NYISO) and the New York State Energy Research and Development Authority (NYSERDA) to co-fund a study which was designed to conduct a comprehensive assessment of wind technology, and to perform a detailed technical study to evaluate the impact of large-scale integration of wind generation on the New York Power System (NYPS). The study was conducted by GE Power System Energy Consulting in fall of 2003 and completed by the end of 2004 (i.e., the 2004 Study).

The overall conclusion of the 2004 Study was the expectation that the NYPS can reliably accommodate up to a 10% penetration of wind generation or 3,300 megawatts (MW) with only minor adjustments to and extensions of its existing planning, operation, and reliability practices. Since the completion of the 2004 Study, a number of the recommendations contained in the report have been adopted. They include the adoption of a low voltage ride through standard, a voltage performance standard and the implementation of a centralized forecasting service for wind plants.

The nameplate capacity of installed wind generation has now increased to 1,275 MW and the NYISO interconnection queue significantly exceeds the 3,300 MW that was originally studied. In addition, the PSC has increased New York States RPS standard to 30% by 2015. As a result, the NYISO has been studying the integration of installed wind plants with nameplate ratings that total from 3,500 MW to 8,000 MW.

From an operational perspective, power systems are dynamic, and are affected by factors that change each second, minute, hour, day, season, and year. In each and every time frame of operation, it is essential that balance be maintained between the load on the system and the available supply of generation. In the very short time frames (seconds-to-minutes), bulk power system reliability is almost entirely maintained by automatic equipment and control systems, such as automatic generation control (AGC). In the intermediate to longer time frames, system operators and operational planners are the primary keys to maintaining system reliability. The key metric driving operational decisions in all time frames are the amount of expected load and its variability. The magnitude of these challenges increases with the significant addition of wind-generating resources.

Variable generation, such as wind and solar, have high fixed costs and very low marginal operating costs which tend to reduce overall production costs and marginal energy prices. However, as will be shown in this study, variable resources require additional resources to be available to respond to the increased system variability, which offsets some of the production cost savings. The primary focus of this report is on the technical impacts of increasing the penetration of wind resources. The impact on production costs, locational-based marginal prices, congestion costs and uplift are presented based on the production costs simulations that were conducted. The study did not conduct, nor did the study scope contemplate, a full economic evaluation of the costs and benefits of wind generation.

NYISO Wind Generation Study l August 2010 i

2. Technical Approach Due to its variable nature and the uncertainty of its output, the pattern of wind generation has more in common with load than it does with conventional generation. Therefore, the primary metric of interest in assessing the impact of wind on system operations is net load, which is defined as the load minus wind. It is net load to which dispatchable resources consisting of primarily fossil fired generation must be able to respond. The study evaluated the impact of up to 8,000 MW of wind-generation resources on system variability. The study process consisted of the following tasks:

Task 1: Develop wind generator penetration scenarios for selected study years including MW output profile and MW load profile.

Task 2: Develop and implement performance-monitoring processes for operating wind generators.

Task 3: Update the review of the European experience conducted for the 2004 study with currently existing wind plants, and review the experiences and studies for wind plants in other regions of the US and Canada.

Task 4: Study the potential impact on system operations of wind generators at various future levels of installed MW for the selected study years as it relates to regulation requirements and the overall impact on ramping.

Task 5: Evaluate the impact of the higher penetration of wind generation from a system planning perspective - including the evaluation of transmission limitations - by identifying specific transmission constraints (limiting element/contingency) for each wind project (or group of projects)

Task 6: Evaluate the impact of the higher penetration of wind generation on the overall system energy production by fuel types, locational-based marginal prices (LBMP), congestion cost, operating reserves, regulation requirements, and load following requirements.

Task 7: Identify the impact of transmission constraints on wind energy that is not deliverable (i.e.,

bottled) and identify possible upgrades for the limiting elements/transmission facilities.

The technical analysis required by the study task includes a set of sequential steps that are needed to successfully conduct a comprehensive analysis of integrating wind into the grid as a function of penetration level. In addition to the traditional planning analysis and economic assessments, the integration of a variable generation resources requires the assessment of operational issues as well.

Operational analyses in conjunction with traditional planning assessments are necessary to fully understand the overall technical implication and potential cost associated with integrating variable generation resources. This process includes the following steps:

Step 1: A determination of the interconnection point of the resources and potential output Step 2: A thorough assessment of the transmission system to determine the contingencies and constraints that could adversely impact wind NYISO Wind Generation Study l August 2010 ii

Step 3: A statistical analysis of the interaction of load and wind as measured by the net load to determine the impact of variable wind resources on overall system variability and operational requirements Step 4: Dispatch simulation with a production cost tool to determine the amount of wind that will be constrained and the impact of wind on the overall dispatchablility such as plant commitment and economics of the system Step 5: An identification and rank ordering of the transmission constraints that impact the dispatchablility of wind Step 6: Development of transmission upgrades to relieve wind constraints for the various penetration levels of wind Step 7: Redo Step 4 with upgrades and needed operational adjustments determined in Step 3 to determine the full impact Step 8: Conduct a dynamic assessment to determine if the planned system with the higher levels of wind will satisfy stability criteria Step 9: Conduct loss-of-load-expectation (LOLE) analysis to determine the impact of installed wind on system load carrying capability or reserve margin requirements.

The study spanned a period of time from the spring of 2008 to the spring of 2010 and involved an extensive review of not only the New York Control Area (NYCA) bulk power system, but the underlying 115 kV transmission system as well. It also involved significant feed-forward and feedback between the power flow analysis and the simulation of NYISO security constrained economic dispatch. This process was used to determine the impact of transmission constraints on the energy deliverability of the wind plants as well as how relieving the transmission constraints affected the energy deliverability of the wind plants. Given the study scope and the plant-by-plant analysis, this study is one of the most comprehensive assessments undertaken for evaluating wind integration for a large balancing area.

NYISO Wind Generation Study l August 2010 iii

3. Study Findings The study has determined that as the level of installed wind plant generation increases, system variability, as measured by the net-load, increases for the system as whole. The increase exceeds 20% on an average annual basis for the 8 GW wind scenario and the 2018 loads. The level of increase varies by season, month, and time-of-day. This will result in higher magnitude ramping events in all timeframes. Ramp is the measure of the change in net load over time to which the dispatchable resources need to respond. Study results are reported for the New York system as a whole and for three superzones (Western load zones A-E, Hudson Valley load zones F-I, and the New York City and Long island load zones J-K). The study resulted in the following findings with respect to system reliability, system operations and dispatch, and transmission planning.

3.1 Reliability Finding:

This study has determined that that the addition of up to 8 GW of wind generation to the New York power system will have no adverse reliability impact. The 8 GW of wind would supply in excess of 10%

of the systems energy requirement. On a nameplate basis, 8 GW of wind exceeds 20% of the expected 2018 peak load. This finding is predicated on the analysis presented in this report and the following NYISO actions and expectations:

The NYISO has established a centralized wind forecasting system for scheduling of wind resources and requires wind plants to provide meteorological data to the NYISO for use in forecasting their output. This item was approved by the Federal Energy Regulatory Commission (FERC) and implemented by the NYISO in 2008.

The NYISO is the first grid operator to fully integrate wind resources with economic dispatch of electricity through implementation of its wind energy management initiative. If needed to maintain system security, the NYISO system operators can dispatch wind plants down to a lower output. This item was approved by the Federal Energy Regulatory Commission (FERC) and implemented by the NYISO in 2009.

The NYISOs wind plant interconnection process requires wind plants: 1) To participate fully in the NYISOs supervisory control and data acquisition processes; 2) To meet a low voltage ride through standard; and 3) conduct voltage testing to evaluate whether the interconnection of wind plants will have an adverse impact on the system voltage profile at the point of interconnection. In addition, the NYISO will continue to integrate best practice requirements into its interconnection processes.

The NYISOs development of new market rules assist in expanding the use of new energy storage systems that complement wind generation. This item was approved by the Federal Energy Regulatory Commission (FERC) and implemented by the NYISO in 2009.

The NYISOs installed resource base will have sufficient resources to support wind plant operations.

As described in this report, the overall availability of wind resources is much less than other resources and their variability (changing output as wind speed changes) increases the magnitude of the ramps.

For a system that meets its resource adequacy criteria (e.g., the 1 day in ten years), the additions of 1 MW of resources generally means that 1 MW of existing resources could be removed and still meet the resource adequacy criteria. However, the addition of 1 MW of wind would allow approximately 0.2 NYISO Wind Generation Study l August 2010 iv

MW to 0.3 MW of existing resources to be removed in order to still meet the resource adequacy criteria. The balance of the conventional generation must remain in service to be available for those times when the wind plants are unavailable because of wind conditions and to support larger magnitude ramp events.

3.2 Operation and Dispatch Simulation Findings:

Analysis of the wind plant output and dispatch simulations resulted in the following findings for the expected impact of wind plant output on system operations and dispatch:

Finding One - Analysis of five minute load data coupled with a ten minute persistence for forecasting wind plant output (i.e., wind plant output was projected to maintain its current level for the next five minute economic dispatch cycle) concluded that increased system variability will result in a need for increased regulation resources. The need for regulation resources varies by time of day, day of the week and seasons of the year. The analysis determined that the average regulation requirement increases approximately 9% for every 1,000 MW increase between the 4,250 MW and 8,000 MW wind penetration level. The analysis for 8 GW of wind and 2018 loads (37,130 MW peak) resulted in the overall weighted average regulation requirement increasing by 116 MW. The maximum increase is 225 MW (a change from a 175 MW requirement up to 400 MW) for the June-August season hour beginning (HB) 1400. The highest requirement is 425 MW in the June-August season HB2000/HB2100.

Finding Two - The amount of dispatchable fossil generation committed to meet load decreases as the level of installed nameplate wind increases. However, a greater percentage of the dispatchable generation is committed to respond to changes in the net-load (load minus wind) than committed to meet the overall energy needs of the system. The magnitudes of ramp or load following events are reduced when wind is in phase with the load (i.e., moving in the same direction). However, for many hours such as the morning ramp or the evening load drop, wind is out of phase with the load (i.e.,

moving in the opposite direction). These results in ramp or net-load following events that are of higher magnitude than those that would result from changes in load alone. It is these ramp or load following events to which the dispatchable resources must respond.

Finding Three - Simulations with 8 GW of installed wind resulted in hourly net-load up and down ramps that exceeded by approximately 20% the ramps that resulted from load alone. It was also determined from the simulations the NYISO security constrained economic dispatch processes are sufficient to reliably respond to the increase in the magnitude of the net-load ramps. This finding is based on the expectation that sufficient resources will be available to support the variability of the wind generation. For example, the data base used for these simulations had installed reserve margins which exceeded 30%.

Finding Four - Simulations for 8 GW of wind generation concluded that no change in the amount of 1

operating reserves was needed to cover the largest instantaneous loss of source or contingency event. The system is designed to sustain the loss of 1,200 MW instantaneously with replacement within ten minutes where as a large loss of wind generation occurs over several minutes to hours. The 1

Operating reserves is the amount of resources that are needed to be available for real-time operations to cover the instantaneous and unexpected loss of resources. The New York power system is operated to protect the system against the sudden loss of 1,200 MW of resources. Operating reserve as stated is an operational concept while the reserve margin discussed in section 3.3 is a planning concept. The required reserve margin is designed to maintain, at an acceptable level, the risk of not having sufficient resources to avoid an involuntary loss-of-load event.

NYISO Wind Generation Study l August 2010 v

analysis of the simulated data found for 8 GW of installed wind a maximum drop in wind output of 629 MW occurred in ten minutes, 962 MW in thirty minutes and 1,395 MW in an hour, respectively.

3.3 Resource Adequacy Findings:

To evaluate the impact of wind resources on NYISO installed reserve requirements, the study started 2

with the New York State Reliability Council (NYSRC) Installed Reserve Margin Study for the 2010-3 2011 Capability Year. The NYSRC base case had an installed reserve margin of 17.9% to meet loss-of-load-expectation (LOLE) criteria of 0.1 days per year. That base case was updated to bring the installed wind resources to the full 8 GW of wind studied. The analysis of a system with this level of installed wind resulted in the following findings.

Finding One - All other things being equal, the addition of 8 GW of wind resources to the NYSRC base case reduced the LOLE from the 0.1 days per year to approximately 0.02 days per year.

Finding Two - To meet the required reliability criteria, the NYISO reserve margin would have to increase from its current level of 18% to almost 30% with 8 GW of nameplate wind as part of the resource mix. This was determined by using the methodology of removing capacity to bring the system to criteria and adding transfer capability in order for the wind plants to qualify for Capacity Rights Interconnection Service (CRIS). However, it should be noted that the NYISOs capacity market requires load serving entities to procure unforced capacity (UCAP) and capacity is derated to its UCAP value for purchase. As a result the total amount UCAP that needs to be purchased to meet reliability criteria remains essentially unchanged. The increase in reserve margin is because on capacity basis 1 MW of wind is equivalent to approximately 0.2 MW to 0.3 MW of conventional generation. Therefore, it requires a lot more installed wind to provide the same level of UCAP as a conventional generator. This results in an increase in the installed reserve margin which is computed on an installed nameplate basis.

Finding Three - The LOLE analysis resulted in an effective load carrying capability (ELCC) for the wind plants studied that exceeded 20%. The ELCC for this study exceeded the ELCC finding in the 2004 study by a factor of 2. Off-shore wind exhibits ELCC that is higher than on-shore wind because a greater percentage of the off-shore wind plants energy production occurs during peak hours. As an example, the GridView wind plant simulations based on 2006 wind data resulted in a 37.4% overall annual capacity factor (CF) for off-shore wind VS 34.3% for on-shore wind. However, the CF for off-shore wind plants during peak hours (the hours between 7am and 11 pm weekdays) was 39.7% for off-shore wind VS 32.5% for on-shore wind.

3.4 Production Cost Simulation Findings:

The production cost simulations conducted with ABBs GridView economic dispatch simulation model and the base case transmission system resulted in the following findings:

2 Reserve margin is the amount of additional capacity above the peak load that is needed so that the risk of not having sufficient capacity available to meet the load meets the minimum reliability criteria. It is expressed as a percentage and is calculated by dividing the required level of resources by the expected peak load. Resources can be unavailable because of equipment failure, maintenance outage, lack of fuel, etc. The higher the unavailability of the overall resource mix, the higher the installed reserve margin will be.

3 http://www.nysrc.org/NYSRC_NYCA_ICR_Reports.asp NYISO Wind Generation Study l August 2010 vi

Finding One - As the amount of wind generation increases, the overall system production costs decrease. For the 2013 study year, the production costs drop from the base case total of almost $6 billion to a level of approximately $5.3 billion for the 6,000 MW wind scenario. This represents a drop of 11.1% in production costs. For the 2018 study year, the production costs drop from the base case total of almost $7.8 billion to a level of approximately $6.5 billion for the 8,000 MW wind scenario. This represents a drop of 16.6% in production costs. The change in production costs reflect the commitment of resources that are needed to support the higher magnitude ramping events but do not reflect the costs of the additional regulating resources.

Finding Two - Based on the economic assumptions used in the CARIS study, locational-based marginal prices (LBMP) or spot prices decline as significant amounts of essentially zero production cost generation that participates as price taker is added to the resource mix. For the 2018 simulations, the NYISO system average LBMP prices are 9.1% lower for the 8 GW wind scenario when compared to the base case or 1,275 MW of installed wind.

Finding Three - The LBMP price impacts are greatest in the superzones where the wind generation is located and tends to increase the price spread between upstate where wind is primarily located in the study and downstate, which implies an increase in transmission congestion.

Finding Four - The primary fuel displaced by increasing penetration of wind generation is natural gas.

For the simulations with 8 GW of wind with 2018 loads, the total amount of fossil-fired generation displaced was approximately 15,500 GWh. Gas-fired generation accounted for approximately 13,000 GWh or approximately 84% of the total. Oil and coal accounted for approximately 2,050 GWh and 465.1 GWh respectively, or approximately 13% and 3% of the total fossil generation displaced.

Finding Five - As suggested by the LBMP trends, the congestion payments in superzones F-I and J-K increase as the level of installed wind generation is increased. The overall increase in congestion payments on a percentage basis as measured against the base case compared to 6,000 MW of wind in 2013 and 8,000 MW in 2018 ranges from a high of 85% for superzone F-I in 2013 to a low of 64%

for superzone J-K in 2018.

Finding Six - The addition of wind resources to superzone J-K in the 2018 case puts downward pressure on LBMPs in those zones, and therefore lowers congestion payments.

Finding Seven - Uplift costs tend to increase in superzones A-E and F-I as the level of installed wind generation increases. Superzone J-K uplift cost are for the most part flat as the level of installed wind increases for 2013 but actually decreases for 2018. This is the result of the offshore wind which has a capacity factor of almost 39% and tends to be more coincident with the daily load cycle and displaces high cost on peak generation in the superzone while requiring less capacity for higher magnitude ramping events. Off shore wind also provides greater capacity benefits.

Finding Eight - The capacity factors for the thermal plants are, as expected, decreased by the addition of wind plants, but this is partially offset by increasing load. The biggest reduction in annual capacity factors from the 2013 base case level of 1,275 MW of wind when compared to the 8 GW scenarios occurs for the combined cycle plants in all superzones with a 30% decline in superzone A-E, 11% decline in superzone F-I and 6% decline superzone J-K. As would be expected the biggest impact is in the superzone with the highest level of installed wind with transmission capacity limitations between the superzones contributing to the reduction.

NYISO Wind Generation Study l August 2010 vii

3.5 Environmental Findings:

For the 2018 load levels, the dispatch simulations with 8 GW of wind resources resulted in the following emissions reductions in comparison to the base case with 1,275 MW of installed wind:

Finding One - A CO2 emission reduction of approximately 4.9 million short tons or a reduction of 8.5%.

Finding Two - Each GWh of displaced fossil-fired generation which primarily consisted of natural gas resulted in an average reduction in CO2 of approximately 315 tons.

Finding Three - A NOx emission reduction of approximately 2,730 short tons or a reduction of 7%.

Finding Four - A SO2 emissions reduction of 6,475 short tons or a reduction of 9.7%.

3.6 Transmission Planning Findings:

Extensive power flow analysis in conjunction with dispatch simulations was conducted to determine the impact of transmission system limitations on the energy deliverability of the wind plant output. The analysis resulted in the following findings:

Finding One - Given the existing transmission system capability, the 6 GW scenario determined that 8.8% of the energy production of the wind plants in three areas in upstate New York would be bottled or not deliverable.

Finding Two - The primary location of the transmission constraints was in the local transmission facilities or 115 kV voltage level.

Finding Three - The off-shore wind energy as modeled was fully deliverable and feeds directly into the superzone J-K load pockets.

Finding Four - The study evaluated 500 miles of transmission lines and 40 substations to determine potential upgrades that would result in the unbottling of the wind energy.

Finding Five - If all the upgrades studied were implemented, the amount of wind energy not deliverable would be reduced to less than 2% of the upstate wind.

Finding Six - Depending on the scope of upgrades required, such as reconductoring of transmission lines compared to rebuilding or upgrading terminal equipment, the cost of the upgrades could range from $75 million to $325 million. However, it should be noted that many of the transmission facilities studied are approaching the end of their expected useful lives.

Finding Seven - Transient Stability Analysis was conducted to evaluate the impact of high wind penetration on NYCA system stability performance. The primary interface tested was the Central East.

The Central East stability performance has been shown historically to be key factor in the dynamic performance of the NYISO power grid. The NYISO power grid (and the Interconnection) system demonstrated a stable and well damped response (angles and voltages) for all the contingencies tested on high wind generation on-peak and off-peak cases. There is no indication of units tripping due to over/under voltage or over/under frequency.

NYISO Wind Generation Study l August 2010 viii

Finding Eight - Wind plants that are in the NYISO interconnection 2008 class year study and beyond may require system deliverability upgrades to qualify for Capacity Resource Integration Service (CRIS). This totals approximately 4,600 MW of new nameplate wind plants that were included in the study. In order to qualify for capacity payments, the wind plants in class year 2009/2010 and beyond in upstate New York would need to increase transmission transfer capability between upstate New York and southeast New York (a.k.a., the UPNY-SENY interface). This transmission interface primarily consists of 345 kV transmission lines in the Mid-Hudson valley region running through Greene County, New York south of Albany to Dutchess County, New York or between Zones E and F and Zone G. The study determined that approximately 460 MW of interface transfer capability needs to be added to this interface for the wind plants that did not qualify for capacity payments to be eligible for them. This does not impact the deliverability of the wind plants energy but only their ability to qualify for capacity payments or CRIS.

NYISO Wind Generation Study l August 2010 ix

4.

Conclusions:

The primary finding of the study is that wind generation can supply reliable clean energy at a very low cost of production to the New York power grid. This energy results in significant savings in overall system production costs, reductions in greenhouse gases such as CO2 and other emissions such as NOx and SO2 as well an overall reduction in wholesale electricity prices. However, wind plants require a significant upfront capital investment. In addition, wind plants, because of their variable nature and the uncertainty of their output, provide a greater challenge to power system operation than conventional power plants. This study determined that the NYISOs systems and procedures (which include the security constrained economic dispatch and the practices that have been adopted to accommodate wind resources) will allow for the integration of up to 8 GW of installed wind plants without any adverse reliability impacts.

This conclusion is predicated on the assumption that a sufficient resource base is maintained to support the wind. The study determined that 8 GW of wind would reduce the need for conventional or dispatchable fossil fired generation on the order of 1.6 to 2 GW or an amount equivalent to 20-25% of the installed nameplate wind. This is the result of the much lower overall availability of wind-produced Energy, when compared to conventional generation. This means an amount of fossil generation equivalent to 75-80% of the nameplate installed wind needs to be available for those times when the wind isnt blowing or the wind plant output is at very low levels. Non-wind generation is needed to respond to the higher magnitude ramps that will result because of winds variable nature.

As wind resources are added to the resource mix, their lower availability could result in an increase in the installed reserve margin and a decline in spot market prices. The impact of these changing conditions has not been analyzed in this report.

The fluctuating nature and the uncertainty associated with predicting wind plant output levels manifests itself as an increase in overall system variability as measured by the net load (load minus wind). In response to these increased operational challenges the NYISO has implemented changes to its operational practices such as being the first ISO to incorporate variable generation resources into security constrained economic dispatch (SCED) and to implement a centralized forecasting process for wind resources. The study concluded that at higher levels of installed wind generation the system will experience higher magnitude ramping events and will require additional regulation resources to respond to increased variability during the five minute dispatch cycle. The analysis determined that the average regulation requirement will need to increase by approximately 9% for every 1,000 MW increase in wind generation between the 4,250 MW and 8,000 MW.

Although the addition of wind to the resource mix resulted in significant reduction in production costs, the reduction would have been even greater if transmission constraints between upstate and downstate were eliminated. These transmission constraints prevent lower cost generation in upstate New York from displacing higher costs generation in southeast New York. This report did not analyze the potential financial impact of an increase in transfer capability from upstate into southeast New York.

Finally, the study determined that almost 9% of the potential upstate wind energy production will be bottled or not deliverable because of local transmission limitations. The study identified feasible sets of transmission facility upgrades to eliminate the transmission limitations. These upgrades were evaluated to determine how much of the wind energy that was undeliverable would be deliverable if NYISO Wind Generation Study l August 2010 x

the transmission limitations were removed. Additional alternatives were suggested and evaluated to address the significant levels of resource bottling that occurs in the Watertown vicinity. The suggested transmission upgrades and alternatives require a detailed physical review and economic evaluation before a final set of recommendations can be determined.

In addition to the findings presented in this Executive Summary, the main body of the report offers other findings as well as additional support for the findings presented in the executive summary. The report also contains an update of the review of the European experience with variable generation that was part of the 2004 study and there are summaries of wind integration studies by the California ISO, the Ontario Power Authority in Canada and the Electric Reliability Council of Texas.

NYISO Wind Generation Study l August 2010 xi

NYISO Wind Generation Study

1. Purpose This document presents the results of a study of 8,000 MW of wind generation on the New York Control Area -

see map below. The purpose of the study was two fold: 1) To update the GE study that was conducted in 2004 for wind generation up to 3,300 MW; and 2) To identify issues that will need to be addressed and initiatives that will be need to be undertaken to integrate several thousand MW of wind generation. The primary focus of the report is on the technical impacts of increasing the penetration of wind resources. The impact on production costs, locational marginal prices, congestion costs and uplift are presented based on the production costs simulations that were conducted. The study did not conduct nor did the study scope contemplate a full economic evaluation of the costs and benefits of wind generation.

NEW YORK CONTROL AREA LOAD ZONES D A - WEST B - GENESE C - CENTRL D - NORTH E - MHK VL E F - CAPITL G - HUD VL H - MILLWD I - DUNWOD F J - N.Y.C.

K - LONGIL B

E A C B

G H

I J K NYISO Wind Generation Study lJune 2010 1

2. Background The implementation of policies and the adoption of regulations designed to encourage the development of renewable energy technologies is resulting in the significant growth in the installed base of wind generation in the New York Control Area (NYCA) as well as throughout the North America. Given wind generations variable and less predictable nature and technology characteristics, industry experience and studies have indicated that large-scale wind generation has a unique set of impacts on power system operation. While these impacts may be relatively small at low penetration levels, as penetration levels increase, physical transmission system reinforcements and special bulk power system planning and operating practices may be required. Therefore, these potential impacts need to be fully understood to guarantee the reliable operation and planning of the New York Power System (NYPS).

In September of 2004, New York State adopted a Renewable Portfolio Standard that requires 25% of New York States electricity needs be supplied by renewable resources by 2013. This requirement resulted in the New York Independent System Operator and the New York State Energy Research and Development Authority (NYSERDA) co-funding a study, which was designed to conduct a comprehensive assessment of wind technology, and to perform a detailed technical study to evaluate the impact of large-scale integration of wind generation on the NYPS. The study was conducted by GE Power System Energy Consulting in fall of 2003 and completed by the end of 2004 (i.e., the 2004 Study).

The overall conclusion of that study was the expectation that the NYPS can reliably accommodate up to 10%

penetration or 3,300 MW of wind generation with only minor adjustments and extensions to its existing planning, operation, and reliability practices - e.g., forecasting of wind plant output. Also, the finding that no major issues were found in the aggregate does not mean that the potential for significant local interconnection issues or engineering challenges specific to particular site would not be encountered. Such issues would need to be identified through the NYISOs interconnection and electric system planning processes. In addition, the NYISO will continue to evolve its operating and interconnection requirements to implement best practices.

Since the completion of the NYISO/NYSERDA wind study, a number of the recommendations contained in the report have been adopted such as a low voltage ride through standard and a centralized forecasting service for wind plants. Installed nameplate wind generation has now grown to in excess of 1,200 MW and the NYISO interconnection queue significantly exceeds the 3,300 MW that was studied in the 2004 Study. In addition, the cap on eligible wind generation exempt from under generation penalties and eligible to be fully compensated for over-generation was increased from 1,000 MW to 3,300 MW. Finally, the State of New York has increased its RPS standard to 30% by 2015.

NYISO Wind Generation Study l August 2010 2

3. Wind Plant Integration - Issues As a result of these changing conditions and ongoing wind integration issues, the NYISO committed to study the impact of wind generation beyond 3,300 MW. As part of the study process the NYISO identified a set of issues that need to be addressed in order to continue the orderly and reliable integration of continuing growth in wind generation into the NYCA power grid and market operations. These issues include the following:

Transmission: Transmission plays a critical role in the large scale integration of variable generation. A significant amount of new transmission and/or enhanced utilization of existing transmission capability will be needed over the next several years to accommodate and integrate higher levels of wind generation.

System Flexibility: The bulk power system will experience higher magnitude ramping events and to accommodate the increased variability and uncertainty of variable generation the system will need to commit proportionately more dispatchable resources to maintain system flexibility. The resource planning and development frameworks must ensure that the bulk power system has the necessary quantity of flexible supply and demand resources necessary to accommodate generation - e.g., storage capability or off-peak load such as plug-in hybrid electric vehicles. Markets, pricing mechanisms and interconnection standards need to provide signals about the characteristics that are valued both to existing generators and to entities that are planning for new generation.

Operator Awareness and Practices: Enhancements are required to existing operator practices, techniques and decision support tools to increase the operator awareness and to enable the operation of the future bulk power 4

systems with large scale penetration of wind generation. Wind generation must be visible to and controllable by the system operator similar to any other power plant to allow the system operator to maintain reliability. Based on current experience with operating wind plants the NYISO has already developed a FERC approved wind resource management proposal which makes wind plants subject to dispatch signals when system constraints exist.

Forecasting: Short term forecasting techniques used for real time operation must be enhanced to more accurately predict the magnitude and phase (i.e. timing) of wind generation plant output. One area needing increased attention is being able to predict extreme weather events that could result in the rapid loss of wind generation - e.g., high-speed wind cutout.

Wind Generation Plant Performance and Standards: Interconnection and generating plant standards must be enhanced to ensure that variable generating plant design and performance contribute to reliable operation of the power system.

System Models: Improved component model development, validation and standardization for all wind technologies are also required, especially for stability and transient analysis.

4 The NYISO interconnection standards already require wind plants to be visible to system operators.

NYISO Wind Generation Study l August 2010 3

4. Study Tasks and Process The study of wind penetrations in excess of 3,300 MW resulted in the following tasks:

Task 1: Develop wind generator penetration scenarios for selected study years including MW output profile and MW load profile.

Task 2: Develop and implement performance monitoring processes for operating wind generators.

Task 3: Update the review of the European experience conducted for the 2004 study with currently existing wind plants, and review the experiences and studies for wind plants in other regions of the US and Canada.

Task 4: Study the potential impacts on system operations of wind generators at various future levels of installed MW for the selected study years as it relates to regulation requirements and the overall impact on ramping.

Task 5: Evaluate the impact of the higher penetration of wind generation from a system planning perspective -

including the evaluation of transmission limitations - by identifying specific transmission constraints (limiting element/contingency) for each project (or group of projects).

Task 6: Evaluate the impact of the higher penetration of wind generation on the overall system energy production by fuel types, locational based marginal prices (LBMP), congestion cost, operating reserves, regulation requirements, and load following requirements.

Task 7: Identify the impact of transmission constraints on wind energy that is not deliverable (i.e., bottled) because of the transmission constraints and identify possible upgrades for the limiting elements/transmission facilities.

The technical analysis required by the study task includes a set of sequential steps that are needed to successfully conduct a comprehensive analysis of integrating wind into the grid as a function of penetration level.

In addition to the traditional planning analysis and economic assessments, the integration of a variable generation resources requires the assessment of operational issues as well. Operational analyses in conjunction with traditional planning assessments are necessary to fully understand the overall technical implication and potential cost associated with integrating variable generation resources. This process includes the following steps:

Step 1: A determination of the interconnection point of the resources and potential output Step 2: A thorough assessment of the transmission system to determine the contingencies and constraints that could adversely impact wind Step 3: A statistical analysis of the interaction of load and wind as measured by the net load to determine the impact of variable wind resources on overall system variability and operational requirements Step 4: Dispatch simulation with a production cost tool to determine the amount of wind that will be constrained off and the impact of wind on the overall dispatchablility such as plant commitment and economics of the system Step 5: An identification and rank ordering of the transmission constraints that impact the dispatchablility of wind NYISO Wind Generation Study l August 2010 4

Step 6: Development of transmission upgrades to relieve wind constraints for the various penetration levels of wind Step 7: Redo step 4 with upgrades and needed operational adjustments determined in step 3 to determine the full impact Step 8: Conduct a dynamic assessment to determine if the planned system with the higher levels of wind will satisfy stability criteria Step 9: Conduct loss-of-load-expectation (LOLE) analysis to determine the impact of installed wind on system load carrying capability or reserve margin requirements.

NYISO Wind Generation Study l August 2010 5

5. Wind Study Results 5.1. Results for Task 1 - Study Assumptions:

This task resulted in three study years being selected. They are 2011, a near-in year; 2013 which is the target year of the 25% RPS; and 2018, which is the tenth year of the 2009 reliability planning cycle, and is also the first year of the Eastern Interconnection Wind Integration study being conducted by the National Renewable Energy Lab (NREL). The starting point or base assumptions for the wind study was the base case for the 2009 5

Comprehensive Reliability Plan (CRP) for the transmission analysis. The starting point for the production cost 6

simulations was the assumptions in the 2009 Congestion Assessment and Resource Integration Study (CARIS).

Section 4.3.1 of the CARIS report presents the New York Control Area transfer limits that were used for the study including a Central East limit of 2,600 MW. The wind study used the nominal planning limit of 2,800 MW.

Section 4.4 of the CARIS report presents the fuel costs assumptions that were used in the production costs simulations which was the GridView modeling tool used for the CARIS study. Section 4.5 of the CARIS report presents the emission costs that were used in the study. The cost for CO2 or green house gas emissions are approximately $3.50 per ton in 2009 and increase to approximately $6.00 per ton in 2018, with 2013 at approximately $5.00 per ton.

For each of the years, two levels or scenarios of installed nameplate wind plant were developed. They are: 1) 3,500 MWs and 4,250 MWs for 2011 which represents approximately 10% and 12% of the projected peak for that year while 4,250 MWs would supply 6.5% of the forecast energy at a 30% capacity factor; 2) 4,250 MWs and 6,000 MWs for 2013 with 6,000 MWs equal to 17% of the projected peak for that year and 8.9% of forecast energy at a 30% capacity factor; and 3) 6,000 MWs and 8,000 MWs for 2018 while 8,000 MWs of wind is equal to 22.4% of the projected peak for that year and 11.6% of forecast energy at 30% capacity factor. AWS Truepower (formerly know as AWS Truewind) who is the contractor for the wind forecasting service, as well as a contractor to NREL for the Eastern Interconnection Wind Integration study, provided the wind output profiles required for the study.

5.2. Results for Task 2 - Wind Plant Performance Monitoring:

One of the observations made in the initial wind study was that much could be learned from operating wind plants as they came on line. To that end, the NYISO developed a reporting process for tracking the performance of operating wind plants. The report entitled: Daily Wind Plant Performance Tracking Report tracks the performance of wind plants on a daily basis for key metrics such as maximum coincident wind plant output, total output at the time of the system peak, Mwh generated, capacity factor, etc. Appendix A-1 contains the daily summary report for 2009.

Besides daily tracking of wind plant performance, the NYISO has experienced and analyzed rare events such as high-speed cutout which is the result of wind conditions that exceed the capability of the wind turbines causing them to shut down rapidly to protect the equipment. Wind plants can also ramp up quickly as the wind speed picks up suddenly. Wind plants may ramp up quickly as a thunder storm approaches a plant site and then shut down as wind exceeds the capability of the equipment. Figure 5.1 is an example of a high-speed cutout event that NYISO operations observed on June 10, 2008. The figure shows how a front containing thunderstorms moved from west to east affecting wind plants at different locations on the system. Wind plant output is 5

http://www.nyiso.com/public/webdocs/services/planning/reliability_assessments/CRP__FINAL_5-19-09.pdf 6

http://www.nyiso.com/public/webdocs/services/planning/Caris_Report_Final/CARIS_Final_Report_1-19-10.pdf NYISO Wind Generation Study l August 2010 6

expressed as a percent of nameplate. For the first set of plants (red line) to encounter the front, the plants ramp up preceding the cutouts from 26% of nameplate to 61% of nameplate over 30 minutes and then ramp downs from cutouts to 5% of nameplate over 10 minutes. After the storm passes, the plants ramp back up to 82% of nameplate over 45 minutes. A similar pattern is observed later for the plants further to the east (green line).

Figure 5.1: High-Speed Cutout Event approx. 12 noon on 6/10/08 In addition, the NYISO has observed the ability of wind plants to adjust the level of their output rapidly in response to changing system conditions which result in price changes. These operating experiences to date indicate a need to communicate dispatch commands to the wind plant operators on an as needed basis to maintain reliability especially as the amount of installed wind plant MWs increased. Experience with existing wind plants resulted in the NYISO moving forward with a resource management initiative to extend its market-based Security Constrained Economic Dispatch (SCED) systems to wind plants.

The integration of increased levels of wind will be facilitated by using the NYISOs market signals (e.g. location-based marginal prices) and the economic offers submitted by the generation resources, including wind plants, to address reliability issues rather than relying upon manual intervention by operators.

Based on the offers submitted by each wind plant and other resources, SCED will determine the most economic mix of resources to meet real-time security constraints. Allowing wind plants to indicate their economic willingness to operate reduces the need for the NYISO or local system operators to take less efficient, out-of-market actions to protect the reliability of the system.

This results in better utilization of wind plant output while maintaining a secure, reliable system and more accurate LBMP signals.

This wind on dispatch initiative was developed in conjunction with stakeholders, approved by the Federal Energy Regulatory Commission, and has now been implemented.

NYISO Wind Generation Study l August 2010 7

5.3. Results for Task 3 - European, US and Canada Experience with Wind Plants:

The purpose of Task 3 was review of the European experience with existing wind plants and review the experiences and studies for wind plants in other regions of the US and Canada that have been conducted since the 2004 Study. Europe is the region of the world that has highest penetration of wind. The NYISO contracted with Dr. Thomas Ackermann of Energynautics GmbH to provide a report of Europes most recent operating experience with wind. Also, the NYISO reviewed the most recent study work from California, Texas and the Province of Ontario. In addition, the NYISO is participating in the North American Electric Reliability Councils, Inc. (NERC) Integration of Variable Generation Task Force (IVGTF) as well as what is known as the Eastern Interconnection Wind Integration Study. This study includes Department of Energy/NREL, MISO (study lead),

NYISO, PJM, SPP, and TVA.

The primary findings of the report prepared by Dr. Ackerman are as follows:

Europe shows that high/very high wind penetration levels are possible, but those high penetration levels are driven by energy policy (subsidies) and not economics for the most part. This also applies to power system integration issues.

Wind power can be successfully included in markets (Spain/UK).

Transmission helps to achieve benefits of aggregating large-scale wind power development and provides improved system balancing services. This is achieved by making better use of physically available transmission capacity and upgrading and expanding transmission systems. High wind penetrations may also require improvements in grid internal transmission capacity.

European regulators and Transmission System Operators (TSOs) have developed a willingness to learn and question existing rules as well as to adjust rules and regulations. In addition, most European countries have shown a flexibility to adjust their energy policy, rules and regulations depending on the technical and economical development in order to create a low-risk environment for renewable energy projects, without allowing windfall profits as it is very difficult to get all relevant regulatory details right at the first attempt. This flexibility for change has been based on a continuous dialogue between policy makers, regulators, network companies and the renewable energy lobby.

Both load and generation benefit from the statistics of large numbers as they are aggregated over larger geographical areas. Larger balancing areas make wind plant aggregation possible. The forecasting accuracy improves as the geographic scope of the forecast increases; due to the decrease in correlation of wind plant output with distance, the variability of the output decreases as more plants are aggregated. On a shorter-term time scale, this translates into a reduction in reserve requirements; on a longer-term time scale, it produces some smoothing effects on the capacity value. Larger balancing areas or coordination agreements with neighboring areas also give access to more balancing units such as hydro units and the ability to bank energy.

Integrating wind generation information into real-time system operations and with updated forecasts for the day-ahead operations will help manage the variability and forecast errors of wind power. Well-functioning hour-ahead and day-ahead markets including having wind plants respond to dispatch signals can help to more cost-effectively provide balancing energy required by the variable-output wind plants and maintain system security.

Appendix B-1 provides an expanded summary of Dr Ackermanns findings.

The overall conclusion from the California study sponsored by the California ISO (CAL-ISO) can best be summarized by the words of California ISO President & CEO Yakout Mansour: The good news is that this study shows the feasibility of maintaining reliable electric service with the expected level of intermittent renewable NYISO Wind Generation Study l August 2010 8

resources associated with the current 20% RPS, provided that existing generation remains available to provide back-up generation and essential reliability services. The cautionary news is the provided part of our conclusion. Appendix B-2 provides an expanded summary of the CAL-ISO study.

The overall conclusion from the Texas study sponsored by the Electric Reliability Council of Texas (ERCOT) is that through 5,000 MW of wind generation capacity, approximately the level of wind capacity presently in ERCOT (on the order of 5% of the peak), wind generation has limited impact on the system. Its variability barely rises above the inherent variability caused by system loads. At 10,000 MW wind generation capacity, the impacts become more noticeable. By 15,000 MW (on the order of 20% of the peak), the operational issues posed by wind generation will become a significant focus in ERCOT system operations. However, the impacts can be addressed by existing technology and operational attention, without requiring any radical alteration of operations.

Appendix B-3 provides an expanded summary of the ERCOT study.

The Ontario study was sponsored by the Ontario Power Authority (OPA). This study concluded that for all wind scenarios, the increase in hourly and multi-hourly variability, as measured by , due to wind is relatively small (not more than 10% for any scenario). From an hourly scheduling point of view, even 10,000 MW of wind would not push the envelope much further beyond the current operating point. However, the amount and magnitudes of extreme one-hour and multihour net-load changes are significantly greater with high wind penetration. With the addition of 10,000 MW of wind, the maximum one-hour net-load rise increases by 34%, and the maximum one-hour net-load drop increases by 30%. This data indicates that with large amounts of wind, much more one-hour ramping capability is needed for secure operation. Clearly the longest sustained ramping (up and down) occurs during the summer morning load rise and evening load decline periods. During these periods (and others) the units may need to ramp continually over three or more hours. For the year 2020 load with 10,000 MW of wind scenario, the maximum positive three-hour load-wind delta increases by 17% and the maximum negative three-hour delta increases by 33%. The detailed results clearly illustrate the fact that units will have to undergo sustained three-hour ramping more often, and ramp further with the addition of large amounts of wind. Appendix B-4 provides an expanded summary of the OPA study.

As noted above, the NYISO also participated in NERCs Integration of Variable Generation Task Force. In December 2008 in anticipation of the growth of wind and other variable generation, NERCs Planning and Operating Committees created the Integration of Variable Generation Task Force charged with preparing a report to include: 1) philosophical and technical considerations for integrating variable resources into the Interconnection, and 2) specific recommendations for practices and requirements, including reliability standards, that cover the planning, operations planning, and real-time operating timeframes.

The goals of this report were to:

Raise industry awareness and the understanding of characteristics of variable generation Raise industry awareness and the understanding of the challenges associated with large scale integration of variable generation Investigate the impacts on traditional approaches used by system planners and operators to plan, design and operate the power system Scan NERC Standards, FERC rules and business practices to identify possible gaps and future requirements to ensure bulk power system reliability in light of large scale integration of variable resource 7

The final document was issued on April 16, 2009 and is available on the NERC website .

In conclusion, the primary insights that can be drawn from the review of the European and other studies and the NERC draft report are as follows:

7 http://www.nerc.com/files/IVGTF_Report_041609.pdf NYISO Wind Generation Study l August 2010 9

Higher levels of installed wind generation above the 3,300 MW from a system operation perspective are feasible.

Achieving a higher level of wind penetration will most likely require the implementation of enhancements to and extension of existing operating protocols, procedures and reliability standards.

The major areas of ongoing concern that are common across all regions tend to focus on the following questions:

Will there be sufficient transmission infrastructure to integrate the higher penetrations of wind?

Will sufficient resources be available when the higher penetration of wind generation are achieved to provide the operational flexibility that will be needed with higher penetration of variable generation?

Validation of wind turbine models needed for system studies.

5.4. Results for Task 4 - Assessing the Impact of Wind Plants on System Operations:

5.4.1. Introduction The focus of Task 4 is to study the impacts on system operations of the penetration of installed wind plants above 3,300 MWs. The impact of increasing wind penetration from its current installed nameplate of 1,274 MW up to 8,000 MW on such operational parameters as regulation requirements, load following, ramping and operating reserves were evaluated. Power systems are dynamic, existing in a continuously changing environment, and are impacted by factors that change from moments-to-seconds, seconds-to-minutes, minutes-to-hours, seasonally and year-to-year. In the various time frames of operation, balance must be maintained between the load on the system and the available generation. In the very short timeframe (seconds-to-minutes),

bulk power system reliability is almost entirely maintained by automatic equipment and control systems such as automatic generation control (AGC). In the intermediate to longer timeframes system operators and operational planners are the primary keys to maintaining system reliability. Figure 5.2 displays the various timescales that impact power systems, the operating and planning processes they impact and the associated issues that need to be addressed.

NYISO Wind Generation Study l August 2010 10

Planning and Technology Operation Process Issues 1 Year Unit Dispatch Slower (Years) 700 Capacity Valuation 600 500 Resource and (UCAP, ICAP) 400 Capacity Planning and MW 300 Long-Term Load 200 (Reliability) Growth Forecasting 100 0

0 2000 4000 6000 8000 Hour 2001 Average Load vs Average Wind 1 Day 30,000 1,600 1,400 25,000 NYISO Load (MW) Wind Output (MW) 1,200 Unit Commitment Day-ahead and 20,000 1,000 and Multi-Day 15,000 800 Day-Ahead Forecasting 600 Scheduling 10,000 Time Frame 400 5,000 July load August load Septem ber load 200 July w ind August w ind Septem ber w ind 0 0 1 6 11 16 21 Hour 3000 2500 3 Hours Hour-Ahead Forecasting 2000 Load Following and MW 1500 (5 Minute Dispatch) Plant Active Power Maneuvering and 1000 Management 500 0

Faster (seconds) 1 61 121 M i nu te s Sept ember Morning A ugus t Morning May Ev ening Oc t ober Ev ening April Af ternoon Real-Time and 10 Minutes Frequency and Autonomous Protection Tie-Line Regulation and Control Functions (AGC) (AGC, LVRT, PSS, Governor, V-Reg, etc.)

Figure 5.2: Power System Time Scales The fact that the load is constantly changing means that its variability must first be understood in order to assess the impact of another variable element, (such as wind), on system operation. Statistics is an extremely useful tool for understanding and describing variation in data. The analysis of system variability for various time scales from minutes to hours is being conducted to assess the impact on such operating parameters as regulation, load NYISO Wind Generation Study l August 2010 11

following, operating reserves, ramping, and scheduling. Figure 5.2 presents the various time scales and the technology issues that are important in that time frame.

AWS Truepower developed wind profiles based on 2004 through 2006 wind data for approximately 35 sites in NY. Utilizing operating wind plants and proposed projects in the interconnection queue the NYISO then developed simulated outputs for wind plants ranging from an installed base of nameplate wind of 3,500 MW up to 8,000 MW of installed nameplate wind. The intermediate steps were nominally 4,250 MW and 6,000 MW. The wind plants from the NYISOs interconnection queue that are included in the study are listed in Table 5-1.

Table 5-1: List of Wind Plant Units Units that Compose the 1275 MW Case Queue Station/Unit Nameplate Rating (MW) Zone I/S Altona Windfield 99.0 D I/S Bliss Windfield 100.5 A I/S Canandaigua II 42.5 C I/S Canandaigua Wind Farm 82.5 C I/S Chateaugay Windpark 106.5 D I/S Clinton Windfield 100.5 D I/S Ellenburg Windfield 81.0 D I/S Fenner Wind Power 30.0 C I/S High Sheldon Windfarm 113.0 C I/S Madison Wind Power 11.6 E I/S Maple Ridge 1 231.0 E I/S Maple Ridge 2 90.7 E I/S Munnsville Wind Power 34.5 E I/S Steel Winds 20.0 A I/S Wethersfield 230kV 126.0 C I/S Wethersfield Wind Power 6.6 B Units Added to Create the 4250 MW Case Queue Station/Unit Nameplate Rating (MW) Zone 113 Prattsburgh Wind Park 55.5 C 119 Prattsburgh Wind Farm 79.5 C 152 Moresville Energy Center 129.0 E 155 Canisteo Hills Windfarm 148.5 C 156 Fairfield Wind Project 120.0 E 157 Orion Energy NY I 100 E 160 Jericho Rise Wind Farm 101.2 D 161 Marble River Wind Farm 88.2 D 166 St. Lawrence Wind Farm 130.0 E 168 Dairy Hills Wind Farm 120.0 C 169 Alabama Ledge Wind Farm 79.2 B 171 Marble River II Wind Farm 140.7 D 182 Howard Wind 62.5 C 186 Jordanville Wind 136.0 E 189 Clayton Wind 126.0 E 197 Tug Hill 78.0 E 198 New Grange Wind Farm 79.9 A NYISO Wind Generation Study l August 2010 12

203 GenWy Wind Farm 478.5 A 207 Cape Vincent 210.0 E 220 Armenia Mountain I 175.0 C 221 Armenia Mountain II 75.0 C 222 Ball Hill Windpark 99 A 234 Steel Winds II 60 A 237 Allegany Windfield 79 A Units Added to Create the 6000 MW Case Queue Station/Unit Nameplate Rating (MW) Zone 150 Cherry Valley Wind Power 70 F 178 Allegany Wind 79.0 A 179 Cherry Hill Windpark 102 D 187 North Slope Wind 109.5 D 215 Noble Burke Windpower 120 D 217 Cherry Flats 90 C 227 Orleans Wind 120 B 236 Dean Wind 150 C 238 Tonawanda Creek Wind 75 B 239 Western Door Wind 100 C 240 Farmersville Windpark 100 A 246 Dutch Gap Wind 250 E 254 Ripley-Westfield Wind 124.8 A 256 Niagara Shore Wind 70.5 A 263 Stony Creek Wind Farm 142.5 C 241 Chateaugay II Windpark 19.5 D Units Added to Create the 8000 MW Case Queue # Station/Unit Nameplate Rating (MW) Zone 270 Hounsfield Wind 268.8 C 282 Concord Wind 101.2 A 285 Machias I 79.2 A 297 Ashford Wind 19.9 A 298 Leicester Wind 57 B 301 Hamlin Wind Farm 80 B 327 Offshore Wind 1400 J, K Summary of Nameplate Rating by Case for each Zone (MW)

Case A B C D E F J, K Total 1275 121 7 394 387 368 1276 4250 917 86 1110 717 1397 4227 6000 1291 281 1593 1068 1647 70 5949 8000 1492 418 1861 1068 1647 70 1400 7955 The simulations were done based on 2005 and 2006 wind data. The AWS site closest to the existing wind or proposed wind plant site was utilized for developing a specific output profile for that wind plant. Output profiles based on 2005 and 2006 wind data were developed for each wind plant. The first 1,500 MW of wind was simulated with wind turbines with a hub height of 80 meters and balance with a hub height of 100 meters.

Simulated wind plant output was developed for one minute, ten minute and one hour for selected sites in NY.

Load profiles were developed internally.

NYISO Wind Generation Study l August 2010 13

Figure 5.3 shows the hourly simulations for 8,000 MW of New York wind plants based on 2006 wind data. Note the variability of the aggregate wind plant output which swings between 90% of nameplate and close to zero.

The figure also contains the thirty day, or 720 hour0.00833 days <br />0.2 hours <br />0.00119 weeks <br />2.7396e-4 months <br />, moving average which shows the seasonality of wind with the highest energy production during the winter capability period and the lowest being during the summer capability period (hours 2880 - 7269).

8000 7000 6000 5000 MW_8000 MW 4000 Trend_8000 3000 2000 1000 0

0 730 1460 2190 2920 3650 4380 5110 5840 6570 7300 8030 8760 Hour of the Year Figure 5.3: Hourly Wind Output for 8,000 MW of Wind The AWS data is available from NREL at http://wind.nrel.gov/public/EWITS/. Load models for each time frame were also developed.

5.4.2. The Critical Importance of Net Load Net load is defined as the aggregate customer load demand minus the aggregate variable generation output.

Why is net load important? It is important because variable generation has more in common with system electrical demand (load) than conventional generation resources, as both are:

Cyclic on an annual (seasonal) basis, and a diurnal (daily) basis Subject to random short-term variations around the multi-hour trends Limited controllability (i.e., limited dispatchablility)

Subject to deviations from predicted day-ahead behavior Mutually dependent on prevailing weather conditions As a result, determining the impacts of variable generation on bulk power system operations and planning cannot be evaluated by examining wind generation output characteristics, such as its variability and predictability, independently from the simultaneous behavior of the load. Thus, analysis of wind variation NYISO Wind Generation Study l August 2010 14

independent of load variation is inadequate and inappropriate to determine impacts of variable generation on the need for flexibility. Operationally, the dispatchable generation output must conform to the characteristics of the net load.

How does variable generation interact with load to affect the variability of net load? The inherent variability and imperfect predictability of variable generation adds to the variability and prediction errors of system load.

Experience has shown that some of the variation in load and wind output cancel each other in a combined series. In other words, given synchronized load and wind generation time series, the net variability of load-wind over a time period is less than the sum of the variability of the individual series over the same time period. In addition, the variabilities cannot simply be combined as if they are independently random, as they are both affected by the common factor of the weather. Nor can they be added algebraically because the correlation is only partial and the coefficient can be either positive or negative, or vary in sign with time or location of the wind resource.

The result is that the net-load is considerably more variable than the load by itself which increases as the amount of variable generation increases. It is the net-load that conventional generation will have to respond to. This will result in a need for greater flexibility from the conventional supply resources. This will translate into a greater need for regulation, ramping, and load following capability in real time operations. These requirements will need to be accounted for in the planning timeframe as well.

5.4.3. Net Load Variability Characterization Net load (Load minus Wind) is the amount of generation required from dispatchable units. This section focuses on the variability of net-load rather than wind generation in isolation because experience has shown that some of the variation in load and wind output cancel each other in a combined series. In other words, given synchronized load and wind generation time series, the net variability of load-wind over a time period is less than the sum of the variability of the individual series over the same time period.

5.4.4. Measuring Variability The variability of net load in different timeframes impacts various aspects of bulk power system operation.

Implications for regulation requirements, ramp and range considerations, and operating reserves issues can be drawn from an analysis of net load variability in the 1-, 5-, 15-, 30-, and 60-minute timeframes, depending on the ancillary service definitions and market rules. This section will focus on the statistical analysis of load and net load variability in the various timeframes. In this section several terms are used to characterize the load and net load variability. They include:

Delta () - The incremental change in a variable such as a period-to-period ramp rate Sigma () - The standard deviation of a dataset which is a measure of how dispersed observations are, relative to the mean ()

Since deltas can be positive or negative depending on the slope of the series at a point in time, the average of the deltas is somewhat meaningless. In fact, for a series of a day or longer, the mean of the deltas is zero or near zero. The standard deviation of the deltas, however, is a good indication of how much the series changes from period-to-period; therefore sigma of the deltas is used as a measure of variability in this study. If the deltas are normally distributed (a rational assumption based on experience) then sigma relates to the proportion of deltas within a certain distance of the mean  as shown in Figure 5.4.

NYISO Wind Generation Study l August 2010 15

Figure 5.4: Normal Distribution The sigma () of the deltas () of the net load for the various time domains and can be summarized in many different ways. For instance, the sigma can be calculated hourly, in groups of hours, day of week, monthly, annually, etc. For the NYCA analysis, the  for the various time domains was calculated by hour, groups of hours, all-days, weekdays, Saturday, Sunday, monthly, and annually. The annual numbers are useful in showing macro trends while the monthly numbers are useful in focusing the analysis. Hourly numbers and groupings are useful for assessing impacts on system operations.

The key driver for net load variability is the time domain relationship between load and wind plant output. Wind can amplify operationally challenging periods. For instance, wind is generally dropping off during the morning load rise which amplifies the morning ramp up requirements. Likewise wind is usually ramping up when the load is dropping off which amplifies the down ramp. How wind interacts with load is the important factor in considering the impact of wind generation on the need for ancillary services, especially regulation, which is discussed later in this report.

The following figures abstracted from the wind simulations for New York clearly demonstrate the time domain interaction of wind and load as described above. Figure 5.5 and 5.6 is a simulation for 2013 for the July system peak load day with 6,000 MW of installed nameplate wind. Figure 5.5 includes the wind as well as the load and net load while Figure 5.6 presents the load and net load without the wind.

NYISO Wind Generation Study l August 2010 16

40000 5000 35000 4500 4000 30000 3500 25000 3000 Load MW 20000 2500 Net Load 15000 2000 1500 Wind Generation 10000 1000 5000 500 0 0 0:00 2:00 4:00 6:00 8:00 10 12 14 16

0
0
0
0 0

0 0

18 20:0

00 0

0 22:00 Time of Day Figure 5.5: Load, Net Load and Wind for the Peak Day of July 2013 with 6,000 MW of Wind 36000 34000 32000 30000 28000 MW 26000 Load 24000 Net Load 22000 20000 18000 16000 0:00 2:00 4:00 6:00 8:00 10 12 14 16

0
0
0
0 0

0 0

18 20:0

00 0

22:00 0

Time of Day Figure 5.6: Load and Net Load for the Peak Day of July 2013 with 6,000 MW of Wind A number of observations can be extracted from these figures. The first is the confirmation of the above discussion which is wind is generally dropping off when load is ramping up and when the load is ramping down the wind is ramping up. The result is the morning load ramp starting earlier and being steeper, and the evening load drop starting a little later and being steeper. Another observation that is not as apparent with this graphic but which is demonstrated with the next graphic, is that change from the MW difference between the night-time net load minimum to the day-time net daily peak load are generally much greater. Finally, this graphic does show a phenomenon that wasnt present in the initial study and is somewhat unique to New York which is the afternoon or secondary peak in wind plant output. The result is that NYs wind plant output has demonstrated higher coincidence with peak loads than the first study found, especially for the peak summer months.

NYISO Wind Generation Study l August 2010 17

The final observation that can be made is that the change in the night time net-load minimum will be lower than load by itself as well as an increase in the MW delta from the daily net-load minimum to the daily net-load peak.

Table 5-2 below presents how the various wind scenarios will impact the net minimum load with no wind curtailment while Table 5-3 presents the minimum to peak maximum increases for summer and winter in 2018.

Also, Figure 5.7 and Figure 5.8 present load and net load for the simulated 2018 summer and winter peak weeks to demonstrate visually the resultant net load which the dispatchable generation would have to follow. The Tables and Exhibit provide simulation results that show that the nighttime net-load minimums will be much lower than load by itself and the amount of movement in dispatchable generation (daily ramp up) that will be required to follow the load from the night-time low to the daily peak will also increase significantly.

Table 5-2: Simulated Minimum Loads and Minimum Net-Loads Load No Wind Low Wind Scenario1 High Wind Scenario2 Study Year (MW) (MW) (MW) 2008 10,790 2011 12,618 10,297 9,692 2013 12,937 10,023 8,560 2018 13,721 9,398 7,574

1) 3,500 MW in 2011, 4,250 MW in 2013 and 6,000 MW in 2018
2) 4,250 MW in 2011, 6,000 MW in 2013 and 8,000 MW in 2018 Table 5-3: Trough to Peak Maximum Increases for Summer and Winter 2018 Load Metric\Season Summer 2018 Winter 2018 Date MW Date MW Peak Load 17-Jul 37102 8-Dec 28231 Max Load Change Trough-Peak 17-Jul 15627 18-Dec 11389 Max Net Load Change Trough-Peak 18-Jul 17464 11-Dec 14734 40000 35000 30000 Load MW Load Net_Load 25000 20000 15000 8 8 8 8 8 8 8 L2 L2 L2 L2 L2 L2 L2 01 01 01 01 01 01 01 14 15 16 17 18 19 20 JU JU JU JU JU JU JU Date NYISO Wind Generation Study l August 2010 18

Figure 5.7: Load and Net Load for the Simulated 2018 Summer Peak Week 30000 28000 26000 24000 22000 Load MW Load 20000 Net_Load 18000 16000 14000 12000 10000 8 8 8 8 8 8 8 20 20 20 20 20 20 20 1 1 1 1 1 1 1 EC EC EC EC EC EC EC 05 06 07 08 09 10 11 D D D D D D D Date Figure 5.8: Load and Net Load for the Simulated 2018 Winter Peak Week 5.4.5. Impact of Wind on System Net-Load Variability Simulations of wind plant output for total installed nameplate wind plant range between 3,500 MW and 8,000 MW for the years 2011, 2013, and 2018. Wind plant output simulations were conducted based on two different weather and load shape years which were 2005 and 2006 to determine the impact of variable wind generation on the net-load. The input and output data created for this analysis totaled in excess of 2 gigabytes. This presented a significant challenge in terms of presenting the results without overwhelming the reader with a significant volume of data but yet presenting the salient points. As presented above, the timescales for the all power system operational processes impacted by wind plants range from less than seconds to minutes, days, and longer.

How does variable generation interact with load to affect the variability of net load? Variations in load and wind output can cancel each other in a combined series. In other words, given synchronized load and wind generation time series, the net variability of load plus wind over a time period is less than the sum of the variability of the individual series over the same time period. In addition, the variability of each cannot simply be combined as if they are independently random, as they are both affected by the common factor of the weather.

The overall outcome is that system variability as measured by the sigma of the net-load deltas increases in all time frames. Figure 5.9 below displays the deltas of the load and net-load for 60 minutes. This is for 8,000 MW of wind and the 2018 load forecast used in this study. The result is generally what is observed in all timeframes. It is also the result that has been in observed in other studies of wind integration. The net-load which is the green or darker bars in the figure has higher variability than the load which has a distribution which is more peaked and less dispersed.

NYISO Wind Generation Study l August 2010 19

1800 1700 1600 1500 1400 1300 Frequency of Load Deltas 1200 1100 1000 900 Load Alone 800 Load - 8000 MW 700 Wind 600 500 400 300 200 100 0

-2

-2 70 < -

0 40 -2 0 0 - 27 40 0 0

10 -2 0

80 -1 0

50 -1 10 80 0 0

0

- 0 20 12 0

00 90 5000 00 60-

-3 300 - 0 0

0 30 30 0 00

- - 0 60 6 90 -

12 - 0 00 0 90 - 0 0 00 12 15 -

00 15 18 -

00 18 0 000 21 - 2 0 0 0 10 24 -

00 24 0 00

- >2727 0 0000 60 Min. Load Delta (in MW) 2006 Wind Figure 5.9: Distribution of 60-minute Deltas for Load and Net Load with 8 GW of Wind As can be seen in Figure 5.9, the frequency of occurrence of the net-load deltas when compared to the load deltas decreases for deltas around zero where wind and load cancel each other. The frequency of higher magnitude net-load delta increases when compared to the load deltas for deltas nearer the extremes of the distribution where wind and load are additive. The result is that the net load is considerably more variable than the load by itself and increases as the amount of variable generation increases.

Given that the variable and uncertain nature of wind plant output results in net-loads that are more variable than the load by itself, the next factor to explore is how does increasing wind plant penetration impact overall net-load variability as measured by the net load deltas? Does it increase linearly with increasing wind plant penetration or exponentially?

To assess this issue a summary of annual sigma of load (without wind) and net-load for various penetrations of wind plants is presented for the 1-minute, 5-minute and 60-minute timeframes. The annual sigma provides a macro overview of how this statistical parameter changes with increasing load and wind penetration. These timeframes are presented because they incorporate the timeframes important to operational processes such as the automatic generation control (AGC), the five minute dispatch cycle and longer term ramping requirements. In the five minute timeframe, the operational burden imposed by wind will certainly translate into more ramp and range requirements as well as potentially increasing regulation requirement. The one-hour timeframe gives a good indication of the longer term ramping requirements that will be required with wind because random variations have less impact in the longer timeframes. Table 5-4 presents the simulated results for how the annual sigma changes with increases in wind penetration and load growth.

Table 5-4: Annual Load/Net-Load   by Timeframes, Load Levels and Wind Penetration NYISO Wind Generation Study l August 2010 20

1-min.  Percent 5-min.  Percent 60-min.  Percent Case Sigma Increase With Sigma Increase With Sigma Increase MW Wind MW Wind MW With Wind Load Alone 2011 36.6 85.4 895.9 Net Load 3500 MW of Wind 37.6 2.8% 89.1 4.4% 916.2 2.3%

Net Load 4250 MW of Wind 37.9 3.5% 90.3 5.7% 924.2 3.2%

Load Alone 2013 37.5 87.5 918.5 Net Load 4250 MW of Wind 38.8 3.4% 92.3 5.5% 946.4 3.0%

Net Load 6000 MW of Wind 39.6 5.6% 95.9 9.6% 967.9 5.4%

Load Alone 2018 39.8 92.8 973.8 Net Load 6000 MW of Wind 41.8 5.0% 100.8 8.6% 1021.5 4.9%

Net Load 8000 MW of Wind 42.8 7.5% 104.8 12.9% 1039.6 6.8%

As expected, overall annual net-load sigma/variability increases as wind generation penetration increases. The increase appears to be linear with a very gradual slope for the penetrations studied which ranged from 10% of peak load up to 21.5% of peak load. Figure 5.10 and 5.11 are plots of the sigmas for the various timescales vs.

MWs of installed wind on a semilog scale and a plot of normalized sigma (MW/min) vs. installed MWs of wind.

1000.0 1-min 

 (MW) 100.0 5-min 

60-min 

10.0 3,000 4,000 5,000 6,000 7,000 8,000 Installed Wind Nam eplate MW Figure 5.10: Net-load  (adjusted for load growth) VS Installed Wind in MWs 45.0 40.0 35.0

 in MW/min.

1-min 

30.0 5-min 

25.0 60-min 

20.0 15.0 10.0 3,000 4,000 5,000 6,000 7,000 8,000 Installed Wind Nameplate MW NYISO Wind Generation Study l August 2010 21

Figure 5.11: Normalized  VS Installed Wind MWs The plots confirm that the increase in annual sigma () of the net load  with the load growth component removed increases linearly with a slight positive slope as the amount of installed wind increases. This result has been observed in other studies, although the slope observed in other studies has been steeper than observed in New York.

The magnitude of the  for the net-load  changes by season, day of week, and hour of the day. Figure 5.12 below presents a plot of the monthly  for the 10-min.  for 6,000 and 8,000 MWs of installed wind based on the 2018 peak load and 2006 weather data.

250 200 150 No Wind

 in MW 6000 MW Wind 8000 MW Wind 100 50 0

1 2 3 4 5 6 7 8 9 10 11 12 Month Figure 5.12: Monthly  of the 10-min. net-load  for 2018 Based on 2006 Wind Data Figure 5.13 shows that the net-load  is highest during months of highest loads and lowest during period of minimal loads. A plot of sigma by hour of the day shows that sigma will also vary by time of day and is generally highest during the morning ramp up and the evening ramp down. Figure 5.14 is a plot of the 5 minute sigma by month and hour of the day for hour beginning 0500, 1300, and 1900 for 2008 actual and 8,000 MW of simulated wind for 2018.

NYISO Wind Generation Study l August 2010 22

90.0 80.0 70.0 60.0

 (MW/5min)

HB 0500 50.0 HB 1300 40.0 HB 1900 30.0 20.0 10.0 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month Figure 5.13: Sigma by the Hour of the Day for 5-min. Net-Load  for 2008 140.0 120.0 100.0

 (MW/5min)

HB 0500 80.0 HB 1300 60.0 HB 1900 40.0 20.0 0.0 1 2 3 4 5 6 7 8 9 10 11 12 Month Figure 5.14: Sigma by the Hour of the Day for 5-min. Net-Load  for 2018 The above graphics have shown that net-load variability increases with increasing wind penetration. The net-load is important because the net-load is what the conventional generation will need to follow. This will result in the need for increased system flexibility. Flexibility can manifest itself in terms of a need for increased regulation requirements in the minute-to-minute timeframe, as well as increased frequency of larger magnitude ramps that occur in the five to ten minute, and hour or longer timeframes. The next three sections discuss the impact of the net load variability on the system regulation, hourly ramping events, and operating reserves.

NYISO Wind Generation Study l August 2010 23

5.4.6. Impact of Increasing Wind Penetration on System Regulation Regulation requirements are established to address the variability of load and wind (net-load variability) that may occur within a 5-minute dispatch interval. This section of the report will outline the methodology used to establish regulation requirements at the specified wind penetration and forecasted load levels included in the study.

Regulation Study Approach In order to evaluate the going-forward regulation requirements, the wind and load data as explained in prior sections of this report is leveraged. Actual 2005 and 2006 meteorological data (e.g., wind speed and direction) is used to simulate NYCA wind generation in 5-minute intervals at the specified wind penetration levels of 3,500MW, 4,250MW, 6,,000 MW, and 8,,000 MW as shown in the Table 5-5 below. In addition, actual 2005 and 2006 load shape data is used to project NYCA 5-minute load for the study years of 2011, 2013 and 2018. The 2005 data forms the basis of the regulation requirements analysis with 2006 data used for validation purposes.

NYISO Wind Generation Study l August 2010 24

Table 5-5: Projected Peak Loads and Wind Plant Penetration Levels Year Projected Peak Projected Wind Penetration (MW)

Load (MW) Level 1 Level 2 2011 34,768 3,500 4,250 2013 35,475 4,250 6,000 2018 37,130 6,000 8,000 The coincident wind and load data is evaluated to determine the net-load on an interval-by-interval basis, as well as the deltas, or differences, between successive intervals. By looking at the net-load variability, situations in which load and wind move in the same direction (resulting in a lesser net change) and situations in which load and wind move in opposite directions (resulting in a greater net change) are considered.

In order to establish the data set of net-load differences between successive intervals, 5-minutes of load and 10-minutes of wind output deltas are considered as shown in the equation below.

Delta Net-Load = Load (t-(t-1)) - Wind (t-(t-2))

where t, t-1 and t-2 represent 5-minute intervals As shown in the Diagram 5-1 below, this equation would for example, take the difference in the load between the 5-minute interval of 6:50 and 6:55 and couple that with the difference in the wind output between the 10-minute interval of 6:45 and 6:55. Given that the NYISO has a 5-minute dispatch and that the NYISO uses a load forecast to project the load in the binding interval of the dispatch, it is appropriate to evaluate the 5-minute load deltas. The NYISO uses a persistence assumption for wind for the next 5-minute binding dispatch interval (leveraging the wind forecast for the further out advisory intervals). Persistence is the most accurate assumption in the near term and it simply applies the current actual wind output to the projected output. As a result, the system is exposed to 10-minutes of wind variability through the dispatch process which is why the regulation analysis takes into account 10-minutes of wind variability.

5-Minute 10-Minute Load Wind Deltas Deltas t = 6:55 t-1 =

6:50 t-2 =

6:45 Diagram 5-1 The standard deviation of the resulting net-load data set is then determined to measure the variability. For each hour of each month, the net-load delta variability corresponding to a 3 sigma level (to incorporate 99.7% of the NYISO Wind Generation Study l August 2010 25

sample set) is calculated. The resulting 3 sigma value represents the amount of regulation resources required to cover system variability.

Study Considerations In order to confirm the validity of the variability data the results were reviewed against historical performance and Operations' knowledge of system patterns. A few adjustments were made to the raw data in order to make the integration with the markets and daily operations more practical and seamless.

Seasonal Definitions - The regulation requirements have historically been broken into four seasonal groupings:

April - May, June - August, September - October, and November - March, as shown in the Table 5-6 below. A comparison of the historical seasonal breakdown with the study results showed similar net-load variability patterns and support retaining the same seasonal divisions.

Table 5-6: Regulation Requirements: Historical (Pre-Study) Sunday & Weekday Requirements April - May June - August Sept - Oct Nov - March Current Current Current Current Current Current Current Current Sunday Weekday Sunday Weekday Sunday Weekday Sunday Weekday HB Req. Req. Req. Req. Req. Req. Req. Req.

0 150 150 175 175 160 180 160 190 1 150 150 175 175 160 180 160 190 2 150 150 175 175 160 180 160 190 3 150 150 175 175 160 180 160 190 4 150 150 175 175 160 180 160 190 5 150 175 175 200 160 250 160 250 6 150 275 175 275 160 275 160 275 7 150 275 175 275 160 275 160 275 8 150 275 175 275 160 275 160 275 9 160 200 160 250 180 260 180 250 10 175 175 175 240 210 250 210 250 11 150 150 175 210 160 210 160 210 12 150 150 175 175 160 180 160 180 13 150 150 175 175 160 180 160 180 14 150 150 175 175 160 180 160 180 15 150 175 175 175 160 190 160 190 16 175 200 230 250 230 250 230 275 17 200 200 250 250 250 250 250 275 18 200 200 250 250 250 250 250 275 19 200 200 250 250 250 250 250 250 20 200 200 250 250 250 250 250 250 21 200 200 250 250 250 250 250 250 22 175 175 225 225 225 240 225 240 23 150 150 175 175 175 190 175 190 Control Performance Requirements - Regulation is required to balance resources and demand, thereby maintaining a satisfactory Interconnection frequency. For certain hours, the raw data study results show that a reduction in the regulation requirement as compared to historical values is possible. For these hours a validation of the new value as compared to the historical (2008/2009) Control Performance was performed. If the historical performance fell below a threshold level of 94% for the hour (NERC CPS2 requires a monthly average Area Control Error of at least 90%), the historical regulation value is maintained.

NYISO Wind Generation Study l August 2010 26

Study Wind Data Basis Year - As explained previously in the report, actual load and wind data from the years of 2005 and 2006 form the basis of the study simulations to project the higher wind penetration values and load levels. In the analysis of the regulation requirements, the base year of 2005 serves as the primary data source because there is some additional volatility in the net-load as compared to the 2006 data. The 2006 net-load variability data was also considered to compare overall results and to adjust particular hours if they appear inconsistent within a day.

Day of Week Requirements - The study data does not support having a unique Sunday requirement; a conclusion supported by a historical control performance review. Therefore, the new requirements are based on common hourly values for all days of the week.

Hourly Increments - The regulation values are set to 25MW increments.

Hourly Ramp - The hour to hour ramping of the regulation requirements is limited to 50MW in order to minimize unnecessary real-time energy price volatility.

Regulation Study Results The final regulation requirements determined in the study are presented along with the historical weekday values (labeled "current") for ease of comparison. Each table represents a single load level (for example study year 2011 with a 34,768MW peak load) displayed with two wind penetration levels (for example study year 2011 includes a wind level of 3500MW and 4250MW).

The study results show that with a 3500MW wind level and 2011 load (34,768MW peak) the regulation requirements increase by 5MW based on a weighted average. The maximum increase is 100MW (a change from a 175MW requirement up to 275MW) for the June-August season HB23. The highest requirement is 300MW in the November-March season HB17.

For the highest wind penetration level of 8,000 MW coupled with a 2018 load (37,130MW peak), the regulation requirements increase by 116MW based on a weighted average. The maximum increase is 225MW (a change from a 175MW requirement to 400MW) for the June-August season HB14. The highest requirement is 425MW in the June-August season HB20/HB21.

The results Tables 5-7 through 5-10 included show the hourly values for each study condition and the shading included helps identify increases and decreases in the requirement as well as modifications due to some of the study considerations previously described.

Key: Results Tables 5-7 through 5-10 NYISO Wind Generation Study l August 2010 27

Increase in Requirement Decrease in Requirement Requirement modified to eliminate > 50MW delta in consecutive hours Underlined Values Requirement modified due to historical CPS 2 performance Table 5-7: Legend Table 5-8: Regulation Requirements: Study Year 2011 (34,768MW Peak Load) 2011 April - May June - August Sept - Oct Nov - March Current Current Current Current Weekday Wind Wind Weekday Wind Wind Weekday Wind Wind Weekday Wind Wind Regulation Level Level Regulation Level Level Regulation Level Level Regulation Level Level HB Requirement 3500MW 4250MW Requirement 3500MW 4250MW Requirement 3500MW 4250MW Requirement 3500MW 4250MW 0 150 175 175 175 225 225 180 175 200 190 200 200 1 150 175 175 175 175 200 180 175 175 190 175 200 2 150 175 175 175 175 175 180 150 175 190 175 175 3 150 175 200 175 175 200 180 175 200 190 150 175 4 150 225 225 175 225 225 180 225 250 190 175 175 5 175 225 225 200 250 275 250 275 300 250 225 225 6 275 225 225 275 275 275 275 275 275 275 275 275 7 275 200 225 275 275 275 275 250 275 275 275 275 8 275 200 200 275 275 275 275 225 225 275 275 275 9 200 175 175 250 225 225 260 200 225 250 225 225 10 175 200 200 240 225 225 250 175 200 250 175 200 11 150 200 225 210 250 275 210 200 200 210 175 200 12 150 175 175 175 225 250 180 200 225 180 175 200 13 150 175 175 175 225 250 180 200 225 180 175 175 14 150 175 175 175 250 275 180 175 200 180 175 175 15 175 175 200 175 225 225 190 175 200 190 225 225 16 200 175 200 250 250 250 250 200 225 275 275 275 17 200 200 225 250 250 250 250 250 275 275 300 300 18 200 225 225 250 250 250 250 275 300 275 250 250 19 200 250 275 250 250 250 250 250 275 250 250 250 20 200 200 225 250 250 250 250 250 250 250 200 225 21 200 200 225 250 250 275 250 250 250 250 225 225 22 175 200 200 225 275 275 240 200 200 240 200 200 23 150 200 200 175 275 275 190 225 250 190 200 200 NYISO Wind Generation Study l August 2010 28

Table 5-9: Regulation Requirements: Study Year 2013 (35,475MW Peak Load) 2013 April - May June - August Sept - Oct Nov - March Current Current Current Current Weekday Wind Wind Weekday Wind Wind Weekday Wind Wind Weekday Wind Wind Regulation Level Level Regulation Level Level Regulation Level Level Regulation Level Level HB Requirement 4250MW 6000MW Requirement 4250MW 6000MW Requirement 4250MW 6000MW Requirement 4250MW 6000MW 0 150 175 200 175 225 275 180 200 225 190 200 225 1 150 175 225 175 200 250 180 200 225 190 200 250 2 150 175 200 175 175 225 180 175 200 190 175 225 3 150 225 250 175 200 225 180 225 250 190 175 200 4 150 275 300 175 250 275 180 275 300 190 225 250 5 175 300 325 200 275 300 250 325 350 250 275 300 6 275 250 275 275 300 325 275 275 300 275 325 350 7 275 250 250 275 275 275 275 275 300 275 275 300 8 275 200 250 275 275 275 275 225 275 275 275 275 9 200 225 250 250 225 275 260 225 250 250 225 275 10 175 225 250 240 225 275 250 175 225 250 200 275 11 150 200 225 210 275 275 210 200 250 210 225 300 12 150 200 250 175 250 300 180 200 250 180 200 250 13 150 225 275 175 225 275 180 200 250 180 200 225 14 150 200 250 175 275 325 180 200 225 180 175 200 15 175 225 275 175 250 300 190 200 225 190 225 250 16 200 175 225 250 250 325 250 225 250 275 275 300 17 200 200 250 250 250 325 250 275 300 275 300 325 18 200 225 275 250 250 275 250 275 325 275 250 275 19 200 275 325 250 250 300 250 275 325 250 250 325 20 200 225 275 250 275 325 250 250 300 250 225 275 21 200 225 275 250 250 325 250 250 275 250 225 275 22 175 225 250 225 275 325 240 225 250 240 250 300 23 150 225 250 175 275 325 190 250 275 190 200 250 Table 5-10: Regulation Requirements: Study Year 2018 (37,130MW Peak Load) 2018 April - May June - August Sept - Oct Nov - March Current Current Current Current Weekday Wind Wind Weekday Wind Wind Weekday Wind Wind Weekday Wind Wind Regulation Level Level Regulation Level Level Regulation Level Level Regulation Level Level HB Requirement 6000MW 8000MW Requirement 6000MW 8000MW Requirement 6000MW 8000MW Requirement 6000MW 8000MW 0 150 225 250 175 250 300 180 225 275 190 250 275 1 150 225 275 175 250 325 180 250 300 190 250 300 2 150 225 275 175 225 275 180 225 250 190 225 300 3 150 275 300 175 250 275 180 275 300 190 225 250 4 150 325 350 175 300 325 180 325 350 190 250 275 5 175 325 350 200 300 325 250 375 400 250 300 325 6 275 275 300 275 350 375 275 325 350 275 350 375 7 275 300 300 275 300 375 275 325 375 275 300 375 8 275 275 300 275 275 325 275 300 350 275 275 350 9 200 250 275 250 275 325 260 250 300 250 275 325 10 175 275 300 240 250 300 250 225 275 250 250 300 11 150 225 275 210 275 325 210 275 300 210 300 300 12 150 250 325 175 275 375 180 250 300 180 250 300 13 150 275 350 175 275 350 180 250 300 180 225 275 14 150 225 300 175 325 400 180 225 275 180 225 275 15 175 250 325 175 300 350 190 250 300 190 250 325 16 200 225 275 250 325 400 250 275 300 275 300 350 17 200 250 300 250 325 400 250 325 350 275 350 400 18 200 275 325 250 275 350 250 300 325 275 300 350 19 200 325 375 250 300 375 250 325 375 250 325 400 20 200 275 325 250 350 425 250 300 375 250 300 375 21 200 275 325 250 350 425 250 300 375 250 275 325 22 175 250 275 225 350 400 240 250 325 240 275 350 23 150 250 300 175 300 350 190 275 325 190 250 325 NYISO Wind Generation Study l August 2010 29

Figures 5-15 through 5-17 display the historical values (labeled "current") requirements along with the requirements as determined in the study. As the load and wind penetration levels increase, the regulation requirement on average also increases.

450 Regulation Requirement (MW) .

400 350 300 250 200 150 100 HB HB HB HB HB HB HB HB HB 5

10 15 20 1

6 11 16 0 HB HB HB HB HB HB HB HB HB HB HB 21 2

7 12 17 22 3

8 13 18 23 April - May June - August September - October November - March Current Req. Req. at 3500MW of Wind Req. at 4250MW of Wind Figure 5.15: Current and Proposed Regulation Requirements 3500/4200 MW of Wind 2011 - 34,768 Peak Load 450 Regulation Requirement (MW) .

400 350 300 250 200 150 100 H

HH HB0 B5 B1 B1 HB H 0 5

20 B1 H

H HHB6 B1 B1 B2 H

HB2 B71 6

1 H

H H

HB1 B1 B2 H

HB3 B8 B1 2 7

2 H

HB1 B2 3 8

3 April - May June - August September - October November - March Current Req. Req. at 4250MW of Wind Req. at 6000MW of Wind Figure 5.16: Current and Proposed Regulation Requirements 4250/6000 MW of Wind 2013 - 35,475 Peak Load NYISO Wind Generation Study l August 2010 30

450 Regulation Requirement (MW) .

400 350 300 250 200 150 100 H

H HH HB0 B5 B1 B1 B2 HB10 5

0 H

H HHB6 B1 B1 B2 H

HB2 B71 6

1 H

H H

HB1 B1 B2 H

HB3 B8 B1 2 7

2 H

HB1 B2 3 8

3 April - May June - August September - October November - March Current Req. Req. at 6000MW of Wind Req. at 8000MW of Wind Figure 5.17: Current and Proposed Regulation Requirements 6000/8000 MW of Wind 2018 - 37,130 Peak Load 5.4.7. Impact of Increasing Wind Penetration on Load Following and Ramping Introduction To evaluate how increasing wind penetration would impact load following and ramping events, both the simulated net load data and simulated dispatch data from GridView were analyzed. In conducting the evaluations, the simulated data were analyzed to determine how the load and wind interact to impact the level of ramping that dispatchable generation needs to follow and how the magnitude of the load delta data compares to the net load delta for five minute, 1-hour, and 4-hour time frames. The five minute timeframe is indicative of the magnitude of the changes that will occur during the economic dispatch cycle, while one to four hours would be indicative of what would occur in the morning up and evening down ramps. Finally, simulated dispatch data generated by the GridView production cost model were analyzed to determine how the increased frequency of higher magnitude ramp changes would impact the dispatch of thermal plants.

Figure 5.18 presents a plot of the hourly loads, wind generation, and resulting net hourly ramps for the week of peak wind generation (week beginning the second Tuesday of February at 0000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />) based on 2018 loads and 8,000 MW of installed wind generation. Figure 5.19 presents a plot of the hourly loads and net-load ramps for that week.

NYISO Wind Generation Study l August 2010 31

20000 30000 25000 15000 Wind Gen. (MW) & Net Ramp (MW/HR) 20000 10000 Load (MW) 15000 5000 10000 0

5000

-5000 0 Net Ramp Wind Gen Load Figure 5.18: 2018 Hourly Loads, Wind Generation and Ramps for the Week of Peak Wind Generation Figure 5.19 presents the hourly ramp resulting from the load and the net-load ramp for the same 2800 2200 1600 1000 Ramp (MW/HR) 400 week. -200

-800

-1400

-2000

-2600 Net-Load Ramp Load Ramp Figure 5.19: 2018 Hourly Net-Load Ramps VS Load Ramps for the Week of Peak Wind Generation NYISO Wind Generation Study l August 2010 32

Figures 5.20 and 5.21 present the same results for the peak load week for 2018.

20000 40000 35000 15000 Wind Gen (MW) & Net Ramp (MW/HR) 30000 25000 10000 Load (MW) 20000 5000 15000 10000 0

5000

-5000 0 Net Ramp Wind Gen Load Figure 5.20: 2018 Hourly Loads, Wind Generation and Ramps for the Week of the Peak Load 4000 3500 3000 2500 2000 1500 Ramp (MW/HR) 1000 500 0

-500

-1000

-1500

-2000

-2500

-3000

-3500 Net-Load Ramp Load Ramp Figure 5.21: 2018 Hourly Net-Load Ramps VS Load Ramps for the Week of the Peak Load NYISO Wind Generation Study l August 2010 33

These graphics demonstrate how the variable nature of wind generation can increase the ramps that the dispatchable generation needs to respond to. When the net-load (green or dashed line) exceeds the load ramp (red line) the wind is increasing the hourly ramp. Likewise, when the net-load is less than the load ramp wind is reducing the ramp that the dispatchable generation needs to follow. Overall, the variable nature of generation tends to increase the range and maximum magnitude of the ramps to which dispatchable generation needs to respond.

Range and Magnitude of Net-Load Events Figures 5.22 through 5.28 present annual duration curves from highest to lowest for up and down ramps for 5 minutes (every 12th point plotted), 1-hour, and 4-hour ramps/load following events. These graphics present the full range of up and down ramping events as simulated for a full year. The 5-minute data is indicative of what the range and magnitude of ramping events for the economic dispatch cycle would be, while the 1-hour and 4-hour ramp durations are indicative of what would be encountered in the morning ramp up and the evening ramp down.

This data is presented based on 8 GW of installed wind. Figure 5.22 presents the 5-minute up and down ramps for 8760 hours0.101 days <br />2.433 hours <br />0.0145 weeks <br />0.00333 months <br />. Figures 5.23 and Figures 5.24 present the 50 highest hours for 5-minute up and down ramps for greater fidelity of maximum ramp events. The Figures 5.25 through 5.27 presents the same data for one hour ramp events. Figure 5.28 presents the annual duration curve for four-hour ramping events.

600 400 200 MW / 5 Min Load Ramp 0

Net-Load Ramp 1 401 801 1201 1601 2001 2401 2801 3201 3601 4001 4401 4801 5201 5601 6001 6401 6801 7201 7601 8001 8401

-200

-400

-600 Figure 5.22: Annual Duration Curve for 5-Minute Ramp Events NYISO Wind Generation Study l August 2010 34

650 600 550 500 MW / 5 Min Load Ramp 450 Net Load Ramp 400 350 300 250 Figure 5.23: Top 50 Hours of 5-Minute Up Ramp Events

-200

-300

-400 MW / 5 Min Load Ramp

-500 Net Load Ramp

-600

-700

-800 Figure 5.24: Top 50 Hours of 5-minute Down Ramp Events NYISO Wind Generation Study l August 2010 35

4000 3000 2000 1000 MW / Hour Load Ramp 0

Net Load Ramp 1 401 801 1201 1601 2001 2401 2801 3201 3601 4001 4401 4801 5201 5601 6001 6401 6801 7201 7601 8001 8401

-1000

-2000

-3000

-4000 Figure 5.25: Annual Duration Curve for 1-Hour Ramp Events 4000 3800 3600 MW / 60 Min Load Ramp 3400 Net Load Ramp 3200 3000 2800 Figure 5.26: Top 50 Hours of 1-Hour Up Ramp Events NYISO Wind Generation Study l August 2010 36

-2500

-2700

-2900 MW / 60 Min Load Ramp

-3100 Net Load Ramp

-3300

-3500

-3700 Figure 5.27: Top 50 Hours of 1-Hour Down Ramp Events 11800 9600 Max Up Load Ramp = 9,952 MW Max Up Net-Load Ramp = 12,230 MW 7400 5200 3000 MW / 4 Hours Load Ramp 800 Net-Load Ramp 1 391 781 1171 1561 1951 2341 2731 3121 3511 3901 4291 4681 5071 5461 5851 6241 6631 7021 7411 7801 8191 8581

-1400

-3600

-5800

-8000 Max Down Load Ramp = 9,332 MW Max Down Net-Load Ramp = 10,051 MW

-10200 Hours Figure 5.28: Annual Duration Curve for 4-Hour Ramp Events NYISO Wind Generation Study l August 2010 37

These graphics, which are based on 2018 loads or a peak load of 35,000 MW and installed wind of 8 GW, demonstrate that the variable nature of wind generation will result in an increase in the magnitude of ramp/load following supplied by dispatchable generation. Based on the simulations, the average 5-minute net-load up ramp will be 81.6 MW compared to an average 5-minute load ramp up of 70.8 MW. The average down 5-minute net-load ramp down of 76.2 MW compared to an average 5-minute load down ramp of 66.0 MW. The maximum 5-minute load ramp up of 560.5 MW increases to 607 MW for the net-load ramp, while the ramp down increases from 469.0 MW to 720.0 MW.

The average 1-hour load ramp up increases from 803 MW to 864 MW for the net-load ramp, while the 1-hour average down load ramp will increase from 714 MW to 769 MW. The maximum 1-hour load ramp up increases from 3,178 MW to 3,929 MW for the net load ramp, while the maximum 1-hour ramp down increases from 3,632 MW to 3,692 MW. The average 4-hour load ramp up increases from 2,752 to 2,895 MW for the net load ramp, while the maximum increases from 9,952 MW to 12,230 MW. The average 4-hour load ramp down increases from 2,649 MW to 2,789 MW for the net load, while the maximum increases from 9,334 MW to 10,052 MW. The overall conclusion is that the dispatchable resources will experience higher magnitude ramping events and be subject to a much wider range of events.

Impact of Net Load Ramping on System Dispatch NYISOs day-ahead scheduling process and security constrained economic dispatch are the primary tools for scheduling sufficient resources to supply the load as well as respond to system changes such as ramping events. The GridView production costing tool simulates the commitment and dispatch process. Data simulated by GridView were analyzed to determine how the integration of wind generation into the resource mix and the resulting increase in net-load ramping events impact dispatch. The addition of 8 GW of nameplate wind generation to the resource base will displace approximately 22,000 GWh of energy previously generated by dispatchable gas and oil fired generation. The focus of the evaluation was the impact of the wind resources on the systems thermal fossil generation which is usually the higher cost resources and is dispatched after hydro and nuclear. In addition to hydro resources, thermal fossil-fuel-fired plants are generally the primary resources that are used for ramping and load following.

The first step is to review the simulated data to determine how the total MW of thermal fossil fired generation changes with increasing installed wind generation. Figure 5.29 is a plot of an hourly duration curve, which displays the MW of thermal generation committed in the GridView simulations by hours from the highest level of hourly commitment to the lowest level of commitment. This graphic presents results for 2013 load levels and the base case of 1,275 MW of installed wind up to 6,000 MW of installed wind. Figure 5.30 presents the results for 2018 loads and the base case of 1,275 MW of installed wind and 8,000 MW of installed wind.

NYISO Wind Generation Study l August 2010 38

30000 25000 1275 MW MW 4250 MW 6000 MW 20000 15000 10000 Figure 5.29: Commitment of Thermal Fossil Generation VS Installed Wind 2013 Loads 30000 25000 1275 MW MW 8000 MW 20000 15000 10000 1

Hours Figure 5.30: Commitment of Thermal Fossil Generation VS Installed Wind 2018 Loads NYISO Wind Generation Study l August 2010 39

As expected, the addition of wind generation reduces the amount of thermal fossil fired generation that is committed to supply energy and provide load following. The simulations for the 2018 loads in the base case resulted in an average commitment of thermal fossil units of 18,287 MW for the base case or 1,275 MW of nameplate wind. The 8 GW wind scenario resulted in an average thermal fossil generating plant commitment of 16,574 MW, or a reduction of 1,713 MW, when compared to the base case. The maximum commitment was reduced from 31,842 MW to 30,571 MW while the minimum commitment was reduced from 12,707 MW to 10,094 MW. The addition of approximately 6,725 MW of nameplate wind in resulted in an average reduction in the amount of thermal fossil fired generation dispatched to meet load of approximately 25% of installed nameplate wind for the 2018 scenario. For the maximum commitment, the reduction was 1,271 MW or 18.9% of nameplate.

Although these numbers provide insight into how wind generation affects the overall level of MW of thermal fossil fired generation that are committed to supply energy and follow load, the analysis does not provide any explicit insight into whether the commitment of thermal fossil-fired plants is higher than it would otherwise be because of the variable nature of wind plant output and the need to follow it (i.e., net-load). To investigate this issue further, an analysis of the committed fossil fired generation relative to the amount energy it supplied. This was done by taking the level of committed MW of fossil fired generation for each hour and determining how much energy was produced by the committed fossil fired generation for that hour. A ratio of the committed MW divided by the MWh produced was created for each level of installed wind. This ratio was then plotted from highest to lowest ratio in the form of an hourly annual duration curve. Figure 5.31 presents the results for 2013 load levels and installed wind ranging from 1,275 MW to 6,000 MW. Figure 5.32 presents the results for 2018 loads and installed wind for 1,275 MW and 8,000 MW.

1.5 1.4 1.3 Capacity/Generation 1275 MW 4250 MW 6000 MW 1.2 1.1 1

Hours Figure 5.31: Ratio of the Committed Fossil Fired Generation MW to the Energy Produced NYISO Wind Generation Study l August 2010 40

Ratio of Fossil Online Capacity to Generation for 2018 1.5 1.4 Capacity/Generation 1.3 1275 MW 8000 MW 1.2 1.1 1

Hours Figure 5.32: Ratio of the Committed Fossil Fired Generation MW to the Energy Produced The graphics show that as you integrate more wind into the system resource mix, the ratio increases. This results because the fossil-fuel generation that has been committed is supplying less energy as wind penetration increases. It also implies that a larger percentage of the fossil fuel generation that is committed is being committed to be available when needed to provide ramping and to follow the net-load.

The final analysis that was conducted was to calculate the MW of ramping capability that was available on an hourly basis from thermal fossil-fired generation and compare it to the hourly net-load ramp requirement. This analysis was done for 2018 loads and 8 GW of wind. The hourly up and down ramping capability was calculated by determining how much capability existed between a generators operating point and its minimum generation level or its maximum generation level depending upon what was needed in that hour.

Figure 5.33 presents a plot of the hourly net-load ramp both up and down against the ramp capability available up and down.

NYISO Wind Generation Study l August 2010 41

6000 4000 2000 0

Ramp Rate ( MW / Hour )

-2000 8000 Hourly Net Load Ramp 8000 Hourly Thermal Ramping Capability

-4000

-6000

-8000

-10000

-12000 Hours Figure 5.33: Fossil Fuel Ramping Capability VS Net-Load Ramp Figure 5.33 show that there is more than sufficient ramping capability available from the fossil fueled generating units that have been committed to cover the hourly ramps. Also, the hourly down-ramp capability is far in excess of the down ramps that are observed. This is because of the need to back generation down at night to balance the load and generation during low night load conditions and yet have the generation available for the next day.

A simulation was conducted where the wind was assumed to be zero for commitment purposes but did show up in real time. This extreme sensitivity did result in an over committed system and the need to curtail wind generation to satisfy minimum load constraints. Because the NYISOs on-line processes do include wind in the commitment process it is expected that curtailment of wind generation because of minimum generation issues will be avoided.

The above analysis leads to the following conclusions regarding the increase in the magnitude of net-load ramping event:

1. The NYISO dispatch processes are sufficient to reliably respond to the increase in the magnitude of the net-load ramps that result from the integration of the MW of installed wind studied.
2. The integration of wind will result in the need to commit less fossil fueled generation for dispatch operations but the variable nature of wind generation will result in a greater percentage of the fossil fuel generation that is committed being committed to cover the increased magnitude of net-load ramping events (i.e., follow the wind).

NYISO Wind Generation Study l August 2010 42

5.4.8. Impact of Increasing Wind Penetration on Operating Reserves 8

Operating reserves are designed to cover the largest instantaneous loss of source or contingency event. The size of the current largest loss of source contingency is 1,200 MW. Reliability standards require the NYISO operations to be able to replace the instantaneous loss of 1,200 MW of energy generation within its balancing area with ten minutes. The analysis of the simulated data found for 8 GW of installed wind found a maximum drop in wind output of 629 MW in ten minutes, 962 MW in thirty minutes and 1,395 MW in an hour. The system is designed to sustain the loss of 1,200 MW instantaneously and replace it within ten minutes. Large loss of wind generation occurs over several minutes to hours. The conclusion is that wind generation will not result in any change in the amount of operating reserves the NYISO would need to have available for operations.

5.4.9. Impact of Increasing Wind Penetration on Resource Adequacy Requirements Power Systems maintain system resources over and above that which are needed to meet the expected peak load. The amount of resources available above the peak load is generally referred to as the system reserve 9

margin . They are generally expressed as a percentage of the peak load. For instance, the NYISOs current installed reserve margin which is set by the New York State Reliability Council (NYSRC) annually is 18%. This means that the NYCA must have installed resources of 118% of the peak load to meet the NYSRC reliability rule.

These reserves are primarily designed to be available when resources are unavailable because of equipment failures or maintenance requirements. They also can be designed to serve as a hedge against unexpected increases in load (i.e., load uncertainty) or transmission outages.

Because wind resources are dependent on wind, they tend to have much lower availability factors than dispatchable resources and their unavailability can be highly correlated over a large area as shown in Figure 5-3 on page 15. The addition of resources with lower unavailability to the overall resource mix will generally result in a higher installed reserve margins to meet the reliability standard. This is because when resources have low availability or high unavailability, the probability of needing to call on other resources increases resulting in an overall increase in the level of reserves needed. Conventional resources generally have overall availabilities of 85 to 90% while variable generation such as wind generally has overall availabilities of around 30%.

The Table 5-11 below presents availability expressed as the capacity factor of the wind plants used in this study, developed from the AWS Truepower simulated wind plant output. Capacity factor is a measure of the actual energy produced by a generator as a percentage of its full potential for every hour of the year. A wind plant with a 30% capacity factor produces energy that totals only 30% of the equipments potential based on its nameplate rating (nameplate times 8760 hours0.101 days <br />2.433 hours <br />0.0145 weeks <br />0.00333 months <br />). This means that the majority of the time, the wind plant is producing well below its rated potential or full nameplate. One minus the capacity factor is a measure of the expected 8

Operating reserves is the amount of resources that are needed to be available for real-time operations to cover the instantaneous and unexpected loss of resources. The New York power system is operated to protect the system against the sudden loss of 1,200 MW of resources. Operating reserve as stated is an operational concept while the reserve margin discussed in section 5.4.9 is a planning concept. The required reserve margin is designed to maintain, at an acceptable level, the risk of not having sufficient resources to avoid an involuntary loss-of-load event.

9 Reserve margin is the amount of additional capacity above the peak load that is needed so that the risk of not having sufficient capacity available to meet the load meets the minimum reliability criteria. It is expressed as a percentage and is calculated by dividing the required level of resources by the expected peak load. Resource can be unavailable because of equipment failure, maintenance outage, lack of fuel, etc. The higher the unavailability of the overall resource mix the higher the installed reserve margin will be.

NYISO Wind Generation Study l August 2010 43

unavailability of the wind plant. Data is presented for the years 2005 and 2006 because wind varies not only over the short term but from year to year as well. Of the three years of wind data AWS, the year 2005 had the lowest wind availability and 2006 had the highest. Also, data for alternative tower heights is presented. In the NYCA, the latest generation of wind plants has 80 meter tower heights while future plants are expected to move to 100 meter tower heights because of the higher capacity factors they can provide. For the study, the first 1,500 MW of installed wind was simulated with 80 meter towers while the balance was with 100 meter towers.

Table 5-11: Expected Capacity Factors for Wind Plants Capacity Factor Capacity Factor Wind Offshore for Wind Plants for Wind Plants with 100 Meter Year with 80 Meter with 100 Meter Tower Heights Tower Heights Tower Heights 2005 26.4% 30.9% 37.9%

2006 29.7% 34.0% 40.4%

To gauge how wind plants would impact the installed reserve margin as the overall percentage of the resource base that are wind generators increases, the most recent NYSRC reserve margin study (see report entitled:

10 New York Control Area Installed Capacity Requirements for the Period May 2010 through April 2011) base case result was utilized as the starting point. The base case resulted in an installed reserve margin of 17.9%

when the system calibrated to the 1 day in ten year criteria or a loss-of-load-expectation (LOLE) of 0.1 days per year. It included wind resources of 1,326 MW. The wind resources were increased to a nominal 8,000 MW. The system now had more resources than required by criteria and the LOLE dropped to 0.017 days per year.

The wind load shapes were updated to reflect the higher capacity factor of 100 meter towers based on 2006 data. Also, the NYISOs interconnection process now requires new generators to demonstrate that their capacity is deliverable to qualify for capacity payment. This is designated as Capacity Resource Interconnection Service (CRIS). Because sufficient transfer capability does not exist from upstate to downstate in order for the wind resources to qualify as capacity, transmission capacity would need to be added. The level of transfer capability necessary under the current rules was determined to be 457 MW. In addition to the updated load shapes, the NYSRC study base case transmission topology was adjusted to reflect an increase of 457 MW across the appropriate interface. The updated case was calibrated to the reliability criteria of the 0.1 days per year using the methodology of removing capacity. This resulted in a required installed reserve margin of 29.9% with 8 GW of installed wind.

The LOLE analysis also provides insight into the effective load carrying capability (ELCC) of wind generators.

ELCC is a methodology to gauge the LOLE benefits that accrue from the additional generating capability relative to its nameplate. A conventional generating resource with a nominal 5% forced outage rate and downtime for maintenance can support a load equivalent of approximately 85 to 90% of its nameplate rating. That is, it has a UCAP value or has an equivalent load carrying capability of 85 to 90% of nameplate. The addition of the incremental wind resources to the NYSRC base case reduced the LOLE to approximately .02 using the existing load shape from the 2004 study. To return the system to the minimum criteria of 0.1 days per year require the removal of 1,440 MW of perfect generation. The result is that by adding 6,654 MW of additional wind (the NYSRC base case modeled 1,346 MW of installed wind) the system was able to support an additional 1,440 MW of load carrying capability or 21.6% of nameplate. This compares to 10% for the 2004 study.

10 http://www.nysrc.org/NYSRC_NYCA_ICR_Reports.asp NYISO Wind Generation Study l August 2010 44

The additions of transfer capability to the UPNY-SENY interface, as well as the update of the wind shapes based on 2006 wind data with more coincident wind profile and to reflect the impact that more of the wind plants will be 11 built with 100 meter towers, resulted in the ELCC of the system increasing by a total of 2,500 MW. However, the addition of the increased transfer capability increases the benefits of all resources above that interface including emergency assistance which means the total increase can not be attributed to the addition of the wind resources alone.

The overall conclusion from the above, all else being equal, is that wind resources need to be supported by a larger installed base of resources because of their higher overall unavailability when compared to other resources in conjunction with their variable and less predictable nature. However, it should be noted that the NYISOs capacity market requires load serving entities to procure unforced capacity (UCAP) and capacity is derated to its UCAP value for purchase. As a result the total amount UCAP that needs to be purchased to meet reliability criteria remains essentially unchanged. The increase in reserve margin is because on capacity basis the simulations indicate that 1 MW of wind is equivalent to approximately 0.2 MW to 0.3 MW of conventional generation. The capacity value equivalent of 1 MW of existing wind plants which include a mix of 65 and 80 meter towers has ranged between 0.14 MW and 0.23 MW. Therefore, it requires a lot more installed wind to provide the same level of UCAP as a conventional generator. This results in an increase in the installed reserve margin which is computed on an installed nameplate basis.

5.4.10. Summary of findings for Task 4 Task 4 resulted in a number of findings which are summarized below:

Because of their variable nature and limited dispatchablility, the addition of wind resources on a large scale basis will result in a system that is more variable than a system without the wind resources.

The increased variability which is measured in term of the net-load deltas (i.e., load minus wind) will result in a greater magnitude ramp event which the dispatchable generation will need to respond to.

The NYISO dispatch processes are already sufficient to reliably respond to the increase in the magnitude of the net-load ramps that result from the integration of up to 8 GW of installed wind studied.

As discussed above, the increased variability will result in increasing the amount regulation capacity that is procured to maintain compliance with reliability criteria.

The addition of wind will result in a reduction of the MW of fossil fuel fired capacity that is needed to operate the system but a greater percentage of the capacity that is committed will be committed to be available to respond to higher magnitude ramping events and produce less energy.

The addition of wind will not alter the NYISOs operating reserve requirement.

The reserve margin requirement will increase as the penetration of wind resource increases because wind has a lower availability relative to other resources and its unavailability is highly correlated.

11 It should be noted that off-shore wind exhibits ELCC that is higher than on-shore wind because a greater percentage of the off-shore wind plants energy production occurs during peak hours. As an example, the GridView wind plant simulations based on 2006 wind data resulted in a 37.4% overall annual capacity factor (CF) for off-shore wind VS 34.3% for on-shore wind. However, the CF for off-shore wind plants during peak hours (the hours between 7am and 11 pm weekdays) was 39.7% for off-shore wind VS 32.5% for on-shore wind.

NYISO Wind Generation Study l August 2010 45

5.5. Results for Task 5 - Impacts of Wind Generation on Transmission Infrastructure:

5.5.1.

Introduction:

The purpose of Task 5 is to evaluate the impact of the higher penetration of wind generation on system planning from a thermal, voltage and stability perspective. This analysis serves as the foundation for determining whether additional transmission infrastructure is needed to support higher penetrations of wind. The analysis of the need for additional transmission infrastructure is discussed under Task 7. The 2009 RNA Summer 2013 peak load case was utilized as the initial study base case. Two cases were prepared to represent 4000 MW and 6000 MW of nameplate installed wind capacity. The represented wind projects were determined by the NYISO interconnection queue as presented in Table 5-1 on page 13, and project interconnection representation data were obtained from the available Interconnection Studies. For each of these cases, a corresponding off-peak case was prepared for evaluating the impact of wind output during light load conditions. The peak load cases represent a NYCA load of 35,900 MW, and the off-peak cases represent 13,400 MW load plus 1,500 MW pumping load (Niagara and Gilboa).

Thermal transfer capability was assessed using the PSS'/MUST program. In addition to the sets of normal and emergency transfer criteria contingency events in the New York bulk transmission system, tower and bus contingencies in the local 115kV area transmission system were also evaluated in the vicinity of the wind project interconnections and key transmission corridors throughout the New York transmission system. Generation source subsystems were defined for wind projects (only), non-wind generation, and all generation, and evaluated by individual zone and groups of zones: West (A+B+C), and North (D+E). Changes in the transfer limits, transfer limiting elements, and limiting contingencies for the different export subsystems help to identify the transmission constraints that need to be modeled in the production simulations in Task 6. The base case power flow wind generation was initially dispatched at 20% or rated nameplate, and the transfer simulations were uniformly increased to 100% nameplate.

5.5.2. Initial Assessment of Transmission Constraints The transfer limit results for the four power flow cases and 24 source subsystem combinations were compared to identify potential transmission constraints and if these constraints were a pre-existing transfer constraint, the result of wind generation (only), or combination of existing generation and new wind generation. Particular attention was given to constraints (limiting elements or events) that are unique to specific generation groups, individual projects, or load level. The analysis to determine limiting transmission facilities and contingencies monitored all New York transmission facilities >100kV, and evaluated critical line, bus and tower contingencies in the local area 115kV transmission system. This expanded contingency list aids in identifying where local area transmission constraints may be more limiting for individual projects or groups of projects within each Zone, and identifying local constraints that could be more limiting than typical EHV transmission system constraints.

NYISO Wind Generation Study l August 2010 46

Figure 5.34: New York Transmission Map Displaying (circles) Where Local Transmission Facilities Limit Wind Plant Output Zone A (West) - all wind projects in this zone can be uniformly increased to about 60-70% of total nameplate level before any transmission constraints are observed. Observed limitations include the Batavia -Golah 115kV for loss of a parallel Niagara - Rochester or Somerset - Rochester 345kV circuit; the actual transfer level is only slightly more limiting than the same constraint in the non-wind generation transfer case.

Zone B (Genesee) - transmission constraints occur at export levels equivalent to 25% above the combined nameplate of the projects.

Zone C (Central) - specific transmission constraints within this zone can be related to a specific project (or group of projects). The most significant constraint appears to be the projects in the vicinity of Bath 115kV being curtailed to as low as 45% nameplate by the local 115kV transmission in the peak load case with 6,000 MW wind; the constraint relaxes to 60% for the off-peak/light load scenario. Another group of projects (2 specific projects connected to the same local 115kV transmission circuit) are locally constrained to 65% nameplate (peak load, 6,000 MW wind) to 75% nameplate (peak, 4,200 MW wind). The group of projects connected along the 230kV path from Stolle Road to Hillside may also be limited by tower contingencies at either Stolle Road or Hillside 230kV at about 65% nameplate. The 115kV path between Hillside and Oakdale is both a constraint to the wind projects and general west-to-east transfers for contingencies involving EHV transmission (230kV and 345kV) at for wind generation levels of 35% nameplate (6,000 MW wind cases) to 50% nameplate (4,200 MW wind cases).

NYISO Wind Generation Study l August 2010 47

Figure 5.35: Transmission Facilities Zones A, B and C Zone D (North) - the transmission constraints - limiting facilities and limiting contingencies - in the North zone are generally centered around the Willis 230kV transmission and connections to the 115kV. The 230/115kV transformers at Plattsburgh limit corresponding projects connected to the same 230kV circuit from Willis for the opening of the Willis end of that line at 40-50% of nameplate. The next constraint is the Willis-Malone-Colton 115kV for the loss of the Moses-Willis 230kV tower at 50-70% nameplate.

Zone E (Mohawk) - the group of three (3) projects radially connected to Coffeen St. 115kV (vicinity of Watertown) are limited by local 115kV transmission radial from the interconnection point to Coffeen St. (Lyme Tap - Coffeen St. and Rockledge Tap - Lyme Tap 115kV) at 35% nameplate. Those 3 projects plus a 4th project connecting at Black River 115kV are collectively constrained by the Lighthouse Hill -- Mallory 115kV for loss of tower Taylorville - Boonville 115kV or loss of tower Black River - Taylorville 115kV; or a Black River -

Taylorville 115kV circuit limiting for the loss of tower Black River - Lighthouse Hill 115kV. These constraints are generally the same for both load levels, however, the single-circuit Taylorville - Boonville appears more limiting for non-wind (hydro and thermal) sourced transfer simulation.

NYISO Wind Generation Study l August 2010 48

Figure 5.36: Transmission Facilities Zones D and E Zone F (Capital) - all transmission constraints are significantly higher than available wind capacity and EHV transmission constraints are more limiting in the non-wind transfer simulations.

5.5.3. Evaluation of Multiple Zone Sources Western New York wind (Zones A, B, and C) - the transmission constraints observed are typical for west-to-east transfers. These include Hillside - Oakdale 115kV (40%), and Lockport - Mortimer 115kV (53%), and Delhi

- Fraser Tap 115kV (35%; light load). Lockport - Mortimer 115kV constraints are related to projects connecting east of Lockport 115kV, however the limitation is responding to loss of transmission elements between the wind project interconnection point(s) and the Mortimer (Rochester end) terminal. Although there is a wind project in the vicinity of the Delhi constraint, it is actually in Zone E and not participating in this transfer simulation; this limitation is based on the contingency loss of tower at Oakdale 345kV (Oakdale - Lapeer and Oakdale - Fraser).

Northern New York wind (Zone D, and E) - the combined wind resources in these two zones are limited by the same Willis exit (Zone D projects) and Watertown area transmission constraints (Zone E projects) that were observed in the individual Zone analyses. Although the constraints and projects affected are independent, this occurs at the combined output of 30-40% nameplate level for both load levels and both wind capacity levels.

NYISO Wind Generation Study l August 2010 49

Since all wind resources (within each defined zone) are participating equally and coincidentally, the constraints may tend to exaggerate the potential for bottling of wind resources as the commitment/dispatch of local thermal resources is not being modified by the presence of the wind generation. The set of transmission contingencies and monitored elements were provided for the analysis conducted in Task 6. The production cost simulation model is used to evaluate the impact of those constraints in the commitment and dispatch process on the potential for bottling of wind production or overall system congestion of wind generating plants.

5.5.4. Transient Stability Analysis Part of the overall evaluation of increased level of wind generation resources in the NYCA is to identify potential impact on the stability performance of the system. The evaluation should consider on- and off-peak load levels with highest expected wind production levels consistent with those load periods.

The NYCA Central East Interface was selected as the primary reference to evaluate the impact of high wind penetration on NYCA system stability performance. Central East stability performance has been shown historically to be key a factor in the dynamic performance of the NYCA as well as the northeastern portion of the Interconnection in general. Selecting this interface also recognizes that the majority of wind projects are located in Zones A through E, the source or upstream side of the Central East interface, and that the lower cost wind generation resources would tend to displace the downstate generation in Zones G through K (SENY).

The Production Cost simulations were reviewed and the hours with the highest dispatch level of wind generation were identified within either off-peak or on-peak hours. These cases represent the highest expected wind production coincident with the load. The actual load, generation commitment and dispatch were obtained from the Production Cost simulation results and imported into the powerflow model. Based on the commitment of the Oswego Complex generation, the Central East transfer level was increased to its margin transfer level for that configuration by increasing all committed generation in the Oswego Complex to maximum capability and additional generation in Zones A through E until the Central East margin test level was achieved. A third case was identified that represented off-peak load with high wind production and no Sithe/Independence generation committed. These 3 cases form the basis of the stability analyses.

Table 5-12: Powerflow Case Summaries Off-Peak/High Wind Off-Peak/High Wind On-Peak/High (without Sithe (with Sithe Units)

Wind units)

NYCA Load + Loses 17202 MW 33559 MW 16113 MW Total Wind Name Plate 7974 MW 7974 MW 7974 MW Total Wind Dispatch 6572 MW 3400 MW 6326 MW Central East Interface 3399 MW* 3390 MW* 3289 MW*

Flow Oswego Complex 3148 MW 5087 MW 2620 MW Dispatch Oswego Units 3/5 5/5 3/5 Commitment Sithe Units Commitment 4/6 6/6 0/6

  • The stressed Central East interface flow is ~110% of Central East interface stability transfer limit based on the commitment of Oswego Complex and Sithe/Independence units.

A subset of contingencies was selected for the Central East interface; these contingencies represent the most severe normal criteria fault tests in NYISO Planning and Operations evaluations of Central East/Total East stability performance.

NYISO Wind Generation Study l August 2010 50

Table 5-13: Description of Contingencies Selected for Testing ID Contingency CE01AR 3ph NC@Edic 345/Edic - N. Scotland #14 with automatic reclosing CE02 3ph NC@Marcy 345/Marcy - N. Scotland #18 CE07AR LLG NC@Edic 345/Edic/Marcy EF40/UCC41 with automatic reclosing CE08AR LLG NC@Coopers Corners 345/Coopers Corners #33/UCC41 CE12 3ph NC@N. Scotland 345/N. Scotland - Edic #14 CE15 SLG-stk@Marcy 345/Marcy #19/UE1-7 CE18AR LLG NC@Rock Tavern 345/ Rock Tavern CCRT34/CCRT42 with automatic reclosing The set of contingencies were simulated using the Siemens/PTI PSS/e program. All simulations were stable, and there were no indications of tripping of wind generation resources due to the severity of the fault or frequency or voltage excursions.

Table 5-14: Simulation Results ID Off-Peak/High Wind Off-Peak/High Wind (with Sithe Units) On-Peak/High Wind (without Sithe units)

CE01AR S S S CE02 S S S CE07AR S S S CE08AR S S S CE12 S S S CE15 S S S CE18AR S S S S - Stable Observations The NYCA (and the Interconnection) system demonstrated a stable and well damped response (angles and voltages) for all the contingencies tested on high wind generation on-peak and off-peak cases. There is no indication of units tripping due to over/under voltage or over/under frequency. The off-peak case without Sithe/Independence units showed a more oscillatory behavior than the corresponding off-peak case with Sithe/Independence units in service. This is an expected result as these units are equipped with Power System Stabilizers (PSS) and the PSS benefit to system damping and overall performance is recognized in the Central East Interface Stability Limits tables. Overall, at the high wind generation levels, the results of the simulations demonstrate that there is no adverse impact on NYCA system stability performance for both on-peak and off-peak conditions.

5.6. Results for Task 6 - Production Simulation Analysis:

5.6.1. Introduction A simulation of security constrained economic dispatch (SCED) was performed for the NYCA system in order to determine the impacts of various levels of wind generation on the balance of the system generation, primarily fossil fuel generation. ABBs GridView was the software that was utilized to simulate SCED. The modeling assumptions used in the simulations were, for the most part, those used in the CARIS except for those NYISO Wind Generation Study l August 2010 51

modifications required to conduct the wind study. The primary focus of the simulations is not to determine the economic value of wind generation, but to answer the following questions:

Question 1: What are the locational based marginal prices (LBMP) for energy or spot prices impact of introducing a large amount of price takers to the system?

Question 2: What types of generation are displaced such as coal, oil or gas?

Question 3: What is the change in production costs?

Question 4: What is the reduction in emissions?

Question 5: What is the change in imports and exports?

Question 6: What is the change in system congestion costs and uplift?

Question 7: What is the change in the capacity factors of the thermal plants?

These simulations were also used to support the evaluation of the ramping issue that was discussed in Task four as well as determining the level of wind bottling.

5.6.2. Locational Based Marginal Prices (LBMP)

Figures 5.37 through 5.45 present the LBMPs that result from the simulations of the different levels of installed wind studied for 2013 and 2018. Results are presented for the NYCA system, by superzones, and for dispatch sensitivities. The simulations were conducted utilizing the CARIS economic assumptions. The first set of LBMP results are for 2013.

200 180 160 140 LBMP ($/MWh) 120 1275 100 4250 6000 80 60 40 20 0

1 501 1001 1501 2001 2501 3001 3501 4001 4501 5001 5501 6001 6501 7001 7501 8001 8501 Hours Figure 5.37: LBMP for 2013 VS Wind Penetration for NYCA NYISO Wind Generation Study l August 2010 52

200 180 160 140 LBMP ($/MWh) 120 1275 100 4250 6000 80 60 40 20 0

1 501 1001 1501 2001 2501 3001 3501 4001 4501 5001 5501 6001 6501 7001 7501 8001 8501 Hours Figure 5.38: 2013 LBMP VS Wind Penetration for Superzones A-E 200 180 160 140 LBMP ($/MWh) 120 1275 100 4250 6000 80 60 40 20 0

1 501 1001 1501 2001 2501 3001 3501 4001 4501 5001 5501 6001 6501 7001 7501 8001 8501 Hours Figure 5.39: 2013 LBMP VS Wind Penetration for Superzones F-I NYISO Wind Generation Study l August 2010 53

200 180 160 140 LBMP ($/MWh) 120 1275 100 4250 6000 80 60 40 20 0

1 501 1001 1501 2001 2501 3001 3501 4001 4501 5001 5501 6001 6501 7001 7501 8001 8501 Hours Figure 5.40: 2013 LBMP VS Wind Penetration for Superzones J-K Table 5-15: Summary of Average LBMP for 2013 Zone Average LBMP ($/MWh) by Installed Nameplate Wind 1,275 MW 4,250 MW 6,000 MW System 69.5 66.8 65. 1 Zone A-E 63.5 58. 9 56.0 Zone F-I 70.2 68.1 66. 9 Zone J-K 73. 8 71. 7 70. 7 The next set of graphics presents the results for 2018. The results for 2018 include two additional sensitivities.

The simulations assume that the wind that is committed is the wind that is available for economic dispatch. To test the impact on prices of an error in the wind commitment, two sensitivities were conducted for the 8 GW of wind scenario. The first sensitivity was the extreme case which did not commit for wind but allowed it to generate during economic dispatch while the second was to simulate the amount of wind that generated during economic dispatch to be available with a 10% mean absolute percent error (MAPE) when compared with the committed wind.

NYISO Wind Generation Study l August 2010 54

250 200 LBMP ($/MWh) 150 1275 8000 100 50 0

1 501 1001 1501 2001 2501 3001 3501 4001 4501 5001 5501 6001 6501 7001 7501 8001 8501 Hours Figure 5.41: 2018 LBMP VS Wind Penetration for the NYCA 250 200 LBMP ($/MWh) 150 1275 8000 100 50 0

1 501 1001 1501 2001 2501 3001 3501 4001 4501 5001 5501 6001 6501 7001 7501 8001 8501 Hours Figure 5.42: 2018 LBMP VS Wind Penetration for Superzones A-E NYISO Wind Generation Study l August 2010 55

250 200 LBMP ($/MWh) 150 1275 8000 100 50 0

1 1001 2001 3001 4001 5001 6001 7001 8001 Hours Figure 5.43: 2018 LBMP VS Wind Penetration for Superzones F-I 250 200 LBMP ($/MWh) 150 1275 8000 100 50 0

1 501 1001 1501 2001 2501 3001 3501 4001 4501 5001 5501 6001 6501 7001 7501 8001 8501 Hours Figure 5.44: 2018 LBMP VS Wind Penetration for Superzones J-K NYISO Wind Generation Study l August 2010 56

250 200 LBMP ($/MWh) 150 8000 8000 Zero UC 8000 10% MAPE 100 50 0

1 501 1001 1501 2001 2501 3001 3501 4001 4501 5001 5501 6001 6501 7001 7501 8001 8501 Hours Figure 5.45: 2018 LBMP VS Wind Penetration for NYCA for the no Unit Commitment and 10% MAPE Commitment Sensitivity Table 5-16: Summary of Average LBMP for 2018 Zone Average LBMP ($/MWh) by Installed Nameplate Wind 1,275 MW 8,000 MW No Commitment 10% MAPE System 86.6 78.7 70.1 79.2 Zone A-E 79.1 67.2 57.5 67.9 Zone F-I 87.3 80.8 71.3 81.2 Zone J-K 91.4 85.6 78.0 85.9 Summary of Findings for LBMP:

What is important in this analysis is not the nominal value of the prices but the overall trend of the prices. The production cost simulations indicate that as significant amounts of essentially zero production cost generation is added to the resource mix, which participate as price takers, LBMP or spot prices decline as expected. For the 2018 simulations, the NYISO average LBMP prices are 9.1% lower for the 8 GW wind scenario when compared to the base case or 1,275 MW installed wind case. The reduction from the base case when compared to the 10%

MAPE sensitivity is 8.5%. The dispatch sensitivities indicate the impact of incorporating wind into the commitment process and the how forecast error of wind resources can affect prices. Also, note the LBMP price impacts are greatest in the superzones where the wind generation is located and tends to increase the price spread between upstate, where the wind resources are primarily located in the study, and downstate which imply increasing congestion costs.

NYISO Wind Generation Study l August 2010 57

The decline in spot prices is generally a positive development for buyers but will result in lower energy revenues and energy production for dispatchable generation. The ultimate impact of these lower LBMPs involves a complex set of issues which are interrelated with other aspects of the NYISO market structure and design. The study of these interactions was beyond the scope of this report and, therefore, was not analyzed in this report.

5.6.3. Fuel Types Displaced by Wind Generation Figures 5.46 through 5.53 present the results from the simulations that display what fuels are displaced by the introduction of wind generation into the resource mix.

60000 50000 40000 Generation in GWh 1275 30000 4250 6000 20000 10000 0

Gas Turbine Steam Gas Combined Cycle Gas Oil Coal Wind Fuel Type Figure 5.46: Fuel Types Displaced for 2013 for the NYCA NYISO Wind Generation Study l August 2010 58

18000 16000 14000 12000 Generation in GWh 10000 1275 4250 6000 8000 6000 4000 2000 0

Gas Turbine Steam Gas Combined Cycle Gas Oil Coal Wind Fuel Type Figure 5.47: Fuel Types Displaced for 2013 for Superzone A-E 20000 18000 16000 14000 12000 Generation in GWh 1275 10000 4250 6000 8000 6000 4000 2000 0

Gas Turbine Steam Gas Combined Cycle Gas Oil Coal Wind Fuel Types Figure 5.48: Fuel Types Displaced for 2013 for the Superzone F-I NYISO Wind Generation Study l August 2010 59

25000 20000 15000 Generation in GWh 1275 4250 6000 10000 5000 0

Gas Turbine Steam Gas Combined Cycle Gas Oil Coal Wind Fuel Types Figure 5.49: Fuel Types Displaced for 2013 for the Superzone J-K 60000 53091.9 50000 44382.2 40000 Generation in GWh 1275 30000 8000 21829.3 20000 19391.6 18926.5 12005.2 10470.3 10000 8358.1 8417.4 3501.6 2307.4 1646.7 0

Gas Turbine Steam Gas Combined Cycle Gas Oil Coal Wind Fuel Type Figure 5.50: Fuel Types Displaced for 2018 for the NYCA NYISO Wind Generation Study l August 2010 60

18000 16000 14000 12000 Generation in GWh 10000 1275 8000 8000 6000 4000 2000 0

Gas Turbine Steam Gas Combined Cycle Gas Oil Coal Wind Fuel Types Figure 5.51: Fuel Types Displaced for 2018 for the Superzone A-E 20000 18000 16000 14000 12000 Generation in GWh 1275 10000 8000 8000 6000 4000 2000 0

Gas Turbine Steam Gas Combined Cycle Gas Oil Coal Wind Fuel Types Figure 5.52: Fuel Types Displaced for 2018 for the Superzone F-I NYISO Wind Generation Study l August 2010 61

25000 20000 15000 Generation in GWh 1275 8000 10000 5000 0

Gas Turbine Steam Gas Combined Cycle Gas Oil Coal Wind Fuel Types Figure 5.53: Fuel Types Displaced for 2018 for the Superzone J-K Summary of Findings for Fuel Displacement:

The primary fuel displaced by increasing penetration of wind generation is natural gas. For the simulations with 8 GW of wind with 2018 loads, the total amount of fossil fired generation displaced was approximately 15,500 GWh. Gas fired generation accounted for approximately 13,000 GWh or approximately 84% of the total, while oil and coal accounted for approximately 2,050 GWh and 465 GWh respectively or approximately 13% and 3% of the total fossil generation displaced.

5.6.4. Wind Generation Impact on System Production Costs The addition of wind resources with virtually zero marginal costs to the NYCA resource mix will result in the reduction of overall system production costs. The Figures 5.54 and 5.55 present the results for the impact of wind generation on system production costs as the level of installed wind generation increases.

NYISO Wind Generation Study l August 2010 62

6200 Production Costs Millions of $s 6000 5800 5969 5600 5400 5516 5200 5308 5000 4800 1275 MW 4250 MW 6000 MW Figure 5.54: Change in Production Costs for 2013 as the Level of Installed Wind Generation Increases 9000 8000 Production Costs Millions of $s 7000 7834 6000 6484 5000 4000 3000 2000 1000 0

1275 MW 8000 MW Figure 5.55: Change in Production Costs for 2018 as the Level of Installed Wind Generation Increases Summary of Findings for Wind Generation Impact on System Production Costs As the amount of wind generation increases, the overall system production costs decrease. For the 2013 study year, the production costs drop from the base case total of almost $6 billion to a level of approximately $5.3 billion for the 6,000 MW wind scenario. This represents a drop of 11.1% in production costs. For the 2018 study year, the production costs drop from the base case total of almost $7.8 billion to a level of approximately $6.5 billion for the 8,000 MW wind scenario. This represents a drop of 16.6% in production costs.

NYISO Wind Generation Study l August 2010 63

5.6.5. Wind Generation Impact on Emissions Production of electricity by wind generators is emissions free. The Figures 5.56 through 5.61 display the changes in emissions for the New York power grid for CO2, NOx and SO2 as a function of increasing wind penetration.

58,000,000 57,419,342 57,000,000 56,000,000 55,000,000 53,857,182 54,000,000 53,000,000 52,005,281 52,000,000 51,000,000 50,000,000 49,000,000 1275 4250 6000 Figure 5.56: Reductions in CO2 (short tons) as Wind Generation Increases for 2013 NYISO Wind Generation Study l August 2010 64

39,500 39,093 39,000 38,500 38,000 37,500 37,061 37,000 36,500 36,049 36,000 35,500 35,000 34,500 1275 4250 6000 Figure 5.57: Reductions in NOx (short tons) as Wind Generation Increases for 2013 65,000 63,981 64,000 63,000 62,000 61,381 61,000 60,000 59,606 59,000 58,000 57,000 1275 4250 6000 Figure 5.58: Reductions in SO2 (short tons) as Wind Generation Increases for 2013 NYISO Wind Generation Study l August 2010 65

58,000,000 57,419,342 57,000,000 56,000,000 55,000,000 54,000,000 53,000,000 52,511,096 52,000,000 51,000,000 50,000,000 1275 8000 Figure 5.59: Reduction in CO2 (short tons) as Wind Generation Increases for 2018 39,500 39,093 39,000 38,500 38,000 37,500 37,000 36,500 36,359 36,000 35,500 35,000 34,500 1275 8000 Figure 5.60: Reduction in NOx (short tons) as Wind Generation Increases for 2018 NYISO Wind Generation Study l August 2010 66

68,000 66,109 66,000 64,000 62,000 60,000 59,632 58,000 56,000 1275 8000 Figure 5.61: Reduction in SO2 (short tons) as Wind Generation Increases for 2018 Summary of Findings for Emission Reductions:

For the 2018 load levels, the dispatch simulations with 8 GW of wind resources when compared to the base case which includes 1275 MW of installed wind resulted in a reduction of 4.9 million short tons of CO2 or an 8.5%

reduction, a reduction of approximately 2,730 short tons of NOx or a 7% reduction and a reduction of approximately 6,475 short tons of SO2 or a 9.7% reduction. Each GWh of fossil fired generation displaced results in an average reduction in CO2 of approximately 315 tons. The total reduction of emissions would be higher except some of the wind generation is bottled by local transmission constraints.

5.6.6. Changes in Imports and Exports The introduction of wind generation into the NYCAs resource mix with its much lower marginal costs of operation generally should tend to reduce imports and increase exports because of the relative price changes.

The Figures 5.62 through 5.67 present the imports, exports and net for the 2013 and 2018 wind scenarios.

NYISO Wind Generation Study l August 2010 67

6,000 5,000 4,000 Power Flow in GWh 1275 MW 3,000 4250 MW 6000 MW 2,000 1,000 0

New England PJM Ontario Neptune HVDC Hydro-Quebec Cross Sound Cable Interface Figure 5.62: Imports for 2013 as Wind Plant Penetration Increases 2,500 2,000 Power Flow in GWh 1,500 1275 MW 4250 MW 6000 MW 1,000 500 0

New England PJM Ontario Neptune HVDC Hydro-Quebec Cross Sound Cable Interface Figure 5.63: Exports for 2013 as Wind Plant Penetration Increases NYISO Wind Generation Study l August 2010 68

6,000 5,000 4,000 3,000 Power Flow in GWh 2,000 1275 MW 4250 MW 6000 MW 1,000 0

-1,000

-2,000

-3,000 New England PJM Ontario Neptune HVDC Hydro-Quebec Cross Sound Cable Interface Figure 5.64: Net Import/Exports for 2013 as Wind Plant Penetration Increases 7,000 6,000 5,000 Power Flow in GWh 4,000 1275 MW 8000 MW 3,000 2,000 1,000 0

New England PJM Ontario Neptune HVDC Hydro-Quebec Cross Sound Cable Interface Figure 5.65: Import for 2018 as Wind Plant Penetration Increases NYISO Wind Generation Study l August 2010 69

2,500 2,000 1,500 Power Flow in GWh 1275 MW 8000 MW 1,000 500 0

New England PJM Ontario Neptune HVDC Hydro-Quebec Cross Sound Cable Interface Figure 5.66: Export for 2018 as Wind Plant Penetration Increases 6,000 5,000 4,000 3,000 Power Flow in GWh 2,000 1275 MW 8000 MW 1,000 0

-1,000

-2,000

-3,000 New England PJM Ontario Neptune HVDC Hydro-Quebec Cross Sound Cable Interface Figure 5.67: Net Import/Exports for 2018 as Wind Plant Penetration Increases NYISO Wind Generation Study l August 2010 70

Summary of Findings for Imports/Exports:

For the 2018 simulations, the net imports for the NYISO show very little change as the level of installed wind increases while net exports increased. On an interface by interface basis, the results vary. Exports to New England and Ontario increase. Imports from Ontario decline while New England imports remain unchanged across all wind scenarios. The Neptune and CSC HVDC cables can only be used for imports and the imports decline as the installed wind increases. The Hydro Quebec interface was modeled as a schedule. The PJM interface flow changes run counter to an expectation of increasing exports and decreasing imports, with imports increasing and exports decreasing. This contradiction is the result of the wind resource addition in New York resulting in additional loop flow. Because of the physics of the power grid, the increased energy production in western New York will show up as an increase in loop flow. These increases in loop flow are evident in the increase in Lake Erie counterclockwise circulation. This increased circulation will show up as an export on the Ontario ties and an import on the PJM eastern ties. This increase in circulation was determined to have no significant impact on the results the study was focused on.

The analysis of production data in neighboring areas that are tied synchronously to New York shows slightly less electrical energy being produced in these areas while NYs total increases. This means as expected that on balance NYs imports decrease while its exports increase.

5.6.7. Changes in Congestion Payments and Uplift It is expected that as more low cost generation is added in Upstate New York relative to the Downstate load centers congestion cost would increase. This outcome was indicated in the LBMP analysis. Also, since the fossil fuel generation that is committed will generate less energy and has to some extent been committed to respond to larger magnitude net-load ramping events, the expectation is that uplift could increase as installed wind generation increases. Figures 5.68 through 5.71 present the congestion payments for the 2013 and 2018 installed wind scenarios for superzones F-I and J-K which are the superzones most impacted by congestion.

NYISO Wind Generation Study l August 2010 71

120.0 Congestion Payments Millions of $s 100.0 80.0 60.0 Superzone F-I 109.5 91.7 40.0 59.6 20.0 0.0 1275 4250 6000 Figure 5.68: Congestion for 2013 by the Level of Installed Wind for Superzones F-I 500.0 450.0 Congestion payments Millions of $s 400.0 350.0 300.0 250.0 Superzone J-K 435.7 200.0 376.5 150.0 256.6 100.0 50.0 0.0 1275 4250 6000 Figure 5.69: Congestion for 2013 by the Level of Installed Wind for Superzones J-K NYISO Wind Generation Study l August 2010 72

160.0 Congestion Payments Millions of $s 140.0 120.0 100.0 80.0 151.0 Superzone F-I 60.0 40.0 81.8 20.0 0.0 1275 8000 Figure 5.70: Congestion for 2018 by the Level of Installed Wind for Superzones F-I 700.0 Congestion Payments Millions of $s 600.0 500.0 400.0 Superzone J-K 300.0 616.8 200.0 377.3 100.0 0.0 1275 8000 Figure 5.71: Congestion for 2018 by the Level of Installed Wind for Superzones J-K NYISO Wind Generation Study l August 2010 73

Figures 5.72 through 5.73 present the uplift cost for 2013 and 2018 as the level of installed wind generation increases. The GridView model calculates the uplift cost on a daily basis. Uplift is the difference between a generators daily production costs and its LBMP payments.

250 215.7 212.9 219.0 200 Uplift Millions of $s 149.2 154.9 150 139.6 1275 MW 4250 MW 100 6000 MW 50 39.9 29.9 31.8 0

A-E F-I J&K Superzones Figure 5.72: Uplift Costs for 2013 as the Level of Installed Wind Increases 350 310.7 300 285.7 Uplift Costs Millions of $s 250 191.9 200 170.3 1275 MW 150 8000 MW 100 49.6 50 38.6 0

A-E F-I J&K Superzones Figure 5.73: Uplift Costs for 2018 as the Level of Installed Wind Increases Summary of Findings for Congestion Payments and Uplift:

As suggested by the LBMP trends, the congestion payments in superzones F-I and J-K increase as the level of installed wind generation is increased. The overall increase on a percentage basis as measured against the base case to 6,000 MW of wind in 2013 and 8,000 MW in 2018 ranges from a high of 85% for superzone F-I in 2013 to a low of 64% for superzone J-K in 2018. Also, the higher loads in 2018 tend to increase congestion payments while the addition of wind resources to superzone J-K in 2018 puts downward pressure on congestion payments.

NYISO Wind Generation Study l August 2010 74

Uplift costs tend to increase in superzones A-E and F-I as the level of installed wind generation increases which is expected. Superzone J-K uplift costs are, for the most part, flat as the level of installed wind increases for 2013 but actually decrease for 2018. This is the result of the offshore wind which has a capacity factor of almost 39%

and tends to be more coincident with the daily load cycle and displaces high cost on-peak generation in the superzone while requiring less capacity for higher magnitude ramping events.

5.6.8. Changes in Thermal Plant Capacity Factors The Tables 5-17 through 5-20 present the average annual capacity factors and how they change for those thermal unit and fuel types displaced by wind generation as the level of installed wind increases.

Table 5-17: Thermal Plant Capacity Factors for the NYISO Installed Wind Fuel 1275(2013) 4250 6000 1275(2018) 8000 GT-NG 8.1% 7.1% 6.7% 11.3% 8.0%

ST-NG 27.1% 23.5% 22.0% 32.3% 22.5%

CC-NG 63.0% 57.6% 57.6% 65.1% 54.9%

OIL 9.8% 9.0% 8.8% 11.7% 9.4%

COAL 83.9% 83.0% 81.9% 83.9% 81.9%

Table 5-18: Thermal Plant Capacity Factors for Superzone A-E Installed Wind Fuel 1275(2013) 4250 6000 1275(2018) 8000 GT-NG 4.9% 2.5% 1.8% 7.7% 2.5%

ST-NG 14.0% 10.4% 9.4% 17.2% 9.6%

CC-NG 49.8% 40.6% 35.0% 52.9% 34.9%

OIL 0.1% 0.1% 0.1% 0.2% 0.1%

COAL 83.3% 82.3% 81.0% 83.4% 81.0%

Table 5-19: Thermal Plant Capacity Factors for Superzone F-I Installed Wind Fuel 1275(2013) 4250 6000 1275(2018) 8000 GT-NG 0.6% 0.4% 0.4% 1.2% 0.6%

ST-NG 13.3% 10.2% 9.0% 19.0% 9.9%

CC-NG 69.0% 64.5% 61.9% 71.1% 61.7%

OIL 0.2% 0.2% 0.1% 0.2% 0.2%

COAL 87.3% 87.3% 87.3% 87.3% 87.3%

Table 5-20: Thermal Plant Capacity Factors for Superzone J-K NYISO Wind Generation Study l August 2010 75

Installed Wind Fuel 1275(2013) 4250 6000 1275(2018) 8000 GT-NG 8.5% 7.6% 7.2% 11.9% 8.7%

ST-NG 32.0% 28.2% 26.6% 37.0% 26.9%

CC-NG 64.9% 62.1% 60.8% 67.7% 60.7%

OIL 13.9% 12.8% 12.6% 16.7% 13.4%

Summary of Findings for Thermal Plant Capacity Factors:

Consistent with the findings for the fuel displacement analysis, the plants with the biggest decline in annual capacity factors are the natural gas fired plants. The capacity factors for the thermal plants are, as expected, impacted negatively by the addition of increasing wind plant penetration but positively by increasing load. The biggest reduction in annual capacity factors from 2013 base case level of 1,275 MW of wind when compared to the 8 GW scenarios occurs for the combined cycle plants in all superzones with a 30% decline in superzone A-E, 11% decline in superzone F-I and 6% decline superzone J-K.

5.7. Results for Task 7 - Identify Transmission System Upgrades:

5.7.1. Identification of Bottled Wind Resources The results of the Task 6 simulations are analyzed to identify the transmission constraints - local and system -

that result in potential wind energy production being limited (i.e., bottled). The active constraint(s) for each instance of energy bottling is reviewed and potential upgrade(s) are applied and the production cost simulation repeated. The iterative process continues until the wind energy bottling in each Zone is below 2% and/or NYCA-wide bottling is below 2%.

The production cost simulation results were evaluated to identify transmission congestion levels and the type(s) of resources that were being curtailed by transmission constraints. Transmission constraints were identified as candidate for upgrade if they resulted in curtailing wind energy production by more than 2% of the potential energy production on a zonal basis, or where an individual projects capacity factor was curtailed by more than 10%. Upgrades that could be considered in the initial evaluation were limited to incremental conductor or line-terminal upgrades (at the same operating voltage) or limited reconfiguration or existing physical plant.

The production cost simulations in Task 6 identified the same three general areas of congestion: southwestern portion of Central (Zone C), Willis (Zone D), and Watertown (Zone E). As the affected projects within these areas are generally independently constrained by different transmission elements and contingencies, the transmission upgrade scenarios could be tested in parallel. The Task 7 process is the step-by-step evaluation of each transmission corridor that constrains wind resources, and the evaluation of the upgrade: applying revised ratings or contingency definitions to the model, running successive simulations, and evaluating changes in wind resource bottling. This process is repeated until the wind resource bottling is less than 2% on a zonal basis, or all projects capacity factors within the zone are curtailed by less than 10% (actual vs. potential capacity factor).

NYISO Wind Generation Study l August 2010 76

Evaluation of the transmission upgrades process was performed primarily on the 6,000 MW level case. The initial production cost simulation identified three zones with high levels of bottling (Central, North and Mohawk).

Within each of these zones, specific constraints were related to individual or groups of wind projects to provide further guidance to the selection of transmission upgrades based on the extent of the bottling and number of effected projects.

The process tested a series of transmission upgrades and reinforcements that was able to minimize the bottling of wind resources in NYCA to less than 2%, and within all zones individually to less than 2% except in the Mohawk Zone. Within the Mohawk Zone the bottling was specific to the projects in the Watertown area.

Of all of the EHV (230kV and above) constraints that were tested in the production cost simulations, only three contributed to any significant level of wind resource bottling. Most of the wind resource bottling for the individual projects is caused by the local (115kV) transmission system between the projects and the EHV transmission system connection points.

At each step in the analysis, estimated costs for the upgrades are indicated. These are generic estimates based on the following assumptions:

Upgrade Estimated Cost Reconductor existing 115kV construction $500,000-750,000/circuit mile Rebuild existing 115kV $750,000-1,500,000/circuit mile Build new 115kV $1,200,000-2,250,000/circuit mile Build new 230kV $2,250,000/circuit mile Line terminal upgrade $250,000-1,250,000/terminal Protection upgrades $250,000-500,000/terminal The cost ranges reflect consideration of conductor size (ampacity or rating) and extent (and number) of tower reinforcement for existing lines, and greater structural strength for new or upgraded construction or complete rebuild of an existing line. Line terminal upgrades range from disconnect switches, station connections, or miscellaneous equipment (metering, CTs, PTs, wavetraps, etc.), to circuit breakers and station bus work. Line terminal upgrades also assume that the upgraded equipment will fit within the footprint of the existing equipment.

Transmission owners have provided information to assist in identifying the limiting element(s) within each circuit and provide guidance as to the extent of structure rebuilding that would be necessary to accommodate any proposed reconductoring.

5.7.2. Overview of Transmission Upgrades The analysis of the production cost simulations conclude that there are no major EHV reinforcements that are needed to accommodate the wind resource nameplate capacity levels of 6,000 MW or 8,000 MW. The existing NYCA EHV system congestion continues to follow historic patterns that follow the west-east/north-south flow patterns. As more wind resources are added to the NYCA the levels of congestion hours does not change significantly, but the relative value of the congestion increases as wind (as a price taker) displaces higher-cost generation resources.

NYISO Wind Generation Study l August 2010 77

Table 5-21: Summary of Base Case Wind Resource Bottling Zone 1275 1275 (2018 load) 4250 6000 8000 A 119 0.0% 119 0.0% 935 0.0% 1309 0.1% 1510 0.1%

B 6 0.1% 6 0.1% 86 0.0% 281 0.1% 418 0.1%

C 393 0.0% 393 0.0% 1110 6.7% 1591 6.3% 1860 6.2%

D 387 3.7% 387 3.7% 717 9.4% 1068 15.0% 1068 15.0%

E 368 0.0% 368 0.0% 1398 6.5% 1648 15.5% 1648 15.6%

F 0.1 0.0% 0.1 0.0% 0.1 0.0% 70 0.1% 70 0.2%

J 700 0.0%

K 700 0.0%

Total 1275 1.1% 1275 1.1% 4247 5.6% 5967 8.8% 7974 6.6%

There were only three identified EHV contingencies that cause significant wind resource bottling; all are double-circuit tower contingencies.

In the Elmira area of Zone C the 230kV double-circuit tower contingency (loss of Canandaigua - Hillside 230kV

  1. 68 and Hillside - Watercure 230kV # 69) generally limits wind resources west of Elmira area by overloading of local 115kV transmission circuits in the vicinity of the Hillside station. In the Binghamton area of Zone C the 345kV double-circuit tower contingency (loss of Oakdale - Fraser 345kV #32 and Oakdale - (Lapeer) Lafayette 345kV #36) generally limits wind resources west of the Binghamton area by overloading the 115kV transmission facilities east of Oakdale to the Delhi 115kV station.

In Zone D, the wind resources in the vicinity of Willis and east toward Plattsburgh are limited for the 230kV double-circuit tower contingency (loss of both Moses - Willis 230kV MW-1 and MW-2) by the 115kV transmission path between Willis and Colton.

5.7.3. Transmission Upgrades for 6,000 MW Buildout The initial transmission upgrades considered were mitigation of these 3 tower contingencies. In each instance the extent of the double-circuit structures is 6 or 7 towers exiting the respective line terminals at Moses/St.

Lawrence 230kV, Hillside 230kV and Oakdale 345kV, and the remaining distance of each circuit is on single-circuit structures. Mitigation of these contingencies would involve limited reconstruction of the 6- or 7-tower sections as individual single-circuit structures.

The 3 tower contingencies noted were removed from the Production Cost model, and the simulation was repeated to identify the next limiting constraints (contingencies and limiting facilities). As each project or group of projects and the corresponding constraints were evaluated in each subsequent production cost simulation, the selection of upgrades were considered based on the lowest cost for incremental transmission capacity benefit. If a limiting element has found a rating that is less than the design conductor rating, the first step would upgrade the facility to allow operation at the design conductor rating. If the facility is still limiting after that upgrade (conductor rating), reconductoring would be considered and the conductor size would be selected to minimize the need to rebuild all structures on the right-of-way end to end, if possible. The following details the process of evaluating necessary upgrades to accommodate 6,000 MW wind capacity and meet the objective of less than 2% state-wide and zonal bottling (when comparing the potential wind energy to the actual constrained energy production).

NYISO Wind Generation Study l August 2010 78

5.7.4. Step 1: Tower contingency mitigations The three double-circuit tower contingencies are reconfigured to single circuit structures:

Moses-Willis 230kV MW1&2 reconfigure (7 towers) at Moses exit Canandaigua-Hillside #68/Hillside-Watercure #69 230kV (7 towers) at Hillside exit Oakdale-Fraser #32/Oakdale-Lafayette #36 345kV (6 towers) at Oakdale exit The approximate cost for each of these upgrades is estimated at $2,000,000 to reconfigure a minimum of 3-4 towers at each existing location, assuming the double-circuit towers are each replaced with two (side-by-side) single-circuit structures on the existing alignment.

The double-circuit contingencies were removed from the model and the production cost simulation was run to identify the next limitations:

Zone C:

Montour Falls - Hillside 115kV for loss of Canandaigua - Hillside 230kV Hillside - No. Waverly 115kV for loss of Watercure - Oakdale 345kV Canandaigua - Avoca - Hillside 230kV pre-contingency loading Zone D:

Plattsburgh 230/115kV transformers #1 and #4 Willis - Plattsburgh 230kV circuits. The loss of the Willis end section of either circuit results in the wind resources connected between Willis and Plattsburgh being radially connected through the transformer at the Plattsburgh end.

Zone E  :

Delhi - Delhi Tap - Colliers 115kV for loss of Oakdale - Fraser 345kV Black River - Taylorville 115kV for loss of tower Black River - Lighthouse Hill Taylorville - Lowville 115kV for loss of tower Black River - Lighthouse Hill The limiting facilities identified above in Zones D and E, and the Canandaigua - Hillside 230kV line are rated lower than the design conductor rating; the upgrade for these elements will identify components within the circuit (terminal equipment including disconnects, wave traps, etc.) that should be replaced to allow operation at design ratings. The 115kV facilities identified in Zone C are proposed to be reconductored based on the observed level of these constraints.

NYISO Wind Generation Study l August 2010 79

5.7.5. Step 2 Upgrades:

The following upgrades are based on the limitations observed after step 1 above:

Zone C:

The Montour - Hillside 115kV (2) circuits are reconductored (replace existing 336ACSR with 795ACSR) from Montour to the Ridge Road Taps; line terminal upgrades at Montour Falls (buswork and disconnect switches) and Hillside (buswork) are necessary to accommodate the higher rated conductor. The cost to rebuild the existing Montour Falls - Ridge Tap line sections (9.5 miles/each) and terminal equipment upgrades is approximately $20,900,000.

The Hillside - North Waverly 115kV is reconductored replacing the existing 2-4/0 ACSR with 795ACSR; and upgrade the buswork at North Waverly and line terminal equipment at Hillside. The cost to rebuild 17 miles of transmission line and terminal equipment is estimated at $17,500,000 The Canandaigua - Avoca - Hillside 230kV circuit is limited by CTs, wavetraps and bus work at the Hillside terminal; replacement of this equipment is estimated to be $1,000,000 to allow operation of the line up to its design conductor rating (1033 ACSR).

Zone D:

The line and transformer terminal connections at Plattsburgh 230/115kV are upgraded to allow operation of the 2-230/115kV transformers (#1 and #4) and the Plattsburgh end sections of the 230kV lines at design ratings. In reviewing these limitations, NYPA staff indicated that the replacement of the bus connections at Plattsburgh 115kV has been completed.

Zone E:

The line terminals for the Delhi - Delhi Tap - Colliers 115kV circuit are upgraded to allow operation of the line to design conductor rating (1033 ACAR). The limitation on this 3-terminal line is the distance protection relay settings; this upgrade is estimated to be $750,000 based on replacement of the existing relays at each terminal.

This upgrade primarily benefits projects in Zone C.

The Black River - Taylorville #1, 2, & 8 115kV lines are upgraded to conductor rating (336ACSR) by upgrading station connections at several locations (est. total $600,000).

The corresponding ratings and circuit impedances in the simulation network model were updated to reflect the ratings and reconductoring of the limiting circuits and the simulation was repeated.

NYISO Wind Generation Study l August 2010 80

The next level of constraints was identified:

Zone C:

Bennett - Howard - Bath - Montour Falls 115kV pre-contingency loading Bennett - Moraine Rd - Meyer 115kV for loss of Howard - Bath 115kV Zone D:

Moses - Willis 230kV MW-1 for loss of Moses - Willis MW-2 Zone E:

Lighthouse Hill - Mallory 115kV for loss of tower Black River - Taylorville Coffeen St. - E.Watertown, Coffeen St. - Black River, and Lyme Tap -Coffeen St. 115kV (all) pre-contingency loading Most of the limiting elements identified in this step can be upgraded to operate to conductor rating by replacing terminal equipment. Based on the bottling levels observed, the Lighthouse Hill - Mallory (Clay) and Lyme Tap -

Coffeen St. 115kV lines should be reconductored; however, the existing structures may not be capable of accommodating the necessary ampacity conductor.

5.7.6. Step 3 Upgrades:

The following upgrades are based on the limitations observed after the steps above:

Zone C:

The Bennett - Howard - Bath 115kV circuit can be upgraded to the 477ACSR conductor rating (780A) by replacing terminal equipment (breaker, CTs, disconnect switches) at Bath. (est. $1,000,000)

The Bath - Montour Falls 115kV circuit can be upgraded to the 602ACSR conductor rating (900A) by replacing terminal equipment (breaker, CTs, disconnect switches) at Bath and Montour Falls. (est. $2,000,000)

The Bennett - Moraine Rd - Meyer 115kV circuit can be upgraded to the 1033ACSR conductor rating (1250A) by replacing terminal equipment (breakers, CTs, disconnect switches and buswork) at Bennett and Meyer. (est.

$2,000,000).

The (normally open) Andover - Palmiter Road 115kV is returned to in-service as recommended in the Canisteo Wind SRIS.

NYISO Wind Generation Study l August 2010 81

Zone D:

The Moses - Willis MW-1 and MW-2 230kV circuits are upgraded to design conductor rating (795ACSR). This will require replacing the bus, breaker and line connections at the St. Lawrence/FDR and Willis terminals (est.

$2,000,000).

Zone E:

The Lighthouse Hill - Mallory 115kV circuit is rebuilt (replace 4/0 CU with 795ACSR); the estimated cost to rebuilt the 26.5 miles of line and upgrade the terminals is $41,855,000.

The Coffeen St. - Black River 115kV circuit is upgraded to design conductor rating (336ACSR) by replacing existing station connections at an estimated cost of $500,000.

The Lyme Tap - Coffeen St. 115kV circuit is rebuilt (replace 336ACSR with 795ACSR) and station connections upgraded for 1140A; est. cost for 6.9mi and line terminal upgrades is $9,588,000.

The Rockledge Tap - Lyme Tap 115kV is upgraded (connections at Lyme Tap) to operate at conductor rating (795ACSR) as identified in the Cape Vincent SRIS (est. $250,000).

As in the previous step (2), the corresponding ratings and circuit impedances in the simulation network model were updated to reflect the new ratings and reconductoring of the limiting circuits and the simulation was repeated.

The constraints identified in Step 3 are:

Zone C:

The Meyer - Eel Pot Rd - ECOGEN/GlobalNY-Flat St-Greenidge 115kV are limited pre-contingency Zone D:

The Plattsburgh 230/115kV transformers limiting pre-contingency and contingency Zone E:

The Taylorville-Boonville 115kV for loss of Lighthouse Hill - Mallory 115kV 5.7.7. Step 4 Upgrades:

The following upgrades are based on the limitations observed after the steps above:

NYISO Wind Generation Study l August 2010 82

Zone C:

The Meyer - Greenidge 115kV path is limiting between the wind projects interconnection point to Greenidge and is only limiting when the 2 projects are in operation; if only one project is ultimately realized, no upgrade is necessary; the line sections from the project interconnection points to Greenidge are upgraded to conductor rating (336ACSR) by replacing existing 4/0CU buswork at Greenidge (est. $250,000).

Zone D:

The existing configuration of the Plattsburgh terminal does not have the high-sides of the 230/115kV transformers paralleled. To mitigate the contingency overloads of the transformers caused by power flowing from the 230kV to the 115kV and back to the 230kV through the other transformer the upgrade would require building out a full 230kV switchyard at an estimated cost of $14-16 million to mitigate the overloading of the transformers.

As the wind resource bottling in the Zone is below the 2% target, this may not be cost-justified based on the remaining level of bottling.

Zone E:

The Taylorville-Boonville 115kV circuits are upgraded to operate at design conductor rating (336ACSR); this will involve replacement of station connections or buswork (est. $1,000,000).

The corresponding ratings in the simulation model were updated to reflect the upgrades and the simulation was repeated. The results indicate that the existing conductor is not adequate. In Zone E essentially all of the wind resource bottling is in the Watertown vicinity. The bottling is of sufficient magnitude that a conductor sized to accommodate the expected level of wind energy production is likely to involve a major rebuilding of the Watertown area transmission.

5.7.8. Step 5 Upgrades:

The following upgrades are based on the limitations observed after the steps above:

Zone C:

Eel Pot Road - ECOGEN/GlobalNY - Flat St - Greenidge 115kV is reconductored (replace 336ACSR with 477ACSR) for 28.9mi and station connections upgraded to 780A (est. $15,400,000).

Zone E:

Black River - Lighthouse Hill 115kV and Taylorville-Boonville 115kV circuits are rebuilt with 795ACSR for the entire length of both double-circuits (total 78.5 right-of-way miles). Including the intermediate and terminal upgrades (1140A) at 8 locations, the cost would be $119,868,000.

After these upgrades, the Watertown area wind resources remain significantly constrained by the Black River -

Taylorville 115kV path for loss of tower Black River - Lighthouse Hill or Lighthouse Hill - Mallory 115kV. In Zone NYISO Wind Generation Study l August 2010 83

C, the reconductoring of the ECOGEN - Greenidge lines relieved the last constraint. The wind resource bottling in Zones C and D is below the 2% target; in Zone E, only the projects in the Watertown vicinity are severely constrained.

5.7.9. Steps 6 - 7 Upgrades:

In Steps 6 and 7 additional reconductoring to identify a feasible transmission upgrade to accommodate the projects in the Watertown area is tested. These steps, when combined with the reconductoring already testing in preceding steps, effectively rebuilds the entire 115kV transmission system in the Watertown area. Review with the transmission owner (NationalGrid) indicates that the necessary conductor sizes cannot be accommodated by the existing structures and, including the Step 5 rebuilds above, will represent a complete rebuilding of over 200 circuit miles of 115kV transmission lines.

Step 6:

Rebuild Black River-N. Carthage #1, Black River-Taylorville #2, and N. Carthage-Taylorville #8 115kV (approx.

26.5mi. double-circuit tower); and replace 4/0 CU with 795ACSR, upgrade station connections to 1140A; estimated cost $38,693,000.

Step 7:

Rebuild Coffeen St - Black River #3 115kV replace 336ACSR with 795ACSR; 8.9mi., and station upgrades (1140A) estimated cost $9,160,000.

Indian River - Black River #9 provides the radial connection for the Dutch Gap project; upgrading this circuit to the conductor rating (795ACSR) involves buswork and station connections at Black River; est. cost $500,000. A limited amount bottling of the project output may still occur but may not justify reconductoring the circuit.

Additionally, the radial transmission behind Coffeen St. to Lyme Tap and the Rockledge Tap - Lyme Tap sections of the #4 115kV circuit severely constrain the output of the 3 projects that connect to that circuit.

Reconductoring Coffeen St. - Black River #3 with 795 ACSR is not sufficient. Alternatives further tested nd rebuilding the existing with conductor sizes up to 1192ACSR, or 2-795ACSR and adding a 2 parallel circuit (additional $10-20,000,000). The additional sub-scenarios (7a, 7b, and 7c) test larger conductor sizing to determine the extent of upgrades to accommodate the remaining bottling.

NYISO Wind Generation Study l August 2010 84

Reconductoring the Rockledge Tap - Lyme Tap - Coffeen St. (1192ACSR) and Coffeen St. - Black River #3 circuit (795ACSR) will reduce the overall resource bottling in Zone E to just under 2%, but at an additional cost of up to $24,545,000 (complete rebuild). The discussion with the Transmission Owner indicated that the reconductoring would require a complete rebuild. An alternative upgrade of the Rockledge - Coffeen St. would nd be the addition of a 2 circuit (15 miles) in parallel with the existing circuit (this assumes the rebuilding the nd Coffeen St. - Lyme Tap section in Step 3). The cost of the 2 (795ACSR) circuit is estimated $24,500,000.

Table 5-22: Summary of Wind Resource Bottling - 6,000 MW Base Case Upgrades Zone Wind Capacity Base Step 1 Step 2 Step 3 Step 4 Step 5 Step 6 Step 7 A 1309 0.1% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%

B 281 0.1% 0.1% 0.0% 0.1% 0.0% 0.0% 0.0% 0.0%

C 1591 6.1% 4.5% 3.9% 1.2% 0.2% 0.0% 0.0% 0.0%

D 1068 15.0% 12.0% 2.5% 1.7% 1.7% 1.7% 1.7% 1.7%

E 1648 15.8% 15.1% 14.0% 11.1% 10.4% 11.0% 8.0% 3.3%

F 70 0.1% 0.1% 0.1% 0.1% 0.2% 0.1% 0.1% 0.2%

Total 5967 8.8% 7.7% 5.4% 3.7% 3.2% 3.4% 2.5% 1.2%

5.7.10. Alternative Solution The upgrades evaluated in Steps 5, 6 and 7 for the Watertown vicinity (Zone E) would require the complete rebuilding of over 200 circuit-miles of existing 115kV transmission lines, and there could still be a significant level of wind resources bottled in the Watertown vicinity. The extent of local 115kV transmission reconstruction may not be feasible.

An alternative solution to rebuilding the existing Watertown area 115kV network is to build a new 230kV transmission line between the existing Coffeen St. 115kV (Watertown) to Adirondack (Taylorville) 230kV stations (approximately 40 miles). The new 230kV line would be connected to the existing Coffeen St. 115kV station with a pair of 300MVA 230/115kV autotransformers.

NYISO Wind Generation Study l August 2010 85

Figure 5.74: Transmission Map with 230kV Upgrade (dotted line)

The 230kV overbuild avoids the rebuilding of the Taylorville - Boonville, Black River - Taylorville, and Black River - Lighthouse Hill lines (approximately 185 circuit-miles or 104 ROW miles). Cost of the new 230kV station at Coffeen St., including the autotransformers and associated switchgear is estimated to be $35-38,000,000; additional 230kV line terminal (2 breakers and associated switchgear) at Adirondack is estimated to be $2-4,000,000. The cost of the 230kV line, including cost of new right-of-way, is approximately $2,250,000/mile, or

$90,000,000 - or $133,600,000 complete. This compares favorably with the estimated $158,561,000 for the nd 115kV rebuild. The Rockledge - Lyme Tap - Coffeen St. 115kV upgrade is necessary in either case with the 2 th 115kV circuit to accommodate the output of the 3 projects that radially connect to Coffeen St., and the 4 project connecting at Black River.

All of the previously indicated upgrades in the other Zones (C and D) were unchanged. These upgrades and any of the indicated line terminal upgrades to conductor rating proposed for Zone E facilities in steps 5 and 6 are included in System Upgrades. The relative resource bottling levels are presented for the 115kV-rebuild or 230kV-overbuild alternatives.

Table 5-23: Comparison of Watertown Alternatives - 6,000 MW Case Wind System Watertown Watertown Zone Capacity Base Case Upgrades 115kV Alt. 230kV Alt.

A 1309 0.1% 0.0% 0.0% 0.0%

B 281 0.1% 0.0% 0.0% 0.0%

C 1591 6.2% 0.4% 0.3% 0.5%

D 1068 11.3% 1.7% 1.7% 1.7%

E 1648 13.7% 8.2% 3.2% 3.6%

F 70 0.1% 0.1% 0.1% 0.1%

Total 5967 7.6% 2.7% 1.3% 1.4%

NYISO Wind Generation Study l August 2010 86

The data reported in Table 5-23 can also be expressed in terms of the bottled energy (MWh) or the unrealized wind energy production. This also provides additional guidance in the determination of the relative value of transmission reinforcements for specific project(s) or for a specific Zones projects.

Table 5-24: Bottled Energy (MWh) Summary - 6,000 MW Case 6000 Base Case - Bottled Energy (MWh)

Wind System Watertown Watertown Zone Capacity Base Case Upgrades 115kV Alt. 230kV Alt.

A 1309 1,965 1,720 1,708 1,684 B 281 682 310 226 398 C 1591 286,368 16,380 16,093 21,438 D 1068 365,160 53,504 53,459 53,278 E 1648 647,623 390,202 153,768 171,055 F 70 217 247 244 295 Total 5967 1,302,014 462,363 225,498 248,149 5.7.11. Transmission Upgrades for the 8,000 MW Buildout The installed wind resource locations of the additional 2,000 MW consist of the 1,400 MW off-shore wind connecting to the NYC and LI Zones (J and K) while the remaining 600 MW is connecting at locations that were previously not constrained. Based on the locations of the additional capacity, the same sets of proposed transmission upgrades developed for the 6,000 MW case were applied to the 8,000 MW case. The resulting wind energy bottling levels and constraints were consistent with the 6,000 MW case results and did not indicate a need for any modification to the upgrade test sequence. The Watertown alternate reinforcement scenarios were also evaluated in the 8,000 MW case.

Table 5-25: Summary of Wind Resource Bottling - 8,000 MW Base Case Upgrades Wind System Watertown Watertown Zone Capacity Base Case Upgrades 115kV Alt. 230kV Alt.

A 1510 0.1% 0.1% 0.1% 0.1%

B 418 0.1% 0.0% 0.0% 0.0%

C 1860 6.2% 0.5% 0.5% 0.6%

D 1068 11.6% 1.7% 1.7% 1.7%

E 1648 13.5% 7.7% 3.0% 2.9%

F 70 0.2% 0.4% 0.4% 0.4%

J 700 0.0% 0.0% 0.0% 0.0%

K 700 0.0% 0.0% 0.0% 0.0%

Total 7974 5.8% 1.9% 1.0% 1.0%

NYISO Wind Generation Study l August 2010 87

5.7.12. Assessment of the Relative Cost/Value of the Benefits of the Transmission Upgrades Studied During the review of study results, the NYISO was asked by stakeholders to provide a measure of the cost or value of the transmission upgrades relative to the benefits they provided in terms of the unbottled wind plant energy. The transmission analysis evaluated the effectiveness of each upgrade in terms of improving the wind energy production as either project (or Zone) capacity factors or total wind energy production by Zone. The incremental wind energy production (i.e., the difference in production before/after a transmission upgrade step) and the estimated cost for each step are summarized in the tables below.

Table 5-26: Cumulative Transmission Upgrade Costs ($1000)

Zone C Zone D Zone E NYCA Step 1 $5,000.00 $2,000.00 $0.00 $7,000.00 Step 2 $45,150.00 $2,000.00 $2,050.00 $49,200.00 Step 3 $50,150.00 $4,000.00 $12,389.00 $66,539.00 Step 4 $50,400.00 $20,000.00 $13,389.00 $83,789.00 Step 5 $65,800.00 $20,000.00 $133,257.00 $219,057.00 Step 6 $65,800.00 $20,000.00 $172,450.00 $258,250.00 Step 7 $65,800.00 $20,000.00 $206,110.00 $291,910.00 The unbottled-incremental wind production is the increase in the total wind resource energy production that resulted from the transmission improvements associated with each upgrade step.

Table 5-27: Cumulative Unbottled Wind Plant Energy Production (MWH)

Zone C MWHr Zone D MWHr Zone E MWHr NYCA Total Step 1 22,743 94,849 33,311 200,710 Step 2 100,646 402,878 89,446 893,574 Step 3 227,663 429,246 222,479 879,845 Step 4 220,893 429,008 259,250 959,857 Step 5 280,279 428,993 229,822 939,697 Step 6 280,158 429,068 393,115 1,082,937 Step 7 280,228 429,088 603,021 1,312,918 To provide a relative measure of the cost or value of the each upgrade, the NYISO developed a metric that relates the transmission upgrade costs to the increased MWH of wind energy production. This metric was developed using a simplified approach which assumes that the capital costs of the transmission upgrade are to be recovered over a period of 15 years and did not account for the time value of money or other carrying charges. The transmission cost/MWH of unbottled wind plant energy was calculated for each transmission upgrade step for each Zone. Table 5-28 presents the results of this calculation.

NYISO Wind Generation Study l August 2010 88

Table 5-28: 15 Year Annualized Transmission Upgrade Cost/MWH Zone C $/MWHr Zone D $/MWHr Zone E $/MWHr NYCA Total Step 1 $4.58 $1.41 $0.00 $2.33 Step 2 $29.91 $0.33 $1.53 $5.53 Step 3 $14.69 $0.62 $3.71 $5.04 Step 4 $12.40 $3.11 $3.44 $5.82 Step 5 $15.65 $3.11 $38.66 $15.54 Step 6 $15.66 $3.11 $30.81 $15.90 Step 7 $15.65 $3.11 $22.79 $14.82 These results can be presented graphically:

Transmission Upgrade $/MWHr (15yr)

$45.00

$40.00

$35.00 Annualized Cost of Upgrade $/MWhr

$30.00

$25.00

$20.00

$15.00

$10.00

$5.00

$0.00 Step 1 Step 2 Step 3 Step 4 Step 5 Step 6 Step 7 Upgrade Step/Scenario C $/MWHr D $/MWHr E $/MWHr Total Figure 5-75: Summary Upgrade Cost/MWH Although the results are presented by Zone, there is limited interaction among transmission upgrades in one Zone impacting (either beneficially or adversely) projects in an adjacent Zone. Some represented projects experience minimal curtailment over the entire range of the study, while certain projects may require substantial transmission upgrades to achieve expected capacity factors. As presented in Table 5-28 there are considerable differences in the relative costs/value between the upgrade steps both within and across zones. Again, this analysis has been conducted to provide some insight into the relative value or costs of the upgrades relative to the unbottled wind energy production.

NYISO Wind Generation Study l August 2010 89

5.7.13. Summary of Transmission Upgrades Analysis Feasible sets of upgrades were developed and tested for the 6,000 MW and 8,000 MW wind resource build-out cases. Additional alternatives were suggested and tested to address the severe levels of resource bottling that occurs in the Watertown vicinity. The suggested transmission upgrades and alternatives require detailed physical review and economic evaluation before a final set of recommendations can be determined. Below is a summery of summary of the findings and upgrades identified:

There were three EHV transmission contingencies identified that should be mitigated.

There were five 230kV transmission lines where upgrading of line terminal facilities are necessary to allow the conductor to be operated at design ratings and will significantly reduce the level of bottled wind resources.

The remaining transmission upgrades identified can be associated with small groups of projects within a locality or Zone.

Some transmission upgrades are only necessary if all projects within a congested group are built.

With the exception of the Watertown vicinity of Zone E, most of the transmission upgrades involve local 115kV transmission circuits, and most of the resource unbottling can be obtained through limited upgrading of the terminal facilities associated with those circuits.

A limited number of 115kV transmission circuits would need to be rebuilt with higher ampacity conductor to accommodate connection of specific project(s).

The most severe wind resource bottling occurs in the vicinity of Watertown (in all projected buildout cases); this is due, in part, to the extent of double-circuit tower construction and relatively light conductor in use.

The existing transmission infrastructure in the Watertown area makes reconductoring upgrades more likely to require a complete rebuilding as part of any effort to upgrade with higher-ampacity (larger size) conductor.

An alternative upgrade for the Watertown vicinity was evaluated consisting of a new 230kV transmission circuit from the existing Coffeen St. 115kV station (Watertown) to the Adirondack 230kV station, and a 2nd 115kV circuit from the Rockledge Tap station to Coffeen St.

The similarity of the bottling patterns in the 6,000 MW and 8,000 MW cases reflects the common locations of the capacity additions being studied. The most significant (1,400 MW) of the new wind resources added to expand from the 6,000 MW case to the 8,000 MW case is located in the NYC and LI zones, and 35% of the remaining resources added in Zones A through E connect directly to existing 345kV There were no significant capacity additions in the vicinity of projects that were already constrained.

The 8,000 MW shows the same severe bottling in the Watertown vicinity as the 6,000 MW, and the transmission upgrades achieve similar results.

NYISO staff and the Transmission Owners are continuing to review the identified transmission constraints and the feasibility of the facility upgrades proposed.

NYISO Wind Generation Study l August 2010 90

6. Wind Study Conclusions 6.1. Overall Study Findings The primary finding of the study is that wind generation can supply reliable clean energy at a very low cost of production to the New York power grid. This energy results in significant savings in overall system production costs, reductions in greenhouse gases such as CO2 and other emissions such as NOx and SO2 as well as results in an overall reduction in wholesale electricity prices. However, wind plants require a significant upfront capital investment. In addition, wind plants because of their variable nature and the uncertainty of wind plant output provide more of a challenge to power system operation than conventional power plants. This study determined that the NYISOs systems and procedures (which includes the security constrained economic dispatch and the practices that have been adopted to accommodate wind resources) will allow for the integration of up to 8 GW of installed wind plants without any adverse reliability impacts.

This conclusion is predicated on the assumption that a sufficient resource base is maintained to support the wind. The study determined that 8 GW of wind would reduce the need for conventional or dispatchable fossil fired generation on the order of 1.6 to 2 GW or an amount equivalent to 20-25% of the installed nameplate wind.

This is the result of wind generations much lower overall availability when compared to conventional generation.

This means an amount of fossil generation equivalent to 75-80% of the nameplate installed wind needs to be available for those times when the wind isnt blowing or the wind plant output is at very low levels of output. Non-wind generation will be needed to respond to the higher magnitude ramps that will result because of winds variable nature.

As wind resources are added to the resource mix, their lower availability could result in an increase in the installed reserve margin and a decline in spot market prices. The impact of these changing conditions has not been analyzed in this report.

The fluctuating nature and the uncertainty associated with predicting wind plant output levels manifests itself as an increase in overall system variability as measured by the net load (load minus wind). In response to these increased operational challenges the NYISO has implemented changes to its operational practices such as being the first ISO to incorporate wind resources into security constrained economic dispatch (SCED) and to implement a centralized forecasting process for wind resources. The study concluded that at higher levels of installed wind generation the system will experience higher magnitude ramping events and will require additional regulation resources to respond to increased variability during the five minute dispatch cycle. The analysis determined that the average regulation requirement will need to increase by approximately 9% for every 1,000 MW increase in wind generation between the 4,250 MW and 8,000 MW level of installed wind.

Although the addition of wind to the resource mix resulted in significant reduction in production costs, the reduction would have been even greater if transmission constraints between upstate and downstate were eliminated. These transmission constraints prevent lower cost generation in upstate New York from displacing higher costs generation in southeast New York. This report did not analyze the potential financial impact of an increase in transfer capability from upstate into southeast New York.

Finally, the study determined that almost 9% of the potential upstate wind energy production will be bottled or not deliverable because of local transmission limitations. The study identified feasible sets of transmission facility upgrades to eliminate the transmission limitations. These upgrades were evaluated to determine how much of the wind energy that was undeliverable would be deliverable if the transmission limitations were removed.

NYISO Wind Generation Study l August 2010 91

Additional alternatives were suggested and evaluated to address the significant levels of resource bottling that occurs in the Watertown vicinity. The suggested transmission upgrades and alternatives require detailed physical review and economic evaluation before a final set of recommendations can be determined.

6.2. Summary of Study Results The study has determined that as the level of installed wind plant MW increases, system variability as measured by the net-load increases for the system as whole. The increase exceeds 20% on an average annual basis from current levels for the 8 GW wind scenario and 2018 loads. The level of increase varies by season, month, and time-of-day. This will result in higher magnitude ramping events in all timeframes whether it is minute-to-minute or hour-to-hour that the dispatchable resources will need to respond to. Study results are reported for the New York system as a whole and for three superzones which are the western load zones A-E, the Hudson Valley load zones F-I, and the New York City and Long island load zones J-K. The study resulted in the following findings with respect to system reliability, system operations and dispatch, and transmission planning impacts of increasing the level of installed wind beyond the 3,300 MW originally studied.

6.2.1. Reliability Finding:

This study has determined that that the addition of up to 8 GW of wind generation to the New York power system will have no adverse reliability impact. The 8 GW of wind would supply in excess of 10% of the systems energy requirement. On a nameplate basis, 8 GW of wind exceeds 20% of the expected 2018 peak load. This finding is predicated on the analysis presented in this report and the following NYISO actions and expectations:

The NYISO has established a centralized wind forecasting system for scheduling of wind resources and requires wind plants to provide meteorological data to the NYISO for use in forecasting their output. This item was approved by the Federal Energy Regulatory Commission (FERC) and implemented by the NYISO in 2008.

The NYISO is the first grid operator to fully integrate wind resources with economic dispatch of electricity through implementation of its wind energy management initiative. If needed to maintain system security, the NYISO system operators can dispatch wind plants down to a lower output. This item was approved by the Federal Energy Regulatory Commission (FERC) and implemented by the NYISO in 2009.

The NYISOs wind plant interconnection process requires wind plants: 1) To participate fully in the NYISOs supervisory control and data acquisition processes; 2) To meet a low voltage ride through standard; and 3) conduct voltage testing to evaluate whether the interconnection of wind plants will have an adverse impact on the system voltage profile at the point of interconnection. In addition, the NYISO will continue to integrate best practice requirements into its interconnection processes.

The NYISOs development of new market rules assists in expanding the use of new energy storage systems that complement wind generation. This item was approved by the Federal Energy Regulatory Commission (FERC) and implemented by the NYISO in 2009.

The NYISOs installed resource base will have sufficient resources to support wind plant operations. As described in this report, the overall availability of wind resources is much less than other resources and their variability (changing output as wind speed changes) increases the magnitude of the ramps. For a system that meets its resource adequacy criteria (e.g., the 1 day in ten years), the additions of 1 MW of resources generally means that 1 MW of existing resources could be removed and still meet the resource adequacy criteria.

However, the addition of 1 MW of wind would allow approximately 0.2 MW to 0.3 MW of existing resources to be removed in order to still meet the resource adequacy criteria. The balance of the conventional generation must NYISO Wind Generation Study l August 2010 92

remain in service to be available for those times when the wind plants are unavailable because of wind conditions and to support larger magnitude ramp events.

6.2.2. Operation and Dispatch Simulation Findings:

Analysis of the wind plant output and dispatch simulations resulted in the following findings for the expected impact of wind plant output on system operations and dispatch:

Finding One - Analysis of five minute load data coupled with a ten minute persistence for forecasting wind plant output (i.e., wind plant output was projected to maintain its current level for the next five minute economic dispatch cycle) concluded that increased system variability will result in a need for increased regulation resources. The need for regulation resources varies by time of day, day of the week and seasons of the year.

The analysis determined that the average regulation requirement increases approximately 9% for every 1,000 MW increase between the 4,250 MW and 8,000 MW wind penetration level. The analysis for 8 GW of wind and 2018 loads (37,130 MW peak) resulted in the overall weighted average regulation requirement increasing by 116 MW. The maximum increase is 225 MW (a change from a 175 MW requirement up to 400 MW) for the June-August season hour beginning (HB) 1400. The highest requirement is 425 MW in the June-August season HB2000/HB2100.

Finding Two - The amount of dispatchable fossil generation committed to meet load decreases as the level of installed nameplate wind increases. However, a greater percentage of the dispatchable generation is committed to respond to changes in the net-load (load minus wind) than committed to meet the overall energy needs of the system. The magnitudes of ramp or load following events are reduced when wind is in phase with the load (i.e.,

moving in the same direction). However, for many hours such as the morning ramp or the evening load drop, wind is out of phase with the load (i.e., moving in the opposite direction). This results in ramp or net-load following events that are of higher magnitude than those that would result from changes in load alone. It is these ramp or load following events to which the dispatchable resources must respond.

Finding Three - Simulations with 8 GW of installed wind resulted in hourly net-load up and down ramps that exceeded by approximately 20% the ramps that resulted from load alone. It was also determined from the simulations the NYISO security constrained economic dispatch processes are sufficient to reliably respond to the increase in the magnitude of the net-load ramps. This finding is based on the expectation that sufficient resources will be available to support the variability of the wind generation. For example, the data base used for these simulations had installed reserve margins which exceeded 30%.

Finding Four - Simulations for 8 GW of wind generation concluded that no change in the amount of operating 12 reserves was needed to cover the largest instantaneous loss of source or contingency event. The system is designed to sustain the loss of 1,200 MW instantaneously with replacement within ten minutes where as a large loss of wind generation occurs over several minutes to hours. The analysis of the simulated data found for 8 GW of installed wind a maximum drop in wind output of 629 MW occurred in ten minutes, 962 MW in thirty minutes and 1,395 MW in an hour, respectively.

12 Operating reserves is the amount of resources that are needed to be available for real-time operations to cover the instantaneous and unexpected loss of resources. The New York power system is operated to protect the system against the sudden loss of 1,200 MW of resources. Operating reserve as stated is an operational concept while the reserve margin discussed in section 6.2.3 is a planning concept. The reserve margin level is designed to maintain the risk of not having sufficient resource to meet the load to the acceptable level.

NYISO Wind Generation Study l August 2010 93

6.2.3. Resource Adequacy Findings:

To evaluate the impact of wind resources on NYISO installed reserve requirements, the study started with the 13 New York State Reliability Council (NYSRC) Installed Reserve Margin Study for the 2010-2011 Capability 14 Year. The NYSRC base case had an installed reserve margin of 17.9% to meet loss-of-load-expectation (LOLE) criteria of 0.1 days per year. That base case was updated to bring the installed wind resources to the full 8 GW of wind studied. The analysis of a system with this level of installed wind resulted in the following findings.

Finding One - The addition of 8 GW of wind resources to the NYSRC base case reduced the LOLE from the 0.1 days per year to approximately 0.02 days per year.

Finding Two - At criteria, the NYISO reserve margin would have to increase from its current level of 18% to almost 30% with 8 GW of nameplate wind as part of the resource mix. This was determined by using the methodology of removing capacity to bring the system to criteria and adding transfer capability in order for the wind plants to qualify for Capacity Rights Interconnection Service (CRIS). However, it should be noted that the NYISOs capacity market requires load serving entities to procure unforced capacity (UCAP) and capacity is derated to its UCAP value for purchase. As a result the total amount UCAP that needs to be purchased to meet reliability criteria remains essentially unchanged. The increase in reserve margin is because on capacity basis 1 MW of wind is equivalent to approximately 0.2 MW of conventional generation. Therefore, it requires a lot more installed wind to provide the same level of UCAP as a conventional generator. This results in an increase in the installed reserve margin which is computed on an installed nameplate basis.

Finding Three - The LOLE analysis resulted in an effective load carrying capability (ELCC) for the wind plants studied that exceeded 20%. The ELCC for this study exceeded the ELCC finding in the 2004 study by a factor of

2. Off-shore wind exhibits ELCC that is higher than on-shore wind because a greater percentage of the off-shore wind plants energy production occurs during peak hours. As an example, the GridView wind plant simulations based on 2006 wind data resulted in a 37.4% overall annual capacity factor (CF) for off-shore wind VS 34.3% for on-shore wind. However, the CF for off-shore wind plants during peak hours (the hours between 7am and 11 pm weekdays) was 39.7% for off-shore wind VS 32.5% for on-shore wind.

6.2.4. Production Cost Simulation Findings:

The production cost simulations conducted with ABBs GridView economic dispatch simulation model and the base case transmission system resulted in the following findings:

Finding One - As the amount of wind generation increases, the overall system production costs decrease. For the 2013 study year, the production costs drop from the base case total of almost $6 billion to a level of approximately $5.3 billion for the 6,000 MW wind scenario. This represents a drop of 11.1% in production costs.

For the 2018 study year, the production costs drop from the base case total of almost $7.8 billion to a level of approximately $6.5 billion for the 8,000 MW wind scenario. This represents a drop of 16.6% in production costs.

The change in production costs reflect the commitment of resources that are needed to support the higher magnitude ramping events but do not reflect the costs of the additional regulating resources.

13 Reserve margin is the amount of additional capacity above the peak load that is needed so that the risk of not having sufficient capacity available to meet the load meets the minimum reliability criteria. It is expressed as a percentage and is calculated by dividing the required level of resources by the expected peak load. Resource can be unavailable because of equipment failure, maintenance outage, lack of fuel, etc. The higher the unavailability of the overall resource mix the higher the installed reserve margin will be.

14 http://www.nysrc.org/NYSRC_NYCA_ICR_Reports.asp NYISO Wind Generation Study l August 2010 94

Finding Two - Based on the economic assumptions used in the CARIS study, locational-based marginal prices (LBMP) or spot Energy prices decline as significant amounts of essentially zero production cost generation that participates in the market by using price taker bids is added to the resource mix. For the 2018 simulations, the NYISO system average LBMP prices are 9.1% lower for the 8 GW wind scenario when compared to the base case or 1,275 MW of installed wind.

Finding Three - The LBMP price impacts are greatest in the superzones where the wind generation is located and tends to increase the price spread between upstate where wind is primarily located in the study and downstate, which implies an increase in transmission congestion.

Finding Four - The primary fuel displaced by increasing penetration of wind generation is natural gas. For the simulations with 8 GW of wind with 2018 loads, the total amount of fossil-fired generation displaced was approximately 15,500 GWh. Gas-fired generation accounted for approximately 13,000 GWh or approximately 84% of the total. Oil and coal accounted for approximately 2,050 GWh and 465.1 GWh respectively, or approximately 13% and 3% of the total fossil generation displaced.

Finding Five - As suggested by the LBMP trends, the congestion payments in superzones F-I and J-K increase as the level of installed wind generation is increased. The overall increase in congestion payments on a percentage basis as measured against the base case compared to 6,000 MW of wind in 2013 and 8,000 MW in 2018 ranges from a high of 85% for superzone F-I in 2013 to a low of 64% for superzone J-K in 2018.

Finding Six - The addition of wind resources to superzone J-K in the 2018 case puts downward pressure on LBMPs in those zones, and therefore lowers congestion payments.

Finding Seven - Uplift costs tend to increase in superzones A-E and F-I as the level of installed wind generation increases. Superzone J-K uplift costs are for the most part flat as the level of installed wind increases for 2013 but actually decrease for 2018. This is the result of the offshore wind which has a capacity factor of almost 39%

and tends to be more coincident with the daily load cycle and displaces high cost on peak generation in the superzone while requiring less capacity for higher magnitude ramping events. Off shore wind also provides greater capacity benefits.

Finding Eight - The capacity factors for the thermal plants are, as expected, decreased by the addition of wind plants, but this is partially offset by increasing load. The biggest reduction in annual capacity factors from the 2013 base case level of 1,275 MW of wind when compared to the 8 GW scenarios occurs for the combined cycle plants in all superzones with a 30% decline in superzone A-E, 11% decline in superzone F-I and 6% decline superzone J-K. As would be expected the biggest impact is in the superzone with the highest level of installed wind with transmission capacity limitations between the superzones contributing to the reduction.

6.2.5. Environmental Findings:

For the 2018 load levels, the dispatch simulations with 8 GW of wind resources resulted in the following emissions reductions in comparison to the base case with1, 275 MW of installed wind:

Finding One - A CO2 emission reduction of approximately 4.9 million short tons or a reduction of 8.5%.

Finding Two - Each GWh of displaced fossil-fired generation which primarily consisted of natural gas resulted in an average reduction in CO2 of approximately 315 tons.

Finding Three - A NOx emission reduction of approximately 2,730 short tons or a reduction of 7%.

NYISO Wind Generation Study l August 2010 95

Finding Four - A SO2 emissions reduction of 6,475 short tons or a reduction of 9.7%.

6.2.6. Transmission Planning Findings:

Extensive power flow analysis in conjunction with dispatch simulations was conducted to determine the impact of transmission system limitations on the energy deliverability of the wind plant output. The analysis resulted in the following findings:

Finding One - Given the existing transmission system capability, the 6 GW scenario determined that 8.8% of the energy production of the wind plants in three areas in upstate New York would be bottled or not deliverable.

Finding Two - The primary location of the transmission constraints was in the local transmission facilities or 115 kV voltage level.

Finding Three - The off-shore wind energy as modeled was fully deliverable and feeds directly into the superzone J-K load pockets.

Finding Four - The study evaluated 500 miles of transmission lines and 40 substations to determine potential upgrades that would result in the unbottling of the wind energy.

Finding Five - If all the upgrades studied were implemented, the amount of wind energy not deliverable would be reduced to less than 2% for the upstate wind.

Finding Six - Depending on the scope of upgrades required, such as reconductoring of transmission lines compared to rebuilding or upgrading terminal equipment, the cost of the upgrades could range from $75 million to $325 million. However, it should be noted that many of the transmission facilities studied are approaching the end of their expected useful lives.

Finding Seven - Transient Stability Analysis was conducted to evaluate the impact of high wind penetration on NYCA system stability performance. The primary interface tested was the Central East. The Central East stability performance has been shown historically to be key factor in the dynamic performance of the NYISO power grid.

The NYISO power grid (and the Interconnection) system demonstrated a stable and well damped response (angles and voltages) for all the contingencies tested on high wind generation on-peak and off-peak cases.

There is no indication of units tripping due to over/under voltage or over/under frequency.

Finding Eight - Wind plants that are in the NYISO interconnection 2008 class year study and beyond may require system deliverability upgrades to qualify for Capacity Resource Integration Service (CRIS). This totals approximately 4,600 MW of new nameplate wind plants that were included in the study. In order to qualify for capacity payments, the wind plants in class year 2009/2010 and beyond in upstate New York would need to increase transmission transfer capability between upstate New York and southeast New York (a.k.a., the UPNY-SENY interface). This transmission interface primarily consists of 345 kV transmission lines in the Mid-Hudson valley region running through Greene County, New York south of Albany to Dutchess County, New York or between Zones E and F and Zone G. The study determined that approximately 460 MW of interface transfer capability needs to be added to this interface for the wind plants that did not qualify for capacity payments to be eligible for them. This does not impact the deliverability of the wind plants energy but only their ability to qualify for capacity payments or CRIS.

NYISO Wind Generation Study l August 2010 96

Appendix A: Summary Wind Plant Performance Metrics For 2009 Nameplate Average Peak Hour Coincidence Number Of Days Month Total MW Capacity Factor Max 1 HR Output MW with Hrs < 0 (avg. daily) Factor (CF) 1,2 January 978.8 29.7% 9.1% 838.4 2 February 1140.3 35.7% 28.4% 997.2 1 March 1273.9 24.0% 28.9% 1002.2 5 April 1273.9 32.2% 38.7% 1058.5 0 May 1273.9 23.1% 8.1% 1070.5 2 June 1273.9 10.2% 25.8% 625.8 7 July 1273.9 15.9% 12.5% 769.7 4 August 1273.9 13.2% 16.5% 716.4 5 September 1273.9 14.7% 9.5% 1001.2 7 October 1273.9 21.7% 8.9% 1171.7 2 November 1273.9 21.9% 10.6% 1144.4 2 December 1273.9 31.5% 10.7% 1114.1 3

1) CF is the ratio of wind plant output at the system peak hour to nameplate
2) Summer Capacity value for wind plant is defined as the capacity factor between the hours of 1400 and 1800 for the summer months of June, July and August. The summer 2007 value was 22.9%, the summer 2008 value was 16.7%

and summer 2009 was 14.1%. Winter Capacity value for wind plant is defined as the capacity factor between the hours of 1600 and 2000 for the winter months of Dec., Jan. and Feb. The winter value for 07-08 was 30.4%. The winter 08 -

09 value was 24.2% and the 09-10 value was 26.4%.

NYISO Wind Generation Study lJune 2010 97

Appendix B: Summary of Other Regions Experiences with Wind Generation Appendix B-1: Summary of Key Lessons from European Wind Integration Experience Europe shows that high/very high wind penetration levels are possible, but those high penetration levels are driven by energy policy (subsidies) and not economics for the most part. This also applies to power system integration issues.

Wind power can be successfully included into markets (Spain/UK).

European regulators and Transmission System Operators (TSOs) have developed a willingness to learn and question existing rules as well as to adjust rules and regulations. In addition, most European countries have shown a flexibility to adjust their energy policy, rules and regulations depending on the technical and economical development in order to create a low-risk environment for renewable energy projects, without allowing windfall profits as it is very difficult to get all relevant regulatory details right at the first attempt. This flexibility for change has been based on a continuous dialogue between policy makers, regulators, network companies and the renewable energy lobby.

Both load and generation benefit from the statistics of large numbers as they are aggregated over larger geographical areas. Larger balancing areas make wind plant aggregation possible. The forecasting accuracy improves as the geographic scope of the forecast increases; due to the decrease in correlation of wind plant output with distance, the variability of the output decreases as more plants are aggregated. On a shorter-term time scale, this translates into a reduction in reserve requirements; on a longer-term time scale, it produces some smoothing effects on the capacity value.

Larger balancing areas also give access to more balancing units.

The development of grid codes played an important role for Europe to ensure a reliable power system operation.

Improvements in wind-plant operating characteristics has enhanced reliable operation of the system through the ability to provide voltage control at a weak point in the system, the ability to provide an inertial response in a stability-constrained system, the ability to participate in providing ancillary services, and the ability to ride through faults (voltage and frequency deviations) without disconnection. Remaining issues in Europe are old wind turbines which do not meet the requirements of the grid codes and validation of turbines/simulation models that fulfill the grid codes.

Integrating wind generation information in system operation both real-time and with updated forecasts up to day-ahead will help manage the variability and forecast errors of wind power. Shortening the gate closure time in market operation practices will help integration but may require improvements in the operating tools. Well-functioning hour-ahead and day-ahead markets can help to more cost-effectively provide balancing energy required by the variable-output wind plants.

Specific wind farm control centers (Spain) combined with power system state estimators provide a powerful tool for large-scale wind integration as wind farms can be remotely adjusted (on/off/part-load/PF control), taking into account real-time conditions in the power system.

Frequency control with wind turbines has been tested in Denmark and the United Kingdom, wind turbines/farms are expected to actively participate in the frequency control task in the future.

Black-start in power systems with high wind penetration level could be problematic. Denmark is developing a new, cell-based power system architecture which will incorporate wind power in the black-start procedure.

Transmission helps to achieve benefits of aggregating large-scale wind power development and provides improved system balancing services. This is achieved by making better use of physically available transmission capacity and upgrading and expanding transmission systems. High wind penetrations may also require improvements in grid internal transmission capacity.

NYISO Wind Generation Study lJune 2010 98

Appendix B-2: Summary of the CAL-ISO Study The planned $1.8 billion of transmission upgrades for the Tehachapi area are sufficient to support up to 4,200 MW of new renewable resources.

New wind generation resources should be Type 3 or Type 4 units as the installation of more Type 1 units in Tehachapi has a negative impact on the reliability of the system.

All new generating facilities, including new wind generation facilities, must meet the California ISO Interconnection Standards, provide 4-second operating data and be prepared to act on dispatch notices from the California ISO Operations.

Integrating 20% renewables in the current generation mix is achievable; however, several market integration and operational changes are required.

Transient stability studies indicated that the new Tehachapi wind generation with Type 3 or Type 4 units, meets WECC LVRT as well as the WECC transient stability standard.

Some of the existing Tehachapi wind generation (Type 1 Units) trips off-line for three phase 500 kV faults in the local area under the full wind scenario.

Post-transient governor power flow analysis results indicate that the WECC standards are met.

A state-of-the-art wind forecasting service is necessary in the Day-Ahead time frame to minimize errors in the unit commitment process. The accuracy of Day-Ahead load and wind generation forecasts will affect the market clearing prices and unit commitment costs.

Approximately 800 MW/hr of generating capacity and ramping capability will be required to meet multi-hour ramps during the morning load increase coupled with declining wind generation. Operations will need to be able to quickly ramp down dispatchable resources during the evening load drop-off and accommodate increases in wind generation.

The amount of regulation required will significantly increase with large amounts of new wind generation.

The size of the supplemental energy stack must significantly increase to meet intra-hour load following needs.

The California ISO must have the ability to curtail wind generation during over-generation conditions.

Short start units must be available to accommodate Hour-Ahead forecast errors and intra-hour wind variations. The quantity of short start units that will be needed requires additional analysis and modeling.

Comments of the California ISO President & CEO Yakout Mansour:

The good news is that this study shows the feasibility of maintaining reliable electric service with the expected level of intermittent renewable resources associated with the current 20% RPS, provided that existing generation remains available to provide back-up generation and essential reliability services. The cautionary news is the provided part of our conclusion. Regulatory actions under active consideration threaten the economic viability of much of this essential generation. Moreover, current regulatory policies assigning high on-peak availability factors to intermittent generation will eliminate the theoretical but not the real need for the essential generation currently provided by existing power plants, and regulators may be unwilling to support sufficient forward procurement of generation. Furthermore, the model used for this study is based on the technical specifications and capabilities of the generation fleet, but does not reflect contractual or other regulatory constraints that are not known to the California ISO.

NYISO Wind Generation Study l August 2010 99

Appendix B-3: Summary of the ERCOT Study Uncertainty and variability are an inherent part of power system operations; power system infrastructure and operating practices have developed around the requirement to accommodate variability and uncertainty. Addition of wind generation capacity increases both, but does not greatly change their nature. The tools of operation used to address these attributes for load alone are expandable to address the net load resulting from wind generation partially offsetting connected system load.

An overall observation in this study is that through 5,000 MW of wind generation capacity, approximately the level of wind capacity presently in ERCOT, wind generation has limited impact on the system. Its variability barely rises above the inherent variability caused by system loads. At 10,000 MW wind generation capacity, the impacts become more noticeable.

By 15,000 MW, the operational issues posed by wind generation will become a significant focus in ERCOT system operations. However, the impacts can be addressed by existing technology and operational attention, without requiring any radical alteration of operations.

While ERCOTs present regulation procurement methodology is adequate in terms of procuring sufficient regulation service, there are improvements that can be made which are expected to reduce the amount of procurement while maintaining sufficiency. Most notable is the inclusion of wind generation forecast information. Also, adjustments are advisable to accommodate year-to-year wind generation capacity growth.

Proper use of wind generation forecasting is of critical importance to reliable and efficient operation of the system. In addition to making efficient unit commitment decisions, wind forecasts allow ancillary services procurements to be adapted to actual conditions. The risks of extreme weather events are generally very predictable, and appropriate operating decisions can be made to pre-emptively reduce their impact.

High penetration of wind generation reduces loading on thermal units while increasing the requirements for these units to provide ancillary services. Beyond ERCOTs present level of wind generation capacity, there will be infrequent periods when unit dispatch and commitment may need to be altered to provide ancillary services. Through the 15,000 MW wind generation capacity scenario investigated, these events become progressively more frequent.

NYISO Wind Generation Study l August 2010 100

Appendix B-4: Summary of the OPA Study The average capacity value of the wind resource in Ontario during the summer (peak load) months is approximately 17%.

The capacity value ranges from 38% to 42% during the winter months (November to February) and from 16% to 19%

during the summer months (June to August). Since 87% of the hits (periods within 10% of the load peak) occur during the summer months, the overall yearly capacity value is expected to be heavily weighted toward the summer. The overall yearly capacity value is approximately 20% for all wind penetration scenarios. In other words, 10,000 MW of installed nameplate wind capacity is equivalent to approximately 2,000 MW of firm generation capacity. The capacity value is generally insensitive to the wind penetration level, mainly due to good wind geographic diversity and the fact that the various wind output levels are derived by scaling the same wind groups.

The results of the regulation analysis show that, in all scenarios, the incremental regulation needed to maintain current operational performance is small. With incremental regulation requirement defined as the increase in 3 of the net-load with and without wind, the increase in regulation is only 11% with 10,000 MW of wind and 4% with 5,000 MW. This additional regulation could be handled within the current system operation framework.

Incremental load following requirements are more substantial due to increased variability in the 5-minute timeframe. The year 2009 load with 1,310 MW of wind scenario could be easily accommodated with the existing generators. The year 2020 load with 5,000 MW of wind scenario shows a 17% increase in load following requirements. It is likely that online generators could provide this incremental requirement. Beyond 5,000 MW of wind, the additional load following requirement may exceed the capability of existing generators. It is important that any future supply mix strategy recognize that wind generators will likely displace more flexible generation resources and the remaining balance-of-portfolio resources must be able to accommodate this additional variability.

The 10-minute operating reserve requirement is specifically tied to a single contingency, meaning that the reserve is meant to accommodate loss of a single unit, but not a simultaneous drop in generation and increase in load. Therefore, the 10-minute wind-alone variability was analyzed as a proxy for operating reserve requirements. The results show that with 5,000 MW of wind, the incremental operating reserve requirement is considered negligible but at higher wind penetrations, the incremental operating reserve requirement becomes more significant. The current largest contingency exposure on the Ontario bulk power system is 900 MW. For the 6,000 MW and 8,000 MW wind penetration cases, the wind output dropped by more than 900 MW in ten minutes 4 times. The wind output dropped by more than 900 MW 10 times With 10,000 MW of wind, The results indicate that an increase operating reserve requirement can be expected in order to accommodate extreme drops in wind generation for the high wind penetration scenarios.

For several of the scenarios, the minimum net-load point (with wind) is significantly reduced as compared to the minimum load-alone point. This has serious implications for the online generation resources during the low load periods and may require curtailment of wind power output or other mitigation measures. For the 10,000 MW scenario, wind energy output below the minimum load point represents 25% of the yearly energy. This is a significant proportion of the yearly energy output. If the minimum load-wind point drops far enough down into the generation stack, then only less maneuverable generation units may be left to serve the load. A complicating factor is that, during these low load-wind periods, the variability of the load-wind deltas is greater than the load-alone deltas. In other words, the maneuverability burden on the units serving the load during these periods is greater.

For all wind scenarios, the increase in hourly and multi-hourly variability, as measured by , due to wind is relatively small (not more than 10% for any scenario). From an hourly scheduling point of view, even 10,000 MW of wind would not push the envelope much further beyond the current operating point. However, the amount and magnitudes of extreme one-hour and multihour net-load changes are significantly greater with high wind penetration. With the addition of 10,000 MW of wind, the maximum one-hour net-load rise increases by 34%, and the maximum one-hour net-load drop increases by 30%.

This data indicates that with large amounts of wind, much more one-hour ramping capability is needed for secure operation. Clearly the longest sustained ramping (up and down) occurs during the summer morning load rise and evening NYISO Wind Generation Study l August 2010 101

load decline periods. During these periods, (and others) the units may need to ramp continually over three or more hours.

For the year 2020 load with 10,000 MW of wind scenario, the maximum positive three-hour load-wind delta increases by 17% and the maximum negative three-hour delta increases by 33%. The detailed results clearly illustrate the fact that units will have to undergo sustained three-hour ramping more often, and ramp further with the addition of large amounts of wind.

The analysis shows that sudden (less than 10-minute) province-wide interruptions of wind generation power output are extremely unlikely and do not represent a credible planning contingency. When sudden changes in wind output do occur, the study shows that the spatial diversity of wind sites and wind groups would tend to limit the impact of individual site or group changes on the aggregate wind output. This includes the impact of extreme weather incidents such as windstorms and ice storms, which are two of the major concerns for wind tower structural integrity. However, windstorms in the form of hurricanes or tornadoes, and ice storms which are capable of severely damaging or toppling a wind structure move at finite speeds and are not capable of sudden wholesale damage to structures across Ontario within ten minutes or less.

NYISO Wind Generation Study l August 2010 102

Retrieved from "https://kanterella.com/w/index.php?title=ML12090A820&oldid=2463274"