Northeast Power Coordinating Council, Inc. Multi-Area Probabilistic Reliability Assessment For Winter 2007/08

Transcription

1 Northeast Power Coordinating Council, Inc. MultiArea Probabilistic Reliability Assessment For Winter 2007/08 Approved by the NPCC Task Force on Coordination of Planning November 20, 2007 Conducted by the NPCC CP8 Working Group

2 This page intentionally left blank.

3 CP8 WORKING GROUP Northeast Power Coordinating Council, Inc. HydroQuébec Distribution HydroQuébec Production Independent Electricity Market Operator ISO New England, Inc. National Grid New Brunswick System Operator New York Independent System Operator New York State Reliability Council Nova Scotia Power Inc. Ontario Power Generation, Inc. Phil Fedora, Chairman Pierre Poirier Thong Nguyen Phat Greg Hine Fei Zeng Jack Martin Scott Brown Gregory Drake Al Adamson Kamala Rangaswamy Kevan Jefferies The CP8 Working Group acknowledges the efforts of Messrs. Glenn Haringa, GE Energy, and Andrew Ford, PJM and thanks them for their assistance in this analysis. CP8 Working Group November 20, Final Report

4 TABLE OF CONTENTS PAGE EXECUTIVE SUMMARY 4 Introduction... 4 Results... 4 Conclusions... 6 INTRODUCTION 7 MODEL ASSUMPTIONS 8 Load Representation...8 Load Shape... 8 Load Forecast Uncertainty... 9 Generation Unit Availability Transfer Limits Operating Procedures to Mitigate Resource Shortages Assistance Priority Modeling of Neighboring Regions WINTER 2006/07 SUMMARY ANALYSIS 18 Winter 2007/08 Results Base Case Scenario Severe Case Scenario Conclusions CP8 Working Group November 20, Final Report

5 APPENDICES PAGE A) OBJECTIVE AND SCOPE OF WORK 26 B) EXPECTED NEED FOR OPERATING PROCEDURES 27 Table 7 Base Case Assumptions (2003/04 Load Shape) Table 8 Severe Case Scenario (2003/04 Load Shape) C) MULTIAREA RELIABILITY SIMULATION PROGRAM DESCRIPTION 29 CP8 Working Group November 20, Final Report

6 EXECUTIVE SUMMARY Introduction This study estimated the use of NPCC Area Operating Procedures designed to mitigate resource shortages for the winter of 2007/08 (November 2007 through March 2008). The CP8 Working Group closely coordinated its efforts with those of the CO12 Working Group s study, "NPCC Reliability Assessment for Winter ", November General Electric s (GE) MultiArea Reliability Simulation (MARS) program was selected for the analysis. GE Energy was retained by the Working Group to conduct the simulations. Results For the November 2007 March 2008 period, Figure EX1 displays the results for the expected load level (the expected load level results were based on the probabilityweighted average of the seven load levels simulated) under the Severe Case assumptions. 2 Estimated Number of Occurrences (days/period) 1 0 NE NY ON MT Q Reduce 30min Reserve Voltage Reduction Reduce 10min Reserve Appeals Disconnect Load Maritimes Area initiates interruptible loads instead of voltage reduction Figure EX1 Potential Range of Use of Indicated Operating Procedures for Winter 2007/08 Considering Severe Case Assumptions (November 2007 March 2008) (Expected Load Level) CP8 Working Group November 20, Final Report

7 For the November 2007 March 2008 period, Figure EX2 shows the estimated use of the indicated operating procedures under the Severe Case assumptions for the extreme load level (representing the second to highest load level, having approximately a 6% chance of being exceeded). 16 Estimated Number of Occurrences (days/period) NE NY ON MT Q Reduce 30min Reserve Voltage Reduction Reduce 10min Reserve Appeals Disconnect Load Maritimes Area initiates interruptible loads instead of voltage reduction Figure EX2 Winter 2007/08 Estimated Use of the Indicated Operating Procedures Severe Case Assumptions, Extreme Load Level (November 2007 March 2008) CP8 Working Group November 20, Final Report

8 Conclusions Use of operating procedures designed to mitigate resource shortages (specifically, reducing 30minute reserve, voltage reduction, reducing 10minute reserve, and public appeals) is not expected for Québec, Ontario, New England, and New York, during the 2007/08 winter period for both the expected and extreme load level under the assumed Base and Severe Case conditions. The expected load level results were based on the probabilityweighted average of the seven load levels simulated. The extreme load level conditions represents the second to highest load level, having approximately a 6% chance of being exceeded. For the Maritimes Area there is an expectation for reducing 30minute reserves in response to a capacity deficiency this winter for the expected load level under the Base Case (approximately once) and Severe Case (approximately twice). These expectations increase approximately seven fold for the extreme load level conditions, coupled with (for the Severe Case only) an additional expectation of also needing to call on interruptible loads seven times. Again, the extreme load level conditions represents the second to highest load level, having approximately a 6% chance of being exceeded. CP8 Working Group November 20, Final Report

9 INTRODUCTION This study estimated the use of NPCC Area operating procedures to mitigate resource shortages for November 2007 through March The Working Group closely coordinated its efforts with the NPCC CO12 Working Group s study, "NPCC Reliability Assessment for Winter ", November The development of this Working Group was in response to the following recommendation from the "NPCC Reliability Assessment for Winter 2004/05", December 2004: The CO12 assessment of the Summer Operating Period is accompanied by a corresponding multi area probabilistic assessment of Loss of Load Expectations and of the projected use of Operating Procedures designed to mitigate resource shortages. This assessment was not performed for this Winter Operating Period. For completeness in the assessment of the Winter Operating Period, the CO12 Working Group recommends that TFCO and TFCP review the merits of having this assessment performed for future Winter Operating Periods. The database developed by the previous CP8 Working Group's "NPCC Summer 2007 MultiArea Probabilistic Reliability Assessment", April 20, 2007, and the NPCC Review of Interconnection Assistance Reliability Benefits, November 2007 was used as the starting point for this analysis. Working Group members reviewed the existing data and made revisions to reflect the conditions expected for the winter 2007/08 assessment period. This report is organized in the following manner: after a brief Introduction, specific Model Assumptions are presented, followed by an Analysis of the results based on the scenarios simulated. The Working Group's Objective and Scope of Work is shown in Appendix A. Tables presenting the corresponding results for the Base Case and Severe Case simulations are listed in Appendix B. Appendix C provides an overview of General Electric's MultiArea Reliability Simulation (MARS) Program. CP8 Working Group November 20, Final Report

10 MODEL ASSUMPTIONS Load Representation The loads for each Area were modeled on an hourly, chronological basis. The MARS program modified the input hourly loads through time to meet each Area's specified annual or monthly peaks and energies. Table 1 summarizes each NPCC Area's winter peak load assumptions for the winter 2007/08. Table 1 Assumed NPCC 2007/08 Peak Loads MW (2003/04 Load Shapes) Area Expected Peak 2003/04 Load Shape Extreme Peak * Month Québec (Q) 36,079 38,063 January Maritimes Area** (MT) 5,667 6,234 January New England (NE) 23,070 24,262 January New York *** (NY) 28,791 29,684 January Ontario **** (ON) 24,123 24,994 December * Extreme Peak based on load forecast uncertainty for peak month. ** Maritimes Area represents New Brunswick, Nova Scotia, Prince Edward Island, and the area administrated by the Northern Maine Independent System Administrator (NMISA). *** More conservative peak load forecast than shown in the NYISO 2007 Load & Capacity Report. **** Represents the IESO monthly weathernormalized peak load forecast. Load Shape The Working Group used one load shape assumption for this analysis. The 2003/04 load shape represents a weather pattern that includes a consecutive period of cold days. The growth rate in each month s peak was used to escalate Area loads to match the Area's winter demand and energy forecasts for both load shapes. The impacts of DemandSide Management programs were included in each Area's load forecast for both load shapes. Figures 1 shows the diversity in the NPCC area load shapes used in this analysis for the 2003/04 load shape assumptions. CP8 Working Group November 20, Final Report

11 40, /08 Projected Coincident Monthly Peak Loads MW 2003/04 Load Shape 35,000 30,000 25,000 MW 20,000 15,000 10,000 5,000 0 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Q MT NE NY ON Figure /08 Projected Monthly Expected Peak Loads for NPCC Areas 2003/04 Load Shape Load Forecast Uncertainty Peak load forecast uncertainty was also modeled. The effects on reliability of uncertainties in the peak load forecast, due to weather and/or economic conditions, were captured through the load forecast uncertainty model in MARS. The program computes the reliability indices at each of the specified load levels (for this study, seven load levels were modeled) and calculates weightedaverage values based on input probabilities of occurrence. While the per unit variations in the load can vary on a monthly basis, Table 2 shows the values assumed for January Table 2 also shows the probability of occurrence assumed for each of the seven load levels modeled. In computing the reliability indices, all of the Areas were evaluated simultaneously at the corresponding load level, the assumption being that the factors giving rise to the uncertainty affect all of the Areas at the same time. The amount of the effect can vary according to the variations in the load levels. For this study, reliability measures are reported for two load conditions: expected and extreme. The values for the expected load conditions are derived from computing the reliability at each of the seven load levels, and computing a weightedaverage expected value based on the specified probabilities of occurrence. The indices for the extreme load conditions provide a measure of the reliability in the event of higher than expected loads, and were computed for the secondtohighest load level. These values are highlighted in Table 2. CP8 Working Group November 20, Final Report

12 Area NPCC WINTER 2007/08 Table 2 Per Unit Variation in Load Assumed for the Month of January 2008 PerUnit Variation in Load Q MT NE NY ON Prob Generation Tables 3 (a) and 3 (b) summarize the winter 2007/08 capacity assumptions for the NPCC Areas used in the analysis for the Base Case and the Severe Case Scenario, respectively. Base Case conditions were chosen to be consistent with the assumptions used in the companion report by the NPCC CO12 Working Group, "NPCC Reliability Assessment for Winter ", November Table 3 (a) NPCC Capacity and Load Assumptions for January 2008 MW Base Case Expected Load Q MT NE NY ON 1 Assumed Capacity 33,113 6,871 33,600 41,089 30,414 (Includes indicated Maximum Wind output Purchase/Sale 6, ,792 0 Peak Load 2 36,079 5,667 23,070 28,791 24,112 Scheduled Maintenance , Capacity shown for Ontario has been seasonally adjusted. 2 Based on 2003/04 Load Shape assumption. 3 Maintenance shown is for the week of the monthly peak load. CP8 Working Group November 20, Final Report

13 Assumed Capacity (Includes indicated Maximum Wind output Table 3 (b) NPCC Capacity and Load Assumptions for January 2008 MW Severe Assumptions Scenario Extreme Load Q MT NE NY ON 1 32,633 6,412 28,457 41,034 29, Purchase/Sale 6, ,792 0 Peak Load 2 38,063 6,234 24,262 29,684 24,662 Scheduled Maintenance , Unit Availability Details regarding the NPCC Area s assumptions for generator unit availability are described in the respective Area s most recent "NPCC Triennial Review of Resource Adequacy" 4 In addition, the following Areas provided the following: Québec The planned outages for the winter period are reflected in this assessment. The volume of planned outages is consistent with historical volumes. Ontario Ontario s generating unit availability was modeled as described in the September 10, Month Outlook: An Assessment of the Reliability of the Ontario Electricity System From October 2007 to March 2009." 5 Ontario market participants provided the majority of generation data. F.O.R. and P.O.R. were based on forecast values for generating units, which reflect past experience and future expectations based on recent maintenance activities. However, for some of the generating units F.O.R. and P.O.R. values were based on North American Reliability Council (NERC) Generator Availability Data System 6 (GADs) data for similar type units. New England This probabilistic assessment reflects New England generating unit availability assumptions 7 based upon historical performance over the prior fiveyear period. Unit availability modeled reflects the projected scheduled maintenance and forced outages. Individual generating unit maintenance assumptions are based upon each unit s historical 4 See: 5 See: 6 See: 7 See: CP8 Working Group November 20, Final Report

14 fiveyear average of scheduled maintenance. Individual generating unit forced outage assumptions were based on the unit s historical data and North American Reliability Council (NERC) average data for the same class of unit. In the previous years reliability assessments, generating unit Equivalent Forced Outage Rate (EFOR) were used to represent the forced unavailability of the generating units. In this assessment, Equivalent Forced Outage Rate Demand (EFORd) was used to represent the forced unavailability of the units. A description of the EFOR and EFORd parameters and their differences can be found at the ISO New England Web site. 8 New York Detailed availability assumptions used for the New York units can be found in the New York ISO February 16, 2007 report 9 "Locational Installed Capacity Requirements Study covering the New York Control Area for the Capability Year" and the New York Control Area Installed Capacity Requirement for the Period May 2007 April 2008 New York State Reliability Council, January 5, 2007 report. 10 Transfer Limits Figure 2 depicts the system that was represented in this Assessment, showing Area and assumed Base Case transfer limits for the winter 2007/08 period. New York Area internal transmission representation was consistent with the assumptions used in the New York ISO February 16, 2007 report 9 "Locational Installed Capacity Requirements Study covering the New York Control Area for the Capability Year" and the New York Control Area Installed Capacity Requirement for the Period May 2007 April 2008 New York State Reliability Council, January 5, 2007 report. 10 New England internal transmission representation was consistent with the assumptions detailed in the 2007 Regional System Plan, ISONE, October 18, See: 9 See: 10 See: 11 See: CP8 Working Group November 20, Final Report

15 MANIT Total Ontario 4,000 In 5,550 OutS 5,900 OutW (all + 1,250 4/09) MRO US 7, S 275 W 90 1,550 S 1,800 W 1, S 343 W West RFC OTH NW Niagara 2,150 S 2,400 W NE Ontario 1,300 S 1,650 W 3,000 Ottawa East RFC 7,500 1,300 S 1,950 W A 550 3, S 84 W 5, , S 110 W C 2, ND 1,998 (4/09) 147 S 1,397 S (4/09) 167 W 800 1,417 W (4/09) 420 S 470 W PJM SW 800 West 200 7, ,015 5, Cent 1,200 S 0 W 1,800 Mtl 300 1,000 1,500 D PJMRTO JB New York 2,000 6,500 8,400 (06/09) 7,500 8,400 (06/09) Chur. Quebec G 1, J 5,200 11, ,500 0 ( 11) 0 Que Cent. 485 S 685 S ( 11) 520 W 17, W ( 11) 100 East NY , (5/08) MAN 660 F K NBM Maritimes NY PEI ( 11) 1,000 S 1,100 W VT 1,200 S 1,525 S 1,325 W 1,600 W Total NYNE (Excludes CSC) NOR , ( 10) 800 NS NB CMA (12/07) NM 700 1,000 ( 08) BHE W MA CT New England Northport to Norwalk Harbor (1385 line K to NOR) transmission capability due to temporary reconductoring project Figure 2 Assumed Transfer Limits Between Areas Tie transfer limits between Areas are indicated in Figure 2 with seasonal ratings (Ssummer, W winter) where appropriate. The acronyms and notes used in Figure 2 are defined as follows: Chur Churchill Falls NOR Norwalk Stamford RFC ReliabilityFirst Corp. MANIT Manitoba BHE Bangor Hydro Electric NB New Brunswick ND NicoletDes Cantons Mtl Montréal PEI Prince Edward Island BJ Bay James C MA Central MA CT Connecticut MAN Manicouagan W MA Western MA NS Nova Scotia 12 Phase angle regulators (PARs) are installed on the Michigan Ontario interconnection but are not available to regulate flows except in emergencies, pending agreement by the International Transmission Company in Michigan to permit full regulation. Before the summer of 2006, the IESO, the Midwest ISO, Hydro One and International Transmission Company, agreed to temporarily bypass the phase angle regulators for normal operation until an agreement is reached to make full use of their regulating capability. Bypassing the PARs increases Ontario s transfer capability to and from Michigan by 300 to 350 MW in the summer and by about 400 MW in the winter. 13 If necessary, transmission congestion between generation in New Brunswick and load in Northern Maine can be overcome by switching additional load in Northern Maine to be radially served from the New Brunswick grid. 12, 13 CP8 Working Group revised November 27, Final Report

16 NE Northeast (Ontario) NBM Millbank NW Northwest (Ontario) MRO Midwest Reliability VT Vermont RFC ReliabilityFirst Corp. NM Organization Que Québec Centre CSC Cross Sound Cable Northern Maine Centre Operating Procedures to Mitigate Resource Shortages Each Control Area takes defined steps as their reserve levels approach critical levels. These steps consist of those load control and generation supplements that can be implemented before load has to be actually disconnected. Load control measures could include disconnecting interruptible loads, public appeals to reduce demand, and voltage reductions. Other measures could include calling on generation available under emergency conditions, and/or reduced operating reserves. The need for an Area to begin these operating procedures is modeled in MARS by evaluating the daily Loss of Load Expectation (LOLE) at specified margin states. The user specifies these margin states for each area in terms of the benefits realized from each emergency measure, which can be expressed in MW, as a per unit of the original or modified load, and as a per unit of the available capacity for the hour. Table 4 summarizes the load relief assumptions modeled for each NPCC Area. The Working Group recognizes that Areas may invoke these actions in any order, depending on the situation faced at the time; however, it was agreed that modeling the actions as in the order indicated in Table 4 was a reasonable approximation for this analysis. Table 4 NPCC Operating Procedures to Mitigate Resource Shortages 2007/08 Winter Load Relief Assumptions MW Actions Q MT NE NY ON 1. Curtail Load / Utility Surplus Appeals LRP/SCR/EDRP * Manual Voltage Reduction % of load 188 1% of load 2. No 30min Reserves Voltage Reduction or Interruptible Loads ** ELRP * 4. No 10min Reserves % of load 1, % of load 222 General Public Appeals 5. EDRP * No 10min Reserves 1, % of load * Since customer participation in these programs varies over time, it is recognized that the actual amount of resources available for this winter may be different than the amount assumed in this study. The values modeled in this study represent a reasonable approximation for this analysis. ** Interruptible Loads for Maritimes Area (implemented only for the Area), Voltage Reduction for all others CP8 Working Group revised November 27, Final Report

17 Assistance Priority Table 5 indicates the priority order followed when allocating reserves and assistance to Control Areas with a resource deficiency. Areas listed with equal priority received assistance on a shared basis in proportion to their deficiency. In this analysis, each step was initiated simultaneously in all Areas and subareas. It was assumed that RFC assists everyone with equal priority. Table 5 Priority Order for Providing Area Emergency Assistance Priority of Assistance Area Providing Assistance 1 ST 2 ND Québec MT ON NE NY Maritimes Area Q ON NE NY New England NY Q MT ON New York NE Q MT ON Ontario Q MT NE NY Millbank Units Q MT PJM RFC MRO NE NY PJM RFC Modeling of Neighboring Regions For the scenarios studied, a detailed representation of the neighboring regions of RFC (ReliabilityFirst Corp.) and the MROUS (Midwest Reliability Organization US portion) was assumed. The assumptions are summarized in Table 6. CP8 Working Group November 20, Final Report

18 Table 6 RFC and MRO 2007/08 Assumptions 14 PJM RFC Other MROUS Peak Load (MW) 114,047 80,594 26,107 Peak Month January January January Assumed Capacity (MW) 168, ,138 37,754 Purchase/Sale (MW) 1, Reserve (%) Weighted Unit Availability (%) Operating Reserves (MW) 3,400 2,206 1,700 Curtailable Load (MW) 2,144 2,000 1,666 No 30min Reserves (MW) 2,099 1,470 1,200 Voltage Reduction (MW) 2, ,100 No 10min Reserves (MW) 1, Appeals (MW) , , / /08 Projected Coincident Monthly Peak Loads MW 2002/ /04 Load Shape MW 120, , , , ,000 80,000 80,000 60,000 60,000 40,000 40,000 20,000 0 Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun NPCC PJM RFCOTH MRO Figure /08 Projected Monthly Expected Peak Loads for NPCC, RFC, PJM and the MRO (2003/04 Load Shape) 14 Load and capacity assumptions based on NERC s Electricity, Supply and Demand Database (ES&D) available at: CP8 Working Group November 20, Final Report

19 Winter 2006/07 Summary According to scientists at the NOAA National Climatic Data Center 15 the December 2006February 2007 winter season was marked by periods of unusually warm and cold conditions in the United States (U.S.), but the overall seasonal temperature was near average. The winter season got off to a late start in much of the U.S. December was the 11th warmest such month on record, and springlike temperatures covered much of the eastern half of the US during the first half of January. The warmerthanaverage seasonal temperatures in the more heavily populated regions of the Midwest and East helped reduce residential energy needs for the U.S. as a whole for the winter season. Using NOAA's Residential Energy Demand Temperature Index (REDTI), it was determined that the nation's residential energy demand was approximately three percent lower than what would have occurred under average climate conditions for the season. Seasonal energy demand would have been lower if not for February's colder temperatures. For the month, temperaturerelated residential energy demand was approximately six percent higher than what would have occurred under average climate conditions for February. February was among the third coldest in the 113year record for the contiguous U.S. Thirtysix states in the eastern twothirds of the U.S. were cooler than average. A complex, widereaching winter storm moved from the MidMississippi Valley into the MidAtlantic and New England February 14 and 15, This storm ranked as a Category 3 event on the Northeast Snowfall Impact Scale (NESIS). The heaviest snow fell in interior regions of the Northeast where amounts over 20 inches were widespread. This event was preceded by a 10day lake effect storm that dumped more than 100 inches of snow on New York's Tug Hill Plateau. A total of 141 inches was reported at Redfield in Oswego County. During the Winter period, only ISONE instituted OP4 Action 6 (begin to allow depletion of 30 minute reserve) on Saturday, February 10, 2007 at 9:40 am 11:00 am. The previous analysis of the winter period by the CP8 Working Group 16 estimated less than one occurrence of the use of these operating procedures for the NPCC areas during the winter period for the expected load level under the Base Case and Severe Cases conditions. 15 See : 16 See: CP8 Working Group November 20, Final Report

20 ANALYSIS Winter 2007/08 Results Base Case Scenario Table 7 (see Appendix B) shows the estimated need for the indicated operating procedures (in days/period) for November 2007 through March 2008 period for the Base Case assumptions for all NPCC Areas for the 2003/04 load shape assumptions. Only the Maritimes Area is expected to need to use these procedures in response to a capacity deficiency for this Scenario. Figure 4 shows the indicated operating procedures occurrences for the NPCC Areas for the expected load (the expected load level results were based on the probabilityweighted average of the seven load levels simulated) for the Base Case assumptions. Figure 5 shows the corresponding results for the extreme load (representing the second to highest load level, having approximately a 6% chance of being exceeded). 2 Reduce 30min Reserve Voltage Reduction Reduce 10min Reserve Appeals Disconnect Load Estimated Number of Occurrences (days/period) 1 0 NE NY ON MT Q Maritimes Area initiates interruptible loads instead of voltage reduction Figure 4 Winter 2007/08 Estimated Use of the Indicated Operating Procedures Base Case Assumptions, Expected Load Level (November 2007 March 2008) CP8 Working Group November 20, Final Report

21 10 Estimated Number of Occurrences (days/period) NE NY ON MT Q Reduce 30min Reserve Voltage Reduction Reduce 10min Reserve Appeals Disconnect Load Maritimes Area initiates interruptible loads instead of voltage reduction Figure 5 Winter 2007/08 Estimated Use of the Indicated Operating Procedures Base Case Assumptions, Extreme Load Level (November 2007 March 2008) The following summary of Base Case assumptions represents system conditions consistent with those assumed in the NPCC CO12 Working Group's "Reliability Assessment for Winter ", November The Base Case assumptions are summarized below: NPCC System AsIs System for the 2007/08 period Transfers allowed between Areas Northeast Reliability Interconnect Project (a/k/a Second New England New Brunswick Tie Line) complete in December (December 1, 2007 inservice date) No imports from Manitoba or Minnesota Load Shape adjusted to Area s year 2007/08 forecast expected and extreme assumptions 17 Ontario Based on IESO 18Month Outlook: An Assessment of the Reliability of the Ontario Electricity System (dated September 10, 2007). Maritimes No fuel disruptions assumed 17 The 2003/04 load shape represents a weather pattern that includes a consecutive period of cold days. CP8 Working Group November 20, Final Report

22 New England ~98 MW capacity additions by December, 2007 New York All Long Island cables inservice 85 MW inservice since the previous winter, including: 30 MW of planned uprate for Gilboa unit 2 55 MW of new units upstate (Plattsburgh Wind Farm) 668 MW of retirements Upstate ~ 952 MW of load reduction due to EDRP & SCR (winter value) Québec Three Eastmain 1 units expected to be inservice (480 MW) Mercier (52 MW), Peribonka (113 MW), and Gaspésie wind farm (100.5 MW) inservice PJMRTO AsIs System for the 2007/08 period Consistent with PJM s 2007 Reserve Requirements Study Report Load Shapes adjusted to the 2007/08 forecast provided by PJM Load forecast uncertainty of % +/ 5.5%, 11%, and 16.5% Operating Reserve 3,400 MW (30min. 2,765 MW; 10min. 634 MW) RFC Other AsIs System for the 2007/08 period Load Shapes adjusted to the 2007/08 forecast provided by PJM Load forecast uncertainty of 91.83% +/ 7%, 14%, and 21% Operating Reserve 2,206 MW (30min. 1,470 MW; 10min. 736 MW) MROUS AsIs System for the 2007/08 period Load Shapes adjusted to the 2006/07 forecast provided by PJM Load forecast uncertainty of 92.75% +/ 6.2 %, 12.5 %, and 18.7% Operating Reserve 1,700 MW (30min. 1,200 MW; 10min. 500 MW) Québec Load management programs are discounted to reflect operational constraints such as number of interruptions per day, week or year. Hydroelectric unit capacity is adjusted to reflect the impact of reservoir level on the head available. Unit capacity is also discounted to take into account operational restrictions such as ice cover formation, runof theriver conditions. 18 See: CP8 Working Group November 20, Final Report

23 New York The Base Case assumes that the New York City and Long Island localities will meet their locational installed capacity requirements as described in the February 16, 2007 New York ISO report "Locational Installed Capacity Requirements Study covering the New York Control Area for the Capability Year" 9 and New York State will meet the capacity requirements described in the New York Control Area Installed Capacity Requirement for the Period May 2007 April 2008 New York State Reliability Council, January 5, 2007 report. 10 The New York unit ratings were obtained from the 2007 Load & Capacity Data of the NYISO (Gold Book 19 ). All inservice New York generation resources were modeled. Special Case Resources and Emergency Demand Response Programs Special Case Resources (SCRs) are loads capable of being interrupted on demand, and distributed generators, rated 100 kw or higher, that are not directly telemetered. SCRs are ICAP resources that only provide energy/load curtailment when activated in accordance with the NYISO Emergency Operating Manual. The Emergency Demand Response Program (EDRP) is a separate program that allows registered interruptible loads and standby generators to participate on a voluntary basis, and be paid for their ability to restore operating reserves. For January, the New York ISO recommended that the SCR programs be modeled as a MW operating procedure step; the EDRPs were modeled as a MW operating procedure with a limit of five calls per month. The amounts modeled in other months vary proportionally to the monthly peak load. Since customer participation in these programs varies over time, it is recognized that the actual amount of SCR/EDRP resources available for this winter may be different than the amount assumed in this study. The New York ISO believes the values modeled in this study represent a reasonable approximation for this analysis. New England The New England generating unit s ratings were consistent with those published in the NEPOOL Forecast Report of Capacity, Energy, Loads and Transmission (CELT Report) April All inservice New England generation resources were modeled. Demand Response Program (DRP) It is anticipated that the New England DRP 21 program will provide additional load relief utilizing market mechanisms in the New England System. As of October 1, 2007 there 19 See: Book_PUBLIC.pdf 20 See: 21 See page 21: CP8 Working Group November 20, Final Report

24 were 2,213 assets registered in the Demand Response program representing approximately 1,417 MW of possible load relief. For this study, ISONE has assumed a total of 909 MW under the ISONE Load Response Program. As NEPOOL Participants continue to sign up additional resources under the DRP, it is recognized that the actual amount of DRP resources may be different than the amount assumed in this study. ISONE believes the value modeled in this study represents a reasonable approximation for this analysis. Southwest Connecticut Emergency Capability RFP On December 1, 2003, ISONE issued a Request for Proposals 22 (RFP) for up to 300 MW or more emergency supplemental capacity to meet critical nearterm electric system reliability needs in southwestern Connecticut (SWCT). For this study, ISONE has modeled the response to the RFP as 250 MW of SWCT load reduction (included in the 909 MW total). Ontario For the purposes of this study, the Base Case assumptions for Ontario are consistent with the IESO 18Month Outlook: An Assessment of the Reliability of the Ontario Electricity System (dated September 10, 2007, available from the IESO web site). 5 All inservice Ontario generation resources were modeled. The following resource additions are also forecast to come into service during the study period. Project Zone Fuel Type Capacity Estimated Effective Date Ripley Wind Power Project Southwest Wind Q4 Nuclear Upgrade N/A Uranium Q4 Great Northern TriGen West Gas Q1 Retirement of Lower Sturgeon Northeast Water Q1 (25 Hz gen) Lac Seul Project Northwest Water Q1 Price sensitive demand response is assumed to be 527 MW during the study period. Severe Case Scenario Table 8 (see Appendix B) shows the estimated need for the indicated operating procedures (in days/period) during November 2007 through March 2008 period for the Severe Case Scenario for all NPCC Areas for the 2003/04 load shape assumptions, respectively. Only the Maritimes Area is expected to need to use these procedures in response to a capacity deficiency for this Scenario. Figure 6 shows the indicated operating procedures occurrences for the NPCC Areas for the expected load (the expected load level results were based on the probabilityweighted average of the seven load levels simulated) for the Severe Case assumptions. Figure 7 22 See: CP8 Working Group November 20, Final Report

25 shows the corresponding results for the extreme load (representing the second to highest load level, having approximately a 6% chance of being exceeded). 2 Reduce 30min Reserve Voltage Reduction Estimated Number of Occurrences (days/period) 1 0 NE NY ON MT Q Reduce 10min Reserve Appeals Disconnect Load Maritimes Area initiates interruptible loads instead of voltage reduction Figure 6 Winter 2007/08 Estimated Use of the Indicated Operating Procedures Severe Case Assumptions, Expected Load Level (November 2007 March 2008) 16 Estimated Number of Occurrences (days/period) NE NY ON MT Q Reduce 30min Reserve Voltage Reduction Reduce 10min Reserve Appeals Disconnect Load Maritimes Area initiates interruptible loads instead of voltage reduction Figure 7 Winter 2007/08 Estimated Use of the Indicated Operating Procedures Severe Case Assumptions, Extreme Load Level (November 2007 March 2008) CP8 Working Group November 20, Final Report

26 The Severe Case Scenario assumptions are summarized below: NPCC System AsIs System for the 2007/08 period Transfers allowed between Areas Transfer capability between Ontario and Michigan reduced by 50% Northeast Reliability Interconnect Project delayed (February 1, 2008 inservice date) No imports from Manitoba or Minnesota 2003/04 Load Shape adjusted to Area s year 2007/08 forecast expected and extreme assumptions Ontario Planned outage returns delayed by 6 weeks (~ 750 MW reduction) Additional dispatchable demand participation does not materialize for winter peak 10% lower hydroelectric capacity and energy assumed Maritimes Natural gas units (without dualfuel capability) unavailable during the months of January and February (~ 459 MW) New England ~98 MW capacity additions delayed Scheduled maintenance overrun by 4 weeks GasFired only capacity not having firm pipeline transportation contracts assumed unavailable (~ 5,000 MW) New York Planned additions delayed (55 MW Up State) Québec Eastmain 1 hydraulic station in outage delayed until end of February 2008 (480 MW) PJMRTO GasFired only capacity not having firm pipeline transportation, assumed unavailable ~ 4,200 MW Increased load forecast uncertainty of % +/ 7.0%, 13%, and 19.0% Ice Storm; ice blocking fuel delivery to all units. ~ 8,400 MW. CP8 Working Group November 20, Final Report

27 Conclusions Use of operating procedures designed to mitigate resource shortages (specifically, reducing 30minute reserve, voltage reduction, reducing 10minute reserve, and public appeals) is not expected for Québec, Ontario, New England, and New York, during the 2007/08 winter period for both the expected and extreme load level under the assumed Base and Severe Case conditions. The expected load level results were based on the probabilityweighted average of the seven load levels simulated. The extreme load level conditions represents the second to highest load level, having approximately a 6% chance of being exceeded. For the Maritimes Area there is an expectation for reducing 30minute reserves in response to a capacity deficiency this winter for the expected load level under the Base Case (approximately once) and Severe Case (approximately twice). These expectations increase approximately seven fold for the extreme load level conditions, coupled with (for the Severe Case only) an additional expectation of also needing to call on interruptible loads seven times. Again, the extreme load level conditions represents the second to highest load level, having approximately a 6% chance of being exceeded. CP8 Working Group November 20, Final Report

28 APPENDIX A Objective and Scope of Work 1. Objective Using the G.E. MultiArea Reliability Simulation (MARS) program, review NPCC Area reliability resulting from the anticipated resource and transmission capacity reported for the winter period under Base Case and Severe Case assumptions, and summarize the range of results for the winter and shoulder season months (the period from November 2007 to March 2008). 2. Scope In meeting this objective, the CP8 Working Group will review the shortterm resource adequacy of NPCC and neighboring regions for the 2007 and 2008 winter period, recognizing uncertainty in forecasted demand, scheduled outages of transmission, forced and scheduled outages of generation facilities, including fuel supply disruptions, and the impact of proposed load response programs. Reliability will be measured by calculating the Loss of Load Expectation (LOLE) and estimated use of Area operating procedures used to mitigate resource shortages. A report summarizing the results of the assessment will be published no later than November 30, The assessment will: 1. Review last winter s CP8 Working Group Winter assessment with respect to actual NPCC Area s experience; 2. Consider the impacts of SubArea transmission constraints; 3. Incorporate, to the extent possible, a detailed GE MARS reliability representation for the regions bordering NPCC; 4. Coordinate assessment assumptions with the NPCC Task Force on Coordination of Operations (CO12 Working Group); and, 5. Examine any impact of evolving market rules on overall NPCC interconnection assistance and other assumptions. CP8 Working Group November 20, Final Report

29 Base Case NPCC WINTER 2007/08 APPENDIX B Table 7 Base Case Assumptions (2003/04 Load Shape Assumption) Expected Need for Indicated Operating Procedures (days/period) (Occurrences 0.5 or greater are highlighted) Québec Maritimes Area New England New York Ontario 30min VR 10min Appeal 30min VR 10min Appeal 30min VR 10min Appeal Disc 30min VR Appeal 10min Disc 30min VR 10min Appeal /Disc /Disc /Disc 03/04 Load ShapeExpected Load Nov Dec Jan Feb Mar NovMar /04 Load ShapeExtreme Load Nov Dec Jan Feb Mar NovMar Notes: "30min" reduce 30minute Reserve Requirement; "VR" and initiate Voltage Reduction (Interruptible Loads for the Maritimes Area); "10min" and reduce 10minute Reserve Requirement; "Appeal" and initiate General Public Appeals; "Disc" and disconnect customer load. CP8 Working Group November 20, Final Report

30 APPENDIX B Table 8 Severe Case Scenario (2003/04 Load Shape Assumption) Expected Need for Indicated Operating Procedures (days/period) (Occurrences 0.5 or greater are highlighted) Severe Case Results Québec Maritimes Area New England New York Ontario 30min VR 10min Apl Disc 30min VR 10 min Apl Disc 30min VR 10min Apl Disc 30min VR Apl 10min Disc 30min VR 10min Apl Disc 03/04 Load ShapeExpected Load Nov Dec Jan Feb Mar NovMar /04 Load ShapeExtreme Load Nov Dec Jan Feb Mar NovMar Notes: "30min" reduce 30minute Reserve Requirement; "VR" and initiate Voltage Reduction (Interruptible Loads for the Maritimes Area); "10min" and reduce 10minute Reserve Requirement; "Apl" and initiate General Public Appeals; "Disc" and disconnect customer load. CP8 Working Group November 20, Final Report

31 APPENDIX C MultiArea Reliability Simulation Program Description General Electric s MultiArea Reliability Simulation (MARS) program 23 allows assessment of the reliability of a generation system comprised of any number of interconnected areas. Modeling Technique A sequential Monte Carlo simulation forms the basis for MARS. The Monte Carlo method allows for many different types of generation and demandside options. In the sequential Monte Carlo simulation, chronological system histories are developed by combining randomly generated operating histories of the generating units with the interarea transfer limits and the hourly chronological loads. Consequently, the system can be modeled in great detail with accurate recognition of random events, such as equipment failures, as well as deterministic rules and policies that govern system operation. Reliability Indices The following reliability indices are available on both an isolated (zero ties between areas) and interconnected (using the input tie ratings between areas) basis:. Daily Loss of Load Expectation (LOLE days/year). Hourly LOLE (hours/year). Loss of Energy Expectation (LOEE MWh/year). Frequency of outage (outages/year). Duration of outage (hours/outage). Need for initiating Operating Procedures (days/year or days/period) The Working Group used both the daily LOLE and Operating Procedure indices for this analysis. The use of Monte Carlo simulation allows for the calculation of probability distributions, in addition to expected values, for all of the reliability indices. These values can be calculated both with and without load forecast uncertainty. The MARS program probabilistically models uncertainty in forecast load and generator unit availability. The program calculates expected values of Loss of Load Expectation (LOLE) and can estimate each Area's expected exposure to their Emergency Operating Procedures. Scenario analysis is used to study the impacts of extreme weather conditions, variations in expected unit inservice dates, overruns in planned scheduled maintenance, or transmission limitations. Resource Allocation Among Areas The first step in calculating the reliability indices is to compute the area margins on an isolated basis, for each hour. This is done by subtracting from the total available capacity in the area for the hour the load demand for the hour. If an area has a positive or zero margin, then it has sufficient capacity to meet its load. If the area margin is negative, the load exceeds the capacity available to serve it, and the area is in a lossofload situation. 23 See: CP8 Working Group November 20, Final Report

32 APPENDIX C If there are any areas that have a negative margin after the isolated area margins have been adjusted for curtailable contracts, the program will attempt to satisfy those deficiencies with capacity from areas that have positive margins. Two methods are available for determining how the reserves from areas with excess capacity are allocated among the areas that are deficient. In the first approach, the user specifies the order in which an area with excess resources provides assistance to areas that are deficient. The second method shares the available excess reserves among the deficient areas in proportion to the size of their shortfalls. The user can also specify that areas within a pool will have priority over outside areas. In this case, an area must assist all deficient areas within the same pool, regardless of the order of areas in the priority list, before assisting areas outside of the pool. Poolsharing agreements can also be modeled in which pools provide assistance to other pools according to a specified order. Generation MARS has the capability to model the following different types of resources:. Thermal. Energylimited. Cogeneration. Energystorage. Demandside management An energylimited unit can be modeled stochastically as a thermal unit with an energy probability distribution (Type 1 energylimited unit), or deterministically as a load modifier (Type 2 energylimited unit). Cogeneration units are modeled as thermal units with an associated hourly load demand. Energystorage and demandside management impacts are modeled as load modifiers. For each unit modeled, the installation and retirement dates and planned maintenance requirements are specified. Other data such as maximum rating, available capacity states, state transition rates, and net modification of the hourly loads are input depending on the unit type. The planned outages for all types of units in MARS can be specified by the user or automatically scheduled by the program on a weekly basis. The program schedules planned maintenance to levelize reserves on either an area, pool, or system basis. MARS also has the option of reading a maintenance schedule developed by a previous run and modifying it as specified by the user through any of the maintenance input data. This schedule can then be saved for use by subsequent runs. Thermal Units In addition to the data described previously, thermal units (including Type 1 energylimited units and cogeneration) require data describing the available capacity states in which the unit can operate. This is input by specifying the maximum rating of each unit and the rating of each capacity state as a per unit of the unit's maximum rating. A maximum of eleven capacity states are allowed for each unit, representing decreasing amounts of available capacity as governed by the outages of various unit components. CP8 Working Group November 20, Final Report

33 APPENDIX C Because MARS is based on a sequential Monte Carlo simulation, it uses state transition rates, rather than state probabilities, to describe the random forced outages of the thermal units. State probabilities give the probability of a unit being in a given capacity state at any particular time, and can be used if you assume that the unit's capacity state for a given hour is independent of its state at any other hour. Sequential Monte Carlo simulation recognizes the fact that a unit's capacity state in a given hour is dependent on its state in previous hours and influences its state in future hours. It thus requires the additional information that is contained in the transition rate data. For each unit, a transition rate matrix is input that shows the transition rates to go from each capacity state to each other capacity state. The transition rate from state A to state B is defined as the number of transitions from A to B per unit of time in state A: TR (A to B) = Number of Transitions from A to B Total Time in State A If detailed transition rate data for the units is not available, MARS can approximate the transition rates from the partial forced outage rates and an assumed number of transitions between pairs of capacity states. Transition rates calculated in this manner will give accurate results for LOLE and LOEE, but it is important to remember that the assumed number of transitions between states will have an impact on the timecorrelated indices such as frequency and duration. EnergyLimited Units Type 1 energylimited units are modeled as thermal units whose capacity is limited on a random basis for reasons other than the forced outages on the unit. This unit type can be used to model a thermal unit whose operation may be restricted due to the unavailability of fuel, or a hydro unit with limited water availability. It can also be used to model technologies such as wind or solar; the capacity may be available but the energy output is limited by weather conditions. Type 2 energylimited units are modeled as deterministic load modifiers. They are typically used to model conventional hydro units for which the available water is assumed to be known with little or no uncertainty. This type can also be used to model certain types of contracts. A Type 2 energylimited unit is described by specifying a maximum rating, a minimum rating, and a monthly available energy. This data can be changed on a monthly basis. The unit is scheduled on a monthly basis with the unit's minimum rating dispatched for all of the hours in the month. The remaining capacity and energy can be scheduled in one of two ways. In the first method, it is scheduled deterministically so as to reduce the peak loads as much as possible. In the second approach, the peakshaving portion of the unit is scheduled only in those hours in which the available thermal capacity is not sufficient to meet the load; if there is sufficient thermal capacity, the energy of the Type 2 energylimited units will be saved for use in some future hour when it is needed. CP8 Working Group November 20, Final Report