Essential Reliability Services Task Force (ERSTF)

Transcription

1 Essential Reliability Services Task Force (ERSTF) Webinar for OC and PC Members on ERS measures August 27, 2015

2 ERSTF Introduction Brian Evans-Mongeon, ERSTF Co-Chair Todd Lucas, ERSTF Co-Chair ERSTF Webinar August 27, 2015

3 ERSTF History The Task force was formed in June 2014 under the umbrella of both Operating and Planning Committees (OC and PC) Taking from the 2001 Building Blocks, Task force produced a Concept Paper as part of the first deliverable explaining Essential Reliability Services (ERS) Developed measures that identify ERS and simultaneously produced the associated framework for those measures Data from volunteer entities was developed, tested and documented (Thanks to those entities) The product of this substantial effort is the draft Framework Report - which was sent out to OC and PC on August 25, 2015 for review and comments 3

4 ERSTF Recommendations Recommend the monitoring of the Measures and investigation of trends. Recommend planning and operating entities to use the Industry Practices. Recommend open sharing of experiences and lessons learned. Recommend all new resources to have the capability to support voltage and frequency. While beyond the formal scope of the ERSTF, the task force recognizes that Distributed Energy Resources (DERs) will increasingly affect the net distribution load that is observed by the Bulk Electric System. 4

5 ERSTF Measures Recommended Measures Measure 1- Synchronous Inertial trend (Interconnection) Measure 2 Initial Frequency Deviation following largest contingency Measure 3 - Synchronous Inertial trend (Balancing Area) Measure 4 Frequency Response Measure 6 Net Load Ramping Variability Measure 7 System Reactive Capability Industry practice recommendation Measure 5 Real time inertial model Measure 9 Voltage-System Performance Measure 10 Voltage: Short Circuit System Strength No longer being pursued Measure 8 Voltage Performance 5

6 ERSTF Webinar Presentation for PC and OC on ERS Measures Julia Matevosyan Frequency Subgroup - Measures 1-5 August 27, 2015

7 Frequency Support Measures The industry recognizes: Frequency support is essential to system reliability Frequency support capability of the generation fleet is changing due to increasing use of non-synchronous generation, other changes in generation resource mix (e.g. coal retirement, increased use of natural gas fired generation). It is important to monitor these changes and their impact on frequency support to address any issues in timely and effective manner (by e.g. incentivizing frequency support from new generation and load resources) 7

8 Frequency Response 8

9 Proposed Frequency Support Measures The ERSTF Frequency Support Sub-group considered the following measures and industry practices for adoption by the appropriate industry entities: Minimum synchronous inertial response (SIR) historically and projected for the future (Measure 1 for interconnection and Measure 3 for BA). Frequency deviation within the first 0.5 seconds following the largest contingency of the interconnection at minimum synchronous inertia conditions (Measure 2 for interconnection). Comprehensive set of frequency response measures following observed contingency events (Measure 4). Measure related to situational awareness monitoring synchronous inertia in real-time (Measure 5) was identified as industry best practice but not recommended as a measure. 9

10 Frequency Support Measures Measure 1 and 3, Synchronous Inertial Response at Interconnection and BA level Synchronous Inertial Response is stored kinetic energy that is extracted from the rotating mass of online synchronous machines following a disturbance in a power system; SIR determines initial rate of change of frequency following a contingency; At any instance SIR is calculated as machine s inertial constant H times it s MVA rating. Total system SIR is calculated as a sum of SIR from all online synchronous machines. The proposed measure would track minimum synchronous inertia, historically and projected for 3 years in the future based on generation interconnection agreements. 10

11 Measure 1, ERCOT Example 3.4 x 105 Kinetic energy, MWs at max wind penetration, historic at max wind penetration, projected based on GIAs at max wind penetration, projected based on GIA&FCs

12 Frequency Support Measures Measure 2, Frequency Deviation Following the Largest Contingency At minimum SIR conditions historic and projected (from Measure 1), determine frequency deviation within the first 0.5 second following the largest contingency (as defined by Resource Contingency Criteria (RCC) in BAL for each interconnection). This measure is only at the interconnection level. This measure would capture changes in frequency rate of change due to changing SIR on a system. Analytical expression can be used to calculate frequency deviation based on SIR and size of contingency (no need for dynamic study) 12

13 Frequency deviation after 2750 MW trip, ERCOT Example 13

14 Frequency Support Measures Measure 5, Inertial Response Monitoring in real time This measure is related to situational awareness, calculating available SIR in real time. The real-time SIR monitoring is useful for systems with high share of nonsynchronous generation where operation at low system inertia becomes a concern. This was identified as industry best practice for Operations and Planning of BAs/Interconnections with high share of non-synchronous generation 14

15 Frequency Support Measures Measure 4, Frequency Response at Interconnection Level Frequency response is the traditional metric used to describe interconnection s performance in arresting and stabilizing frequency after the loss of resources or load. Frequency response performance evaluation has been limited to BA level with, traditionally, only 2 6 seconds scan-rate frequency data; Conventional definition of frequency response is based on stabilizing frequency (Value B); Advancements in high resolution synchronized measurement technology allow examining frequency response at higher scan-rates; The proposed Measure 4 is based on sub-second resolution measurements (10-60 samples/s) from PMUs and FDRs. The measure is split into several sub-measures covering the full spectrum of frequency response. 15

16 Existing ALR1-12 Metric This metric is used to track and monitor interconnection average Frequency Response where frequency drops more than interconnection s defined threshold. It is defined in MW/0.1Hz as: FFFFFFFFFFFFFFFFFF RRRRRRRRRRRRRRRR AAAAAAA 12 = GGGGGGGGGGGGGGGGGGGG LLLLLLLL (MMMM) FFFFFFFFFFFFFFFFFF AA FFFFFFFFFFFFFFFFFF BB Value A is the average frequency from t 0 16 to t 0 2 and Value B is the average frequency from t 0+20 to t The time windows used for calculating these values accounts for variability in Supervisory Control and Data Acquisition (SCADA) scan rates, ranging from 2-6 seconds between Balancing Authorities. 16

17 Measure 4, Event Triggers It is proposed to keep the same event triggers as for ALR

18 Measure 4 A to C frequency response Captures the impacts of inertial response, load response (load damping) and initial governor response; Ratio of net MW lost to difference between Value A and Value C. FFFFFFFFFFFFFFFFFF RRRRRRRRRRRRRRRR NNNNNNNNNN = GGGGGGGGGGGGGGGGGGGG LLLLLLLL (MMMM) FFFFFFFFFFFFFFFFFF AA FFFFFFFFFFFFFFFFFF CC Trending this measure year to year will capture effects of changes in generation mix and load characteristics, and help to identify needs for synchronous inertia and fast frequency response (e.g., from storage, load with under-frequency relays, synthetic inertia). 18

19 Measure 4 A to B frequency response Captures the effectiveness of primary frequency response in stabilizing frequency following a large event (existing ALR1-12). Ratio of net MW lost to difference between Point A and Point B. FFFFFFFFFFFFFFFFFF RRRRRRRRRRRRRRRR AAAA = GGGGGGGGGGGGGGGGGGGG LLLLLLLL (MMMM) FFFFFFFFFFFFFFFFFF AA FFFFFFFFFFFFFFFFFF BB Trending this measure year to year will capture effects of changes in generation mix on governor response. 19

20 A to B frequency response (blue) and A to C frequency response (red), ERCOT 20

21 Measure 4 C to B Ratio Captures the difference between maximum frequency deviation and settling frequency. This measure is related to governor responsiveness with respect to frequency deviation reading, and their capability to arrest and stabilize system frequency CC: BB RRRRRRRRRR = FFFFFFFFFFFFFFFFFF CC FFFFFFFFFFFFFFFFFF AA FFFFFFFFFFFFFFFFFF BB FFFFFFFFFFFFFFFFFF AA Trending this measure year to year provides insight into the amount of generation providing primary frequency response. 21

22 C to B Ratio in EI 22

23 Measure 4 C to C Ratio The ratio between the absolute frequency minimum (Point C ) caused by governor withdrawal and the initial frequency nadir (Point C). CCC: CC RRRRRRRRRR = FFFFFFFFFFFFFFFFFF CCC FFFFFFFFFFFFFFFFFF AA FFFFFFFFFFFFFFFFFF CC FFFFFFFFFFFFFFFFFF AA Point C is of concern in the EI due to governor response withdrawal. Trending C to C ratio in EI for similar size events will capture if generators are working with vendors to adjust plants Distributed Control Systems (DCS) load controllers to mitigate the impact of governor response withdrawals. 23

24 C to C Ratio in EI 24

25 Measure 4 Time Based Measures Are used to capture the speed of inertial response and governor response. These measures include: t C -t 0 is the difference in time between the frequency nadir and initial event, and captures rate in which system inertia and governor response arrest system frequency. Trending t C -t 0 can be useful for ensuring that the defined times for BAL-003 fit the actual event data. In addition, trending this with respect to event size and initial frequency can help identify how deadband settings play a role in frequency arrest. 25

26 Measure 4 Time Based Measures (continued) t C -t C measure is the difference in time between the governor withdrawal minimum and the initial frequency nadir. This measure captures the time in which governor stabilization and withdrawal occurs prior to secondary controls beginning to return frequency to its initial value. t C -t 0 measure is the difference in time between the governor withdrawal minimum and the initial event. This provides a comprehensive picture of the overall time in which frequency declines and continues to fall due to the initiating event. 26

27 In summary Measures 1 4, Inertial and Frequency Response Measures, should become approved measures assigned to NERC Frequency Working Group Measure 5, Real Time Inertial Monitoring is recommended industry practice for Operations and Planning of BAs/Interconnections with high share of non-synchronous generation. 27

28 28

29 ERSTF Webinar Presentation for PC and OC on ERS Measures Clyde Loutan Ramping Subgroup Measure 6 August 27, 2015

30 Ramping Variability High penetrations of non-dispatchable resources and/or VERs may require increased system ramping capability. ERSTF analysis indicates that ramping capability is not currently an issue for many BAs, but in CAISO, with a significant amount of VERs & non dispatchable resources, ramping is already a challenge. Greater risk of over-generation during periods of low demand because some resources cannot be shut down due to long start -up times or contractual limits Need to mitigate steep intra-hour net demand ramps and multi-hour net demand ramps Need for more flexible resources with faster ramping capability Need for resources to have the capability to stop and start multiple times per day Greater difficulty in accurately forecasting operating needs of the system Potential for rapid change in the intra-hour ramp direction Any non-dispatchable resources can exacerbate minimum generation concerns 30

31 Actual wind and solar production 2,200 Wind & Solar Production --- 3/1/2014 2,000 1,800 Wind/Solar (MW) 1,600 1,400 1,200 1, Wind Solar 31

32 Actual load and net-load 27,000 Load & Net Load --- 3/1/2014 Load & Net Load (MW) 26,000 25,000 24,000 23,000 22,000 21,000 20,000 19,000 18,000 17,000 16,000 Load NetLoad 32

33 Expected increased challenges 24,000 Expected Minimum on-line resources in a typical spring months in ,000 20, Actual Minimum Net Load 2015 Actual Minimum Net Load 18,000 16,000 14,000 MW 12,000 10,000 8,000 6, Expected Minimum Net Load 4,000 2, Nuclear Non_Dis_QF Geothermal Dedicated Imports Biomas Biogas Dry_Hydro Year Min_CCYC_Prod Expected_Min_Net Load_2021 Min_Net_Load_2015 Min_Net Load_

34 ERSTF Recommended NET Load Ramping Variability Measure The TF recommends Measure 6 as a measure that should be monitored & evaluated at the BA level and the data provided to NERC annually for industry wide trending and analysis. Measure 6 provides both a historical and future view of the maximum one-hour up/down & three hour up/down ramping requirements. The data requirements include one minute data (or the smallest sample rate available) and a projected build out of generation & load over the next three years. Additional considerations are provided in the framework report for more detailed analysis by BAs. 34

35 Net-load is a NERC accepted metric 1 for evaluating additional flexibility needs to accommodate VERs Net load is the aggregate of customer demand reduced by variable generation power output Net-load is more variable than load itself and it increases as VER production increases The monthly three-hour flexible capacity need equates to the largest up-ward change in net-load when looking across a rolling three-hour evaluation window Some BAs dispatched flexible resources including VERs to meet net-load Some BAs may treat VERs as must-take and dispatch flexible resources to meet net-load 1 NERC Special Report Flexibility Report Requirements and metrics for Variable Generation: Implications for System Planning Studies, August Slide 35

36 Flexibility capacity assessment is based on current load and expected renewable build-out data Uses most current data available for renewable build-out obtained from all load serving entities For new renewable installation may use NREL s simulated production data or scale CREZs located in close geographic proximity Profiles account for both technology and location of all solar resource types Solar Thermal; Solar Thermal with storage; Solar PV Tracking & Nontracking; Distributed Rooftop Solar PV etc. Generate net-load profiles for 2016 through 2018 Generate load profiles for 2016 through 2018 Generate solar profiles for 2016 through 2018 Generate wind profiles for 2016 through Slide 36

37 The monthly 3-hour ramping need is calculated using the largest ramp in each 180 minute period The maximum monthly three-hour net load ramp within a three-hour period is the highest MW value reached within any three-hour moving window Maximum 3-hour up ramp change MW A B C The maximum netload change in three-hours can occur in less than three hours t=0 t=180 Upward Ramp = Average(t+4 min) Average(t-4min) Down Ramp = Average(t+4min) < Average(t-4min) 37 Slide 37

38 Measure 6: CAISO Example 18,000 16,000 18,000 14,000 16,000 Max 3-Hour Down Ramps, MW Max 3-Hour Up Ramps, MW 12,000 10,000 8,000 6,000 4,000 2, ,000-4,000-6,000-8,000-10,000-12,000-14,000 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Hourly Net Load Ramps, MW 14,000 12,000 10,000 8,000 6,000 4,000 2, ,000-4,000-6,000-8,000-10,000-12,000-14,

39 39

40 Back up slides 40

41 Measure 6 should be measured at a BA level to ensure the BA has adequate ramping capability within its resource mix to address net-load intra and inter-hour variability Additional considerations for BAs expecting a high penetration of wind, solar, and non-dispatchable resources: BAs may choose to track net demand variability on an annual basis. BAs may choose to trend their control performance standard 1 (CPS1, see Appendix C) scores on shortened timeframes, such as hourly or daily, to identify any correlation between significant intra-hour or multi-hour ramps and CPS1 excursions below 100% across the same timeframe BAs may choose to begin tracking the frequency and duration when their Balancing Authority ACE Limit (BAAL, see Appendix C) exceeds predefined limits BAs may choose to review their net inadvertent interchange to determine if ramping deficiencies within a BA results in inadvertent flows to neighboring entities during ramp deficiencies BAs may choose to develop day-ahead and real-time forecasting tools to better predict VER output changes 41

42 Additional Observations Greater risk of over-generation during periods of low demand because some resources cannot be shut down due to long start -up times or contractual limits Need to mitigate steep intra-hour net demand ramps and multi-hour net demand ramps Need for more flexible resources with faster ramping capability Need for resources to have the capability to stop and start multiple times per day Greater difficulty in accurately forecasting operating needs of the system Potential for rapid change in the intra-hour ramp direction Any non-dispatchable resources can exacerbate minimum generation concerns 42

43 CAISO: Actual and estimated 3-hour upward ramping needs Maximum 3-Hour Upward Ramps 20,000 18,000 16,000 14,000 12,000 MW 10,000 8,000 6,000 4,000 2,000 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 2011_3hr 7,319 6,766 6,067 5,688 5,942 6,732 7,822 7,702 7,251 6,767 6,433 7, _3hr 7,657 7,173 7,028 5,774 6,278 5,543 6,367 7,410 6,591 6,422 6,062 7, _3hr 7,524 7,171 6,736 5,881 6,096 8,745 6,426 6,024 6,591 6,609 7,355 8, _3hr 6,354 6,170 5,755 5,363 6,394 6,177 6,559 5,879 7,862 5,952 5,844 6, _3hr 13,216 15,048 14,100 11,332 11,022 10,769 10,390 12,143 14,174 12,509 15,190 17,179 43

44 ERSTF survey results Table 3: Summary of the Survey Results BA/ISO Installed Capacity of Nonsynchronous generation (NSG), 2014 Installed Capacity of NSG, 2017 Total solar with respect to overall VERs Increased Need for Flexible Capacity? CAISO 12,471 17,176 65% Yes ERCOT 11,066 21,130 30% No ISO NE 3,155* 5,591* Insignificant No IESO 4,075* 5,607* Insignificant No MISO 13,726 18,526 Insignificant No BC Hydro Insignificant No Southern BA 454 2,324 Insignificant No Duke: DEF 0 0 Insignificant No Duke: DEC Insignificant No Duke: DEP Insignificant No 44

45 ERSTF Webinar Presentation for PC and OC on ERS Measures John Simonelli Voltage Subgroup Measure 7-10 August 27, 2015

46 Voltage and Reactive Measures The industry recognizes: Voltage management is essential to reliability, Voltage management must encompass both baseline operations as well as contingency conditions, Voltage management is best done at a sub-area/cluster level due to the inability to move reactive support long distances on the transmission system, Voltage management can effectively and efficiently be done through a combination of static and dynamic reactive power resources coupled with adequate load power factor control, The ERSTF Voltage and Reactive Sub-group considered the following measures and industry practices for adoption by the appropriate industry entities. 46

47 Voltage and Reactive Measures Measure 7: Reactive Capability on the System The proposed measure would track system static and dynamic reactive resources as well as load power factors for distribution at the low side of transmission buses. These quantities would be tracked by the BA on either the entire BA footprint or as appropriate at a sub-area/cluster level within the BA footprint. These quantities would be gathered for peak, shoulder, and light load levels. The NERC Performance Analysis Subcommittee (PAS) would develop data collection protocols and the NERC System Analysis and Modeling Subcommittee (SAMS) would review the data to develop industry trends. 47

48 Measure 7: Reactive Capability 48

49 Voltage and Reactive Measures Measure 8: Voltage Performance of the System The ERSTF considered a proposal to track real-time operational voltage limit exceedances and monitoring buses with low short-circuit strength or susceptibility to fault-induced delayed voltage recovery (FIDVR) conditions. The ERSTF concluded Measure 8 had significant overlap with Measure 9 relative to real-time operational voltage limit exceedances. The ERSTF further concluded Measure 8 had significant overlap with Measure 10 relative to monitoring buses with low short-circuit strength. At this time there is no need to pursue Measure 8. 49

50 Voltage and Reactive Measures Measure 9: Overall System Performance The ERSTF considered a proposal to track system events that suggest stressed reactive capability or degraded voltage profiles. Reactive and voltage performance for these events would be evaluated across all time horizons (planning, seasonal, real-time) to compare planned performance with real-time operations. This evaluation would provide useful insight into the success of the planning process, document as built systems, and confirm effective operational management of resources in real time. The ERSTF concluded this type of post mortem analysis comports with various requirements in existing and proposed NERC standards and would fall under the existing Event Analysis Subcommittee (EAS) responsibilities. 50

51 Voltage and Reactive Measures Measure 10: System Voltage and Reactive Strength Performance The ERSTF considered a proposal to measure and track system strength based on calculating short circuit ratios for sub-areas/clusters in the system. Part One: o Planning/Transmission Coordinator will annually perform traditional short circuit evaluations per TPL standard on their system short-circuit capability for the purpose of circuit breaker fault duty analysis. o The short circuit data from this assessment can be used to calculate the short circuit ratio (SCR) at busses as defined in IEEE standard o Where SCR in sub-areas/clusters are identified as low (typically < 3), entities should utilize additional analytical study techniques to further analyze the potential for FIDVR and voltage stability issues. 51

52 Voltage and Reactive Measures Measure 10: System Voltage and Reactive Strength Performance Where there is a significant amount of inverter-based resources or other non-synchronous resources and the SCR is low an additional study process beyond the traditional short circuit ratio calculation in Part One is recommended. Part Two: o Both ERCOT and GE have developed detailed study processes specifically designed to identify potential voltage control instability amongst groups of inverter-based resources or other non-synchronous resources. o The Planning Coordinators using these suggested study processes can then identify problem areas and develop remedial actions to prevent reliability issues. ERSTF proposes that Study Process One and Two be considered as an industry practice. 52

53 Voltage and Reactive Measures In summary Measure 7 Reactive Capability on the System, should become an approved measure and assigned to NERC PAS and SAMS Measure 8 Voltage Performance of the System, does not need to be pursued at this time Measure 9 Overall System Performance, is all ready covered under the NERC EAS responsibilities so no further action is needed Measure 10 System Voltage and Reactive Strength Performance, is a recommended industry practice for Planning and Transmission Planners 53

54 54

55 Next Steps-Framework Report Framework Report Incorporate comments from OC/PC Comments due October 1 Finalize Framework Report Deliver Framework Report to OC/PC October 13 Seek OC/PC acceptance of Framework Report October 20 December 7 BOT report review and vote December 9 Release final report 55

56 Next Steps-Abstract Report Abstract Abstract is intended to be a summary of key points for Policy Makers. Release draft Abstract for OC/PC review September 8. OC/PC comments on Abstract due October 20. Release final Abstract to OC/PC for review November 20 Seek OC/PC acceptance of Abstract December 15/16 standing committee meetings. December 17 deliver Abstract to BOT. 56

57 Abstract Key Messages Goal is to inform, educate, and build awareness on the implications of the changing resource mix and how industry can evolve the system in a reliable manner Resource adequacy changing due to a variety of reasons, technology neutral, some technology advancing new capabilities Consider technical aspects of ERS when making decisions related to interconnecting new resources or market and tariff oversight Policy decisions have direct influence on changes in the resource mix, and thus can also affect the reliability of the bulk power system 57