ERCOT Historic Synchronous Inertia (Kinetic Energy) and Future Projections. Julia Matevosyan, PhD Sr. Planning Engineer Resource Adequacy ERCOT

Transcription

1 ERCOT Historic Synchronous Inertia (Kinetic Energy) and Future Projections Julia Matevosyan, PhD Sr. Planning Engineer Resource Adequacy ERCOT

2 Historic Analysis Assumptions Kinetic energy is calculated as a sum of H*MVA of all online synchronous generators on hourly basis; Jan-May, Nov-Dec data for and January-May data for 2014 are included in the analysis; Peak wind penetration hour (i.e. max P wind /P load ) in each year is analyzed in details (load, wind, wind capacity factor)

3 Historic Kinetic Energy Kinetic energy, MWs Installed Capacity, MW 3.4 x ,116 9,452 10,034 10,570 11,066 Max P wind /P load 25.5% 27.4% 29.8% 35.8% 39.4% P wind, MW 6,483 6,772 7,247 8,773 9,699 Capacity Factor 71% 72% 72% 83% 88% P load, MW 25,427 24,745 24,328 24,488 24,617

4 Future Inertia Projection Methodology Expected installed wind capacity in a future year (with SGIA or SGIA&FC) Expected wind capacity in a future year X: P wind capacity Project an hour with highest instantaneous penetration of wind based on historical trends Projected peak wind penetration hour: P wind =CapacityFactor hist *P wind capacity P load = Avg. historical load at wind penetration peaks Penetration = P load /P wind Net Load=P load P wind Project system inertia for this hour, based on historical inertia trendline Synchronous inertia at projected wind penetration peak: SI = a*net Load +b

5 Kinetic Energy Trend used for Future Projections

6 Historic Kinetic Energy and Future Projections 3.4 x 105 Kinetic energy, MWs at max wind penetration, historic at max wind penetration, projected based on SGIAs at max wind penetration, projected based on SGIA&FCs Installed Capacity, MW ,116 9,452 10,034 10,570 11,066 19,443 20,630 21,130 Max P wind /P load 25.5% 27.4% 29.8% 35.8% 39.4% 69% 73.2% 75% P wind, MW 6,483 6,772 7,247 8,773 9,699 17,041 18,082 18,520 Capacity Factor 71% 72% 72% 83% 88% 88% 88% 88% P load, MW 25,427 24,745 24,328 24,488 24,617 24,700 24,700 24,700

7 Frequency deviation after 2750 MW trip (in 0.5 s)

8 Additional metrics Maximum permissible RoCoF, Hz/s: based on frequency deviation to under-frequency load shed (UFLS) and time until first fast frequency response (FFR) in a system is fully deployed (e.g. 0.5 seconds for Responsive Reserve provided by Loads in ERCOT). RoCoF max =Δf UFLS /t FFR, From this metric based on largest contingency and load damping constant, minimum inertia requirement can be calculated for a system. Hz/MW metric: this is frequency nadir per MW generation trip, this metric does not only consider inertia but also includes governor response, load damping and fast frequency response. This metric can be tracked based on historic events and projected for the future (in use in ERCOT).

9 Frequency nadir for system load conditions MW (interpolated based on historical events)

10 Appendix: Supporting data for slide 6 Installed Capacity, MW (w. FC) ,116 9,452 10,034 10,570 11, ,443 20,630 21,130 Max P wind /P load 25.5% 27.4% 29.8% 35.8% 39.4% 61% 69% 73.2% 75% P wind, MW 6,483 6,772 7,247 8,773 9, ,041 18,082 18,520 Capacity Factor 71% 72% 72% 83% 88% 88% 88% 88% 88% Net Load, MW Inertia, MWs Estimated RoCoF, Hz/s

11 Inertia Southern MW/Sec Hourly I May SCIT Inertia /1 0 5/1 10 5/1 20 5/2 6 5/2 16 5/3 2 5/3 12 5/3 22 5/4 8 5/4 18 5/5 4 5/5 14 5/6 0 5/6 10 5/6 20 5/7 6 5/7 16 5/8 2 5/8 12 5/8 22 5/9 8 5/9 18 5/10 4 5/ /11 0 5/ / /12 6 5/ /13 2 5/ / /14 8 5/ /15 4 5/15 14 SC

12 a Max and Min n California May 2014 Max Min Inertia / /16 0 5/ / /17 6 5/ /18 2 5/ / /19 8 5/ /20 4 5/ /21 0 5/ / /22 6 5/ /23 2 5/ / /24 8 5/ /25 4 5/ /26 0 5/ / /27 6 5/ /28 2 5/ / /29 8 5/ /30 4 5/ /31 0 5/ /31 20 SCIT Inertia (MW/Sec)

13 Net Imports /

14 NERC Essential Reliability Service Task Force Clyde Loutan, Senior Advisor Renewable Energy Integration October 29, 2014 NERC Office Atlanta

15 Take away from Vancouver Key focus for our group is to develop the process of predicting future ramping needs, specifically the three hour ramping need as illustrated by CAISO s example Develop hourly upward ramping needs Develop hourly downward ramping needs Develop three-hours upward ramping needs Develop three-hours downward ramping needs Slide 2

16 Loads and Resources Team Subgroup Lead Company Clyde Loutan CAISO Subgroup Members Amir Najafzadeh NERC Brendan Kirby Kirby Consulting Dave Devereaux IESO Ed Scott Duke Energy Jay Ruberto First Energy Layne Brown WECC Michael McMullen MISO Michael Milligan NREL Noha Abdel-Karim NERC Pooja Shah NERC Ron Carlsen Southern Company Todd Lucas Southern company Tom Siegrist Brickfield, Burchette, Ritts & Stone, P.C. Dariush Shirmohammadi California Wind Energy Association Aidan Tuohy EPRI Slide 3

17 2014 Monthly load vs. net-load profiles January Load vs. Net Load February Load vs. Net Load March Load vs. Net Load April Load vs. Net Load 28,000 27,000 26,000 25,000 24,000 23,000 22,000 21,000 20,000 19,000 18, ,000 26,000 25,000 24,000 23,000 22,000 21,000 20,000 19,000 18, ,000 26,000 24,000 22,000 20,000 18,000 16, ,000 26,000 24,000 22,000 20,000 18,000 16, May Load vs. Net Load June Load vs. Net Load July Load vs. Net Load August Load vs. Net Load 30,000 28,000 26,000 24,000 22,000 20,000 18,000 16, ,000 32,000 30,000 28,000 26,000 24,000 22,000 20,000 18, ,000 36,000 34,000 32,000 30,000 28,000 26,000 24,000 22,000 20,000 18, ,000 36,000 34,000 32,000 30,000 28,000 26,000 24,000 22,000 20,000 18, September Load vs. Net Load 36,000 34,000 32,000 30,000 28,000 26,000 24,000 22,000 20,000 18, Slide 4

18 Option 1 CAISO s determination of flexible capacity for future years NREL s 2005 VERs profiles were used to develop the wind profiles and Clean Power data was used to develop the solar profiles Obtain the latest PTOs assumption of VERs build-out and CPUCs RPS calculator Calculate 1-minute net-load Load Use minute actual load data Wind Develop 1-minute wind production profiles for CREZs based on their geographic location using NREL s 2005 wind profiles Solar installation Develop 1-minute solar production profiles for CREZs based on their geographic location and technology using NREL s 2005 solar profiles (i.e. solar thermal, solar PV tracking & solar PV fixed) Aggregate all new solar 1-minute production data by technology New CREZs does not have the load/solar correlation but the maximum 3-hour ramps during the non-summer months are highly influenced by sunset which is consistent with existing solar data Slide 5

19 Option 2 - The monthly flexibility capacity requirement is calculated using the most recent full year of the CAISO s load, wind, and solar 1-minute data Use 2013 actual CAISO s load, wind and solar 1-minute data For new VERs installation, use NREL s simulated production data for CREZs located in close geographic proximity to develop minuteby-minute production profiles Solar profiles were created using both technology type and location of the new resources Generate net-load profiles for 2020 Generate 1-minute load profiles for 2020 Generate 1-minute solar profiles for 2020 Generate 1-minute wind profiles for 2020 Slide 6

20 Option 2 - Wind growth assumptions Use actual 1-minute wind production data for the most recent year e.g actual 1-minute data was used to build minute data 1-minute wind profiles for projects installed in 2020 were created using 2013 actual data for the months the projects were not inservice (i.e. profiles for projects installed in May 2013 were created for January through April) Wind 1-minute profiles for 2020 were created by scaling the 1- minute wind data for 2013 based on installed capacity 2020 W 1-min = 2013W Actual_1-min * 2020W Installed Capacity /2013W Installed Capacity Slide 7

21 Option 2 - Solar growth assumptions Existing solar Use actual solar 1-minute production data for the most recent year (e.g actual 1-minute solar data was used to develop 2020 profile 1-minute solar profiles for projects installed in 2013 were created for the months the projects were not in-service using 2013 actual solar data New solar installation Develop 1-minute solar production profiles for CREZs based on their geographic location and technology using NREL s 2005 solar profiles (i.e. solar thermal, solar PV tracking & solar PV fixed) Aggregate all new solar 1-minute production data by technology New CREZs does not have the load/solar correlation but the maximum 3-hour ramps during the non-summer months are highly influenced by sunset which is consistent with existing solar data Sum the actual 1-minute existing solar production data with the aggregated simulated solar data for new installation Total solar min = 2013 Actual_1-min Simulated_1-min data Slide 8

22 Calculating 1-hour and 3-hour upward/downward ramp capacity using net load (NL) Option 1 One minute moving window 1-Hour Ramp: NL 61 -NL 1, NL 62 -NL 2, NL 63 -NL 3.NL n+61 -NL n+1.n 0 3-Hour Ramp: NL 181 -NL 1, NL 182 -NL 2, NL 183 -N L3.NL n+180 -NL n Option 2 Five minute moving window 1-Hour Ramp: NL 61 -NL 1, NL 66 -NL 6, NL 71 -NL 11.NL 5n+61 -NL 5n+1. n 0 3-Hour Ramp: NL 181 -NL 1, NL 186 -NL 6, NL 191 -NL 11.NL 5n+181 -NL 5n+1 Option 3 Average of one minute moving window 1-Hour Ramp or 3-Hour Up Ramp If(Avg(NL n+6 +NL n+7 +NL n+8 +NL n+9 +NL n+10 ) - Avg(NL n+1 +NL n+2 +NL n+3 +NL n+4 +NL n+5 )) >0 Down Ramp If(Avg(NL n+6 +NL n+7 +NL n+8 +NL n+9 +NL n+10 ) - Avg(NL n+1 +NL n+2 +NL n+3 +NL n+4 +NL n+5 )) <0 Slide 9

23 Maximum monthly 1-hour 1-Hour upward ramping capacity for 2014 and expected 1-hour ramping capacity for ,000 Monthly 1-Hour Up Ramp Capacity & ,000 7,000 6,000 MW 5,000 4,000 3,000 2,000 1,000 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec ,084 3,844 3,279 3,641 2,308 3,063 2,340 2,464 2, ,152 6,944 6,688 5,852 4,830 4,517 4,283 5,131 6,362 6,499 7,199 7, Slide 10

24 Maximum monthly 1-hour downward ramping capacity for 2014 and expected 1-hour ramping capacity for Monthly 1-Hour Down Ramp Capacity & ,000-2,000-3,000 MW -4,000-5,000-6,000-7,000 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec ,943-2,671-3,389-3,822-4,131-4,180-3,674-3,992-4, ,355-4,443-5,821-4,989-4,956-4,736-5,009-5,375-5,721-5,452-5,084-4, Slide 11

25 Maximum monthly 3-hour upward ramping capacity for 2014 and expected 3-hour ramping capacity for ,000 18,000 16,000 14,000 12,000 Monthly 3-Hour Up Ramp Capacity & 2020 MW 10,000 8,000 6,000 4,000 2,000 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec ,288 7,078 6,843 6,613 5,574 5,150 5,750 5,577 7, ,788 15,830 14,917 12,071 11,800 11,516 11,127 12,867 15,053 13,194 15,899 17, Slide 12

26 Maximum monthly 3-hour downward ramping capacity for 2014 and expected 3-hour ramping capacity for Monthly 3-Hour Down Ramp Capacity & ,000-4,000-6,000 MW -8,000-10,000-12,000-14,000 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec ,490-6,734-7,625-8,461-10,615-9,831-9,729-10,449-10, ,228-9,882-9,651-9,565-12,205-10,837-11,830-13,042-13,310-10,362-8,657-9, Slide 13

27

28 WWSIS III: Western Frequency Response and Transient Stability Study System Inertia Discussion GE Energy Nicholas W. Miller (PM) Miaolei Shao Slobodan Pajic Rob D Aquila NREL Kara Clark (PM) The draft report is under review by the TRC and by DOE. Therefore, all of the results and statements in this presentation MUST be regarded as preliminary and subject to further review and modification. Presented by Jason MacDowell for ERSTF Frequency Response TF Atlanta, GA October 29-30,

29 Light Spring Base and High Renewable Cases ( 22) ~25% of US generation ~ 25 GW of wind and solar total in reference case ~ 52.8 GW of wind and solar total ~54% of US generation US only WECC CALIFORNIA DSW NORTHEAST NORTHWEST Wind (GW) PV (GW) CSP (GW) DG (GW) Others (GW) total (GW) Preliminary Results of Western Wind and Solar Integration Study, Phase III : Transient Stability and Frequency Response. Subject to Final Review and Approval by DOE

30 Inertial Impact of Increasing VG on ROCOF (rate of change of frequency) Light Spring Base Light Spring Hi-Mix Light Spring Hi-Mix CSP changed to PV ROCOF (1 sec 3 sec): Blue (Base): Hz/Sec Red (+25GW Wind + Solar): 0.113Hz/Sec = +18% Green (7GW CSP -> PV): Hz/Sec =+22% Initial ROCOF proportional to 1/Inertia Trip 2 Palo Verde units (~2,750MW) Preliminary Results of Western Wind and Solar Integration Study, Phase III : Transient Stability and Frequency Response. Subject to Final Review and Approval by DOE

31 Thank you! Preliminary Results of Western Wind and Solar Integration Study, Phase III : Transient Stability and Frequency Response. Subject to Final Review and Approval by DOE