System Reliability Modeling Datasets 2023
This page lists system reliability modeling datasets used by Energy Division to model the CAISO balancing area electric system, typically in support of the Resource Adequacy (RA) and Integrated Resource Plan (IRP) proceedings. Prior versions of this webpage were titled “Unified RA and IRP Modeling Datasets.” CPUC staff have changed the name to more directly describe the content and avoid naming specific CPUC proceedings since the modeling datasets have several use cases across multiple CPUC proceedings.
All data were exported from the “Strategic Energy & Risk Valuation Model” (SERVM), a probabilistic system reliability and production cost model developed by Astrapé Consulting. In 2023, Energy Division updated inputs to support modeling primarily for the IRP 2023 Preferred System Plan development work. Key inputs to SERVM in support of this work are provided on this page.
Electricity Consumption Profiles
Hourly MW consumption profiles for study years 2024, 2026, 2030, and 2035, for all regions modeled, based on 1998-2020 weather. Median annual energy and peak demand for all California regions are based on the CEC’s 2022 IEPR Update Demand Forecast. SERVM’s modeled consumption is developed from the CEC’s managed forecast by backing out the effects of BTM PV, BTM storage, AAEE, AAFS, TOU rates, and electric vehicle (light/medium/heavy duty) charging load. Each of these demand modifier components are separately modeled in SERVM. Energy and peak demand for non-California regions are derived from the WECC Anchor Dataset 2032.
Hourly Consumption Demand, all regions, all weather years (updated June 2023)
The above profiles are aligned to start on the day of week corresponding to the calendar year, e.g. 2026 will start on a Thursday. Leap years drop Dec 31 such that all years are 8760 hours long.
For modeling results posted after October 2023, including the modeling informing the October 5, 2023 ALJ Ruling in the IRP proceeding, hourly MW consumption profiles were aligned to start on Mondays for all study years. This was done to ensure each future year only varied from load, resource, and cost projections, and not from which day-of-week start was used. Monday start profiles for 2026, 2030, 2034, 2035, and 2039 are included in the file below.
Hourly Consumption Demand, all regions, all weather years, Monday start (updated January 2024)
Generating Units
The following list of generating units modeled in SERVM includes “baseline” and “LSE Plan” units. “Baseline” means online or in development, consistent with the definition in the IRP proceeding. “LSE Plan” means planned/new resources aggregated from LSE IRP filings in November 2022. The “baseline” was derived from the CAISO Master Generating Capability Lists through January 2023, the January 2023 NQC list, the 2032 WECC Anchor Data Set, and the November 2022 LSE IRP Compliance Filings. "LSE Plan" units were derived from aggregating all planned/new contracts in the November 2022 LSE Compliance Filings. Both 25 MMT by 2035 and 30 MMT by 2035 LSE Plan portfolio units are included in the list. Notes in the column headings explain each column where necessary.
SERVM Generator List (updated 10/5/2023)
The below list of generating units identifies the specific units modeled to inform the 2023 Proposed PSP 25 MMT Core Scenario included in the IRP proceeding January 10, 2024 Proposed Decision. It includes the named units from the SERVM Generator List above, plus RESOLVE Planned and Selected Capacity.
SERVM Modeled Unit List for 2023 Proposed PSP 25 MMT Core Scenario (updated January 2024)
Hydro Generation Profiles (June 2023 vintage)
Hourly MW profiles for the indicated study year. The profiles are provided covering 23 weather years (1998-2020) for a relatively low hydro availability year (based on 2015) and a relatively high hydro availability year (based on 2005). SERVM models hydro data from 1998-2020 but only 2015 and 2005 are posted here for brevity. The files are June 2023 vintage.
The above profiles correspond to the 25 MMT portfolio results presented on slides 35-40 only of this slide deck: 2023 Proposed PSP Reliability & Emissions Slide Deck and other SERVM results within this deck have different hydro profiles. For different scenarios with different net load conditions, SERVM may schedule the hydro differently resulting in different hydro output profiles.
Hydro SERVM-specific input variables
This file contains SERVM-specific input variables defining hydro unit inputs. It can be used to develop hydro unit inputs for other production cost models. The inputs were developed from 23 years (1998-2020) of monthly data from the EIA and 4 years of hourly data from the CAISO, BPAT, and EIA. The source data was translated into monthly generation, daily minimum, average, and maximum generation, and monthly maximum output parameters. SERVM schedules the hydro according to the net load conditions of a given scenario. For different scenarios with a different net load conditions, SERVM may schedule the hydro differently resulting in different hydro output profiles.
Renewables Generation Profiles (June 2023 vintage)
Hourly MW profiles for the indicated study year, for all unit categories in the three CAISO regions modeled, based on 1998-2020 weather patterns. These profiles correspond to the 25 MMT portfolio results presented on slides 35-40 onlyof this slide deck: 2023 Proposed PSP Reliability & Emissions Slide Deck and other SERVM results within this deck have different profiles. These files are June 2023 vintage.
Hourly MW profiles for study years 2024, 2026, 2030, and 2035 for non-CAISO California regions, based on 1998-2020 weather patterns. This file is June 2023 vintage.
Normalized Renewable Hourly Profiles
The (large) file below contains NORMALIZED renewables hourly profiles based on 1998-2020 weather patterns. It can be used to develop renewables profiles for other production cost models. The June 2023 vintage profiles above were created by scaling up the normalized profiles to match the installed capacities in the portfolio for each year. The Generator Lists linked above include the weather station name and the installed capacity of each renewables unit to match with the normalized profile name to create the final scaled up profile corresponding to that unit. The file below includes a README describing each file briefly.
Normalized solar, wind, and load modifiers profiles, 1998-2020 weather updated November 2023
The following file maps weather year profiles to weather stations to latitude and longitude as well as resource zones as defined in the RESOLVE model.
ProfileToWeatherStationMap 20230920
Besides utility-scale solar and wind categories, the normalized profiles file includes the following categories of demand modifiers:
- AAEE – Additional Achievable Energy Efficiency
- AAFS – Additional Achievable Fuel Switching
- BTMPV – Behind-The-Meter solar PV
- BTMStorageShapeCharge – Charging profile of BTM storage
- BTMStorageShapeDischarge – Discharging profile of BTM storage
- EV – Light and med-heavy-duty Electric Vehicles including AATE
- TOU_LoadIncrease – Time-Of-Use rate impacts that increase hourly load
- TOU_LoadDecrease – Time-Of-Use rate impacts that decrease hourly load
Day of week dependent demand modifier profiles such as AAEE and EV were shifted to start on a Monday. Leap years drop Dec 31 such that all years are 8760 hours long. This is important to maintain day of week alignment with the consumption demand profiles listed above.
Additional 2039 data for load modifier profiles that are not included in the above normalized profile is provided below. This additional data is required to model the 2039 study year.
2039 normalized load modifier proflies (updated January 2024)
Representation of the CAISO Transmission System and Neighbors
Transmission capability in MW
Burner-Tip Fuel and Carbon Prices