1. High and Low Load Sensitivities TEPPC 2024 CC Data Work Group January 27, 2015 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

2. Overview Background Information on Running Sensitivity Analysis Existing Assumptions for High & Low Loads – 10 Year Loads – 20 Year Loads Assumptions for Load Sensitivities – Discuss assumptions for high and low load conditions used by the California Energy Commission – Discuss the impact of high efficiency programs on the load forecast Assumptions for high loads – Loads used by the ISO for Reliability Analysis – Loads used by BPA for Reliability Analysis Where do we go from here? 2 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

3. TEPPC\TAS Meeting 021611 Running Sensitivities with Every Study Cycle DWG participants see value in running sensitivities (High and Low) with every study cycle (e.g., Gas Prices, Hydro and Loads) A regular schedule will give DWG and others sufficient time to prepare for getting the data and assumptions delivered on time. The question about submitting an associated study request to TEPPC been settled by Brad. The request is internal to the TEPPC process and should be settled as such. 3 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

4. 4 Already planning to run low load and high DSM sensitivities o No consideration been given to run high load sensitivity. The LRS loads are one in two year risk exposure, whereas, transmission planning is typically performed with one in five and one in ten. TEPPC planning is transmission planning. TEPPC\TAS 013012 DWG report on Load Sensitivities:

5. Existing Assumptions for High & Low Dan Beckstead, WECC 5 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

6. Assumptions for High & Low Load Conditions Angela Tangahetti, CEC 6 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

7. Introduction Topics to Cover: Developing CEC IEPR “Common Cases” Overview of Common Case Methodology Common Case Input Assumptions 7 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

8. Purpose of IEPR “Common Cases” Energy sectors serving California are complex, interdependent systems lead to three common cases that “easily” translate across sector Lack of common case assumptions led to sectors “fractured” analytical approaches Stronger analytical basis for policy discussions 8 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

9. Common Cases Require Common Definitions Defining cases key to coordination “High” & “Low” not specific enough Three worldviews chosen to model Reference/Mid Case or Business as Usual High Energy Consumption Low Energy Consumption 9 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

10. Graphical Representation of Iterative Modeling Process WECC Electricity Dispatch Model North American Gas Model Updated Economic/ Demographic Assumptions CA Transportation Demand Models CA Electricity Demand Models WECC Electricity Dispatch Model

11. Common Case Input Assumptions Gross Domestic Product Growth CPI Inflation Gross State Product Population Growth Energy Efficiency Improvements Demand Response Carbon Prices Weather (HDD/CDD)

12. Trade-Offs in High and Low Energy Consumption Cases High and Low Consumption Scenario for one sector comes at expense of other sectors Some trade-offs necessary in defining high and low cases Chosen approach was “Major Driver” test If input value was major driver in one model but not others, value set by model where major driver

13. Resolution of Conflicting Variables Variable Electricity Price NG Price Crude Oil Price EV Penetration Coal Price NGV Penetration Controlling Model Electricity Natural Gas Transportation Electricity Transportation

14. Understanding Case Development Reference/Mid Case reasonably expected trajectory given best available input High and Low Cases Energy Consumption cases are reasonable range High and Low Cases Energy Consumption are NOT most extreme possible

15. High and Low Common Case Assumptions Relative to Mid/Reference Case Input CategoryHigh DemandLow Demand Natural GasElectricityTransportationNatural GasElectricityTransportation GDPHigh Low Wholesale Electricity PriceLow High Low Retail Electricity RatesLow High Low NG Supply CostLow High Low Crude Oil PriceHighNo EffectLow No EffectHigh Pop/DemographicsHigh Low Renewables (Gen)LowHigh Low Energy EfficiencyLow No Effect (High)High No Effect (High) Demand ResponseLow No Effect (High)High No Effect (High) CHPLow No Effect (High)High No Effect (High) Carbon PriceLow High Weather (HDD/CDD)High No EffectLow No Effect Low Carbon Fuel StandardHigh Low High Electric Vehicle PenetrationHigh Low High Ren. Fuel VehiclesNo Effect LowNo Effect High CAFÉ StandardsNo Effect LowNo Effect High Coal PriceLow No Effect (Low)High No Effect (High)

16. Notes For Chart on Previous Slide Terminology - "High" and "Low" case metrics is demand/consumption High - Represents a value above those identified in the Mid Case Low - Represents a value below those identified in the Mid Case Highlighted cells represent values which have a conflict between what settings the modeling teams would each use Parentheticals represent theoretical setting if the model allowed it. Actual values are outside parenthesis and represent limitations of model complexity or setting. Major/Minor drivers - Not all models are strongly sensitive to all of these factors. The goal is to identify which models see these inputs as "Major" drivers and use those models to determine which of these values should be used for the common cases.

17. The Impact of High Efficiency Programs on the Load Forecast Galen Barbose, LBNL 17 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

18. 2022 High DSM Case Assume “all cost-effective EE potential” is achieved throughout the West – Intended as a boundary case – Agnostic about what types of policies are used to get there (codes, standards, customer-funded programs, etc.) Rely on recent existing EE potential studies to estimate potential for individual utilities/regions – Extrapolate to regions for which recent potential studies are unavailable – Lots of complexities (e.g., differences in study vintage/scope, baseline definition, etc.) and approximations 18

19. 2022 High DSM Forecast Adjust Only for Incremental EE Beyond What Is Contained in Common Case 19

20. 2022 WECC-Wide EE Impacts Common Case vs. High DSM Case 20 GWhMW Cumulative Savings (2011-2021) Common Case112,46624,874 High DSM Case203,52543,592 Cumulative Savings (% of 2021 Load) Common Case 10.2%11.5% High DSM Case 18.4%20.1% CAGR (2010-2021) Common Case 1.4%1.3% High DSM Case 0.5%0.3% Common Case EE savings represent roughly 10% of total WECC-wide load High DSM Case represents roughly an additional 8% reduction in WECC- wide annual energy and 9% reduction in non-coincident peak

21. 2022 High DSM Impacts: Reduction from Common Case by BA 21

22. 2032 High DSM Case Partnered with Itron to apply their Statistically Adjusted End-Use (SAE) load forecasting framework – Econometric model with stock efficiency and saturations specified for 30 separate end-uses – Produces monthly energy and peak demand, disaggregated by end-use Key Efficiency Assumption for High DSM Forecast: Average stock efficiency for each end-use reaches level equivalent to the most efficient models commercially available today Requires a structured sequence of load forecasts Calibrated to the official WECC Reference Case 22

23. 2032 High DSM Impacts Reduction from Reference Case 23 Roughly 20% reduction in WECC-wide annual energy in 2032, relative to reference case Variation across states reflects differences in end-use mix, reference case efficiency level, weather, etc.

24. A Few Take-Aways and Questions to Consider Increasing or decreasing EE by 50% relative to what is embedded in Common Case forecast is equivalent to about a +/- 5% adjustment in total WECC-wide loads over 10 years Prior high DSM study cases suggest aggressive EE alone could reduce loads by almost 10% over 10 years, compared to Common Case – Is an aggregate 10% reduction for Low Load Sensitivity too modest, given additional potential drivers for low load beyond EE? EE impacts differ greatly by load zone (e.g., 2-20% of load) – Is a one-size-fits-all 10% reduction for the Low Load Sensitivity too coarse; are region- or BA-specific adjustments warranted? EE impacts on peak demand are not proportional to energy impacts, but perhaps close enough for level of precision desired 24 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

25. Loads Used by the CAISO for Reliability Analysis Irina Green, CAISO 25 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

26. Load Forecast for Annual Transmission Planning Process (TPP) CEC Load forecast is used as the starting point – For 2014-2015 TPP, the mid-case California Energy Demand Forecast 2014-2024 released by California Energy Commission (CEC) dated January 2014 with the Mid-Case LSE and Balancing Authority Forecast spreadsheet updated as of February 8, 2014. – Energy Efficiency Adjustments based on 2013 IERP final report from January 23, 2013 Using Mid Additional Achievable Energy Efficiency scenario http://www.energy.ca.gov/2013_energypolicy/documents/demand- forecast_CMF/Additional_Achievable_Energy_Efficiency/January_2014 _files/ http://www.energy.ca.gov/2013_energypolicy/documents/demand- forecast_CMF/Additional_Achievable_Energy_Efficiency/January_2014 _files/ 26 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

27. Load Forecast (continued) Methodologies used by Participating Transmission Owners (PTOs) to create bus-level load forecast are documented in the Study Plan 1-in-10 year heat wave load projection for individual local area studies 1-in-5 year heat wave load projection for bulk system studies PTOs subtract losses from the CEC forecast Pumping loads are modeled as generators with negative output CAISO studies summer peak, summer off-peak, minimum load cases, winter peak cases for some areas, and other cases as needed (spring peak, summer partial peak) 27 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

28. Study Areas 28 W ESTERN E LECTRICITY C OORDINATING C OUNCIL Northern Area: PG&E – bulk system and local areas Southern area: SCE, SDG&E, VEA

29. Load Forecast Methodology PG&E PG&E creates bus-level load forecast (using CEC forecast as the starting point) – PG&E loads in the base case Determination of Division Loads based on historical data and projected load growth. Total PG&E load growth allocated to divisions and adjusted based on 1-in-10 or 1-in-5 temperatures Allocation of Division Load to Transmission Bus Level depending on types of loads: conforming, non- conforming, self generation and plant (not included in division loads) Municipal loads in the base case – municipal utilities provide their load forecast to PG&E 29 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

30. Load Forecast Methodology SCE 30 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

31. Load Forecast Methodology SDG&E Utilize CEC’s latest load forecast as the starting point SDGE’s methodology to create bus-level load forecast – Actual peak loads on low side of each substation bank transformer from historical data – Normalizing factors applied for achieving weather normalized peak – Adverse temperature adjustment factor applied to get the adverse peak – Adverse load and coincident load determined for each substation 31 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

32. Load Forecast Methodology Valley Electric (VEA) Historical SCADA data and load plans Adjusted to CEC forecast 32 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

33. Questions? Comments? Irina Green igreen@caiso.com 916-608-1296igreen@caiso.com 33 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

34. Loads Used by the BPA for Reliability Analysis Reed Davis, BPA 34 W ESTERN E LECTRICITY C OORDINATING C OUNCIL

35. BPA’s Transmission System 35

36. 36 BPA’s methodology to create BUS forecasts Historical data used to create monthly energy and peak forecasts Weather normalizing factors included as appropriate Forecasted with a 34 year average peak producing temperature. Considered a 1 in 2 peak. Temperatures are updated every 10 years. Forecasts reviewed with local utility staff to make sure it aligns with local construction plans. New large loads are reviewed and added directly to forecast if load has a greater than 70% chance of occurring. Permanent load transfers between BUS are included. Specific large conservation projects are included directly at the BUS level.

37. 37 BPA’s methodology for special BUS forecasts Using the base BUS forecast we – Adjust to 1 in 20 weather conditions – Add economic uncertainty based on economic conditions surrounding the BUS and load delivery over the BUS based on load type: conforming, non-conforming. – Adjust for changing conservation plans

38. 38 W ESTERN E LECTRICITY C OORDINATING C OUNCIL