SUMMARY .gov

?PUBLIC UTILITIES COMMISSION OF THE STATE OF CALIFORNIA Item no: 4 (Rev.1) Agenda ID: 18386ENERGY DIVISION RESOLUTION E-5077 June 25, 2020 RESOLUTIONResolution E-5077. Adopts updates to the Avoided Cost Calculator for use in demand-side distributed energy resources cost-effectiveness analyses.PROPOSED OUTCOME:Adopts certain major modeling adjustments for the Avoided Cost Calculator for use in distributed energy resource cost-effectiveness analyses.SAFETY CONSIDERATIONS:Based on the information before us the Resolution does not appear to result in any safety impacts.ESTIMATED COST: No incremental cost. Funds necessary for updates to the Avoided Cost Calculator were authorized in Decision (D.)16-06-007.Authorized by D. 16-06-007, issued on June 15, 2016 and D.20-04-010 issued on April 16, 2020.__________________________________________________________SUMMARYD.20-04-010 authorized Energy Division to issue a resolution providing the final updated 2020 Avoided Cost Calculator (ACC), consistent with the policies adopted in the decision. The ACC is used in cost effectiveness analysis of distributed energy resource (DER) programs and policies. The Decision adopted major and minor changes to the ACC.This Draft Resolution provides the final 2020 ACC and related documentation, consistent with policies adopted in D.20-04-010. This Draft Resolution describes the methodological updates to the 2020 ACC, including details of increased alignment with the IRP and DRP proceedings, major changes to the electric avoided cost calculator, major changes to the natural gas avoided cost calculator, and several minor changes.BACKGROUNDThe Avoided Cost Calculator (ACC), first adopted in D.05-04-024, was originally used to measure Energy Efficiency (EE) cost-effectiveness. The assumptions, data, and models used in the ACC require periodic updates to stay current with market conditions, prices, and trends. Thus, semi-regular improvements to the ACC modeling software and data input updates were adopted in several Energy Efficiency proceedings by D.06-06-063, D.09-09-047, and D.12-05-015. D.09-08-026 modified and adopted the ACC for use by customer generation (then called distributed generation) programs.D.10-12-024 modified and adopted the ACC for use by demand response programs, and adopted Demand Response Cost-Effectiveness Protocols, which detailed those ACC modifications. The Demand Response Cost-Effectiveness Protocols were subsequently updated in D.15-11-042, including updates to the ACC. In 2014, the IDER proceeding opened with a focus on developing policy to facilitate the use of DERs. Among its goals was to establish a unified cost-effectiveness framework that would apply to all DER programs, technologies, and proceedings. The IDER proceeding established a four-phase plan to accomplish this, the first phase of which was to establish one Avoided Cost Calculator for use in all DER-related proceedings, and define a process to regularly update the ACC.D.16-06-007 authorized annual updates to the ACC, consisting of minor changes, corrections and data updates, via Resolution drafted by Energy Division staff. D.16-06-007, OP 2 states:The Commission’s Energy Division, no later than May 1st each year, shall draft a Resolution recommending data updates and minor corrections to the avoided costs calculator and, when appropriate the inputs, as described in this decision. Energy Division may issue a draft Resolution updating the Avoided Cost Calculator for 2016 after this Decision is adopted.D.19-05-019 revised D.16-06-007, authorizing biennial processes for making both major and minor changes to the ACC. This decision modified the schedule set out in D.16-06-007, by authorizing Resolution adopting minor changes to the ACC to be released for public comment no later than May 1st of every odd-numbered year, as well as establishing a process for making major changes (in addition to minor changes and updates) during even-numbered years. That process began with a workshop on August 30, 2019, to discuss proposals for both major and minor changes to the 2020 ACC. Parties filed testimony on October 7, 2019, which included proposals for major and minor changes to the Avoided Cost Calculator. On October 21, 2019, parties filed rebuttal testimony. The assigned Administrative Law Judge presided over an evidentiary hearing on November 18, 2019. On November 20, 2019, the Administrative Law Judge issued a ruling, inviting parties to file comments on the Energy Division Staff Proposal for 2020 Avoided Cost Calculator Update (Staff Proposal) along with opening briefs and reply comments with reply briefs. On April 16, 2020, D.20-04-010 (the Decision) adopted a modified Staff Proposal, as summarized in Attachment A of the Decision. The Decision authorizes Energy Division to issue a draft resolution providing the final ACC within 30 days of the Decision issuance. The Decision approved major changes to both the electric and natural gas calculators to create greater alignment between the ACC, the Integrated Resource Planning (IRP) Rulemaking (R.) 16-02-007, and the Distributed Resource Planning R.14-08-013. Additionally, the Decision approved the addition of a new avoided cost for high global warming potential (GWP) gases. The Decision also authorized the Director of Energy Division to host workshops or webinars to: (a) provide calculations for the Net Cost of New Entry for battery storage, (b) review the post-2030 greenhouse gas values, (c) educate parties and stakeholders on the greenhouse gas emissions avoided costs, (d) discuss the final values for the unspecified distribution avoided costs, (e) provide parties with the details of the method to derive avoided transmission costs, and (f) review details of the avoided cost of high global warming potential gases.Energy Division’s consultant, Energy and Environmental Economics, Inc. (E3) performed the update of the ACC under direction from Energy Division staff. E3 has issued a draft ACC spreadsheet and documentation that details the proposed set of changes to the ACC. Energy Division staff has posted these files to the CPUC’s Public Documents Area website, as described in Appendix A.In accordance with OP 1 of the Decision this Resolution adopts the changes to the ACC. According to D.16-06-007, Conclusion of Law 2, all DER proceedings are required to use the ACC adopted in the IDER Rulemaking (R.) 14-10-003 when performing cost-effectiveness analysis of DER programs. Hence, any direction or guidance provided by the Decision in concert with this Resolution supersedes any contradictory provisions of previously discussed Decisions, Resolutions, or other documents adopted by the Commission, such as the Demand Response Cost-Effectiveness Protocols.DISCUSSIONWe have reviewed the Avoided Cost Calculator (ACC) updates made by staff’s consultant E3 and find that the proposed ACC updates are within the scope ordered by D.16-06-007, D.19-05-019, and D.20-04-010. The ACC updates are found to be necessary to more accurately reflect Commission policies and priorities related to resource planning, as well as to better reflect market conditions, trends and prices. We have determined that it is reasonable to adopt these changes. Greater Alignment with the IRP and DRP ProceedingThe 2020 ACC update brings greater alignment between the IDER and the IRP and DRP proceedings, as detailed in this section. Data from IRP was used to update energy, ancillary services, and greenhouse gas avoided costs. Data from DRP was used to update transmission and distribution avoided costs. As a matter of policy, the Decision stated that the ACC will reflect the IRP proceeding’s modeling inputs and outputs. The IRP proceeding uses RESOLVE capacity expansion and SERVM production cost modeling to determine the least-cost resource portfolios for meeting electricity sector GHG emissions targets.The ACC uses RESOLVE model inputs for financial assumptions, natural gas prices and energy storage technology costs. RESOLVE model results are used to estimate the GHG adder and forecast cap and trade values. The adopted values are used in SERVM production cost modeling to estimate the hourly avoided energy and ancillary services costs and marginal GHG emissions used in the ACC. Additionally, the “No New DER” scenario, developed in the RESOLVE model and based on the Reference System Portfolio adopted in the IRP proceeding, will be the basis for most of the avoided cost inputs. The “No New DER” scenario is a counterfactual load forecast that includes no new distributed energy resources installed after 2018. It represents what the forecasted load would be if no new distributed energy resources were to be installed.The IRP Reference System Portfolio includes forecasts for energy efficiency, demand response, and behind-the-meter solar and energy storage. The portion of these DERs attributed to utility-sponsored programs is removed from the load forecast to create the No New DER scenario. Thus, in the No New DER scenario all energy efficiency, behind-the-meter solar and storage, and other demand-side resources would remain at the 2018 level and demand response resources are assumed to be zero. A table showing the exact amounts of DERs removed to create the No New DER scenario is included in the ACC documentation referenced in Appendix A.The Decision also called for greater alignment with the Distributed Resources Planning (DRP) proceeding, R.14-08-013. In D.20-03-005, the DRP proceeding developed methods for modeling transmission and distribution avoided costs for consideration in the ACC. D.20-03-005 adopted a staff proposal entitled Staff Proposal on Avoided Cost and Locational Granularity of Transmission and Distribution Deferral Values. (Staff White Paper). The ACC will reflect the Staff White Paper’s proposed framework for using unspecified distribution costs as the basis for the avoided cost of distribution. The ACC will also continue to use the current method of estimating the avoided cost of transmission, as recommended in the Staff White Paper. The current method uses marginal transmission values filed by the utilities in their general rates cases. However, currently only PG&E includes a marginal transmission value. Hence, the method used by PG&E to develop this value has been extended to SCE and SDG&E, and marginal transmission capacity costs for those two utilities have been derived based on utility-specific data. Major Changes to the Electric Avoided Cost Calculator Avoided CostCurrent MethodNew MethodData SourceGeneration CapacityCombustion Turbine Cost of New Entry Battery Storage Cost of New EntryRESOLVE input assumptionsEnergyEnergy futures and gas turbine modelingRESOLVE and SERVM modelingSERVM outputsAncillary Servicespercentage of energyRESOLVE and SERVM modelingSERVM outputsGHG ValueBased on RESOLVE GHG shadow price and cap & tradeBased on RESOLVE GHG shadow price and cap & tradeRESOLVE outputs, cap & trade pricesGHG emissionsImplied market heat rate short-run marginal emissionsSERVM short-run marginal emissions and RESOLVE long-run grid emissions intensityRESOLVE and SERVM outputs, cap & trade prices, annual electric sector GHG goalsRenewable Portfolio StandardIncorporated into avoided GHG in 2019NANATransmissionGRC marginal cost filingsFrom DRP guidanceGRC filings and historical utility cost and financial dataDistributionGRC marginal cost filingsFrom DRP guidanceGNA dataHigh GWP gasesNAMethane & refrigerant leakage modelingCARB dataThe table states the previous method (as reflected in the 2019 ACC), new method (as is included in the 2020 ACC), and data source for each avoided cost. The Decision added the avoided cost of high GWP gases, and separated transmission and distribution avoided costs. Note that the table also includes, for historical reference, a previously used avoided cost, the avoided cost of meeting Renewable Portfolio Standard (RPS) goals, which was absorbed into the avoided GHG cost as of the 2019 ACC update.This section addresses major changes to each of the avoided costs listed in the table. Additional technical details of each change can be found in the ACC documentation linked in Appendix A.Generation Capacity Avoided Cost. Previously, the ACC estimated the avoided cost of generation capacity using a natural gas combustion turbine as a proxy. The annual capacity values were allocated to each hour of the year, for 30 years, using E3’s RECAP model. The results of the RESOLVE model show that a battery storage resource better represents the marginal capacity unit. To create greater alignment with IRP, the generation capacity value will now use a new 4-hour battery storage resource as a proxy. The Avoided Cost Calculator uses RESOLVE model input assumptions for the fixed costs of a new 4-hour storage battery and calculates the annual levelized fixed cost of a battery over its expected useful life. The revenue that batteries earn from the energy and ancillary markets will be based on SERVM production cost modeling, and subtracted from the leveled fixed costs to calculate a Net Cost of New Entry in $/kW-yr. Energy Avoided Cost. Previously, the avoided cost of energy was forecasted using energy futures and gas turbine modeling. The average energy cost in the short run was based on the last 22 trading day average on-peak and off-peak market prices forecasts for NP-15 and SP-15. For the long run, energy costs were forecasted using last available futures market price and long-run energy market price. The avoided cost of energy will be now be determined by hourly values from the SERVM model, based on the No New DER case. Because SERVM models the dispatch of all generators, it produces more accurate values for future energy prices than the previous methodology. Ancillary Services Avoided Cost. Previously, the avoided cost of ancillary services was forecasted as a percentage of wholesale energy costs. Estimates of hourly avoided ancillary services costs, will come from SERVM production cost modeling. The SERVM modeling uses data from the No New DER case to forecast ancillary service prices. Because the SERVM model simulates the dispatch of electric resources, it is a more accurate indicator of actual ancillary services prices than the previous method.Greenhouse Gas Avoided Cost. The avoided cost of GHG is an estimate of the cost ratepayers will incur to achieve the electric sector’s share of California’s GHG goals. This avoided cost estimates the total cost that will be incurred, including both cap and trade allowance prices and the additional electric sector supply costs for delivered renewable energy needed to meet the GHG goals. Previously, greenhouse gas impacts were based only on hourly marginal emissions and calculated using an implied heat rate incorporating market price forecasts for electricity and natural gas. This approach does not reflect the GHG intensity of the electric grid, which must decline each year, as determined in the IRP proceeding, to reach the GHG goals. When energy usage decreases due to DERs such as energy efficiency, or increases due to electrification programs, this has the short-run impact of changing a utility’s cap and trade obligation. The short-run impact is calculated in the various resource cost-effectiveness tools by multiplying the hourly marginal electric grid emissions (in tonnes/kWh) by the change in load in kWh. However, in the long-run, changes in load will result in changes in a utility’s planning and procurement of renewable energy, as the utility must rebalance their supply portfolio to achieve their GHG goals.The 2020 ACC uses a combination of hourly marginal emissions and resource portfolio rebalancing to more accurately project hourly GHG emissions over time. Hourly marginal emissions will be estimated for each year from SERVM production simulation modeling. Portfolio rebalancing to achieve the annual target for average GHG intensity of the electric grid will be estimated for each year from RESOLVE modeling. The GHG costs avoided by demand-side actions will be calculated in two steps, based on the annual energy sector GHG intensity target. In the first step, hourly marginal emissions up to the annual grid intensity target will be valued at the cap and trade allowance price. In the second step the supply portfolio rebalancing necessary to achieve annual grid intensity target will be valued at the energy sector GHG value. The energy sector GHG avoided cost reflects the marginal cost of GHG abatement based on the additional supply costs needed to meet the GHG goals. The GHG avoided cost is based on GHG shadow prices modeled in RESOLVE for the Reference System Portfolio. The GHG adder is the difference between the GHG avoided cost and the cap and trade allowance price forecast. The 2020 ACC includes separate categories for the cap and trade allowance price and the GHG adder, the sum of which equal the GHG avoided cost.The GHG shadow prices are very low in early years and very high in later years. The GHG shadow price curve has been modified to a straight line to ensure steady deployment of distributed energy resources. This approach also reflects that making cost-effective GHG reductions in early years is preferable to making them in later years. In developing the GHG adder used in the 2020 ACC, Staff considered the RESOLVE model’s 2020-2030 GHG shadow price values, as well as post-2030 values, as per the Decision. Staff determined that the GHG adder proposed in the Staff Proposal represents the best estimation of the marginal GHG avoided cost. The ACC documentation provides more information on the various GHG adders that were considered.Distribution Avoided Costs. Previously, the ACC used the marginal transmission and distribution capacity costs from utilities’ General Rate Case Phase Two proceedings for the avoided cost of distribution and transmission, as a combined value. The Decision adopted a methodology which calculates transmission and distribution avoided costs separately, and includes only unspecified costs, following the specific guidance in the Staff White Paper, as per D.20-03-005. Unspecified distribution deferral avoided costs (transmission deferral avoided costs are discussed below) reflect the cost of distribution capacity projects that are likely to be needed in the future but are not specifically identified in current utility distribution planning. Unspecified distribution deferral avoided costs will be calculated using a system-average approach. The ACC will use a counterfactual forecast to determine the impact of distributed energy resources on load. The ACC will extrapolate the avoided cost estimates from the Distribution Deferral Opportunity Report and the Grid Needs Assessment, as filed in the DRP proceeding. Transmission Avoided Costs. As mentioned above, the ACC previously calculated transmission and distribution jointly using values from utility General Rate Case Phase 2 Proceedings. The Decision acknowledged that DERs avoid transmission costs but stated that, as D.20-03-005 determined, the record of the DRP proceeding provided no specific method for determining unspecified transmission costs, other than recommending continued use of the existing method.Therefore, the Decision directed that the ACC continue to use marginal cost of transmissions values from the General Rate Case Phase 2 proceedings. As PG&E is the only utility to file transmission-level costs in their general rate case, transmission values for San Diego Gas & Electric and Southern California Edison will be modeled using PG&E’s method and data specific to each utility. The ACC documentation provides the details of these calculations.High Global Warming Potential Gases Avoided Costs Previously, the ACC did not include avoided costs associated with high global warming potential (GWP) gases. The ACC will include a new avoided cost associated with leakage of refrigerants and methane, which are high GWP gases. Considering the avoided cost of high GWP gases is essential, due to the increased statewide focus on programs designed to replace natural gas appliances with electric appliances. The new avoided cost includes three components, or “use cases.” Two use cases will apply to methane reductions in the electric and gas sectors, respectively. The third use case will apply to refrigerant leakage emissions and will be used for programs that change the amount or type of appliances that use refrigerants. The impacts of methane leakage will be estimated by increasing avoided GHG emissions for all DERs, using an upstream in-state methane leakage adder. This new avoided cost also includes an additional behind-the-meter adder, which will increase the avoided GHG emissions only for those programs which eliminate natural gas appliances from residential buildings. The upstream in-state methane leakage adder has been determined to be 5.57%, and the behind-the-meter adder is 3.78%, based on data from the California Air Resources Board (CARB). These adders take into account both the 100-year global warming potential of methane, which has 25 times the global warming impact of CO2, and different molar mass of CO2 and methane.All methane and refrigerant leakage data are inferred from CARB’s databases, and the leakage rates and adders were reviewed by CARB staff. Major Changes to the Natural Gas Avoided Cost CalculatorThis resolution makes the following Major Changes to the natural gas ACC:Simplify methodology for developing natural gas price forecastUtilize same IRP-based GHG adder as the electric sectorPreviously, natural gas prices were forecasted using NYMEX natural gas futures prices for the most recent 22 days, long-term natural gas forecast using the revised 2019 Integrated Energy Policy Report Mid-Demand case. The method for calculating natural gas price forecasts has been simplified. The ACC natural gas price forecasts will be developed using forward prices for five years, then transition to the California Energy Commission IEPR mid gas price forecast, which is currently used in the IRP proceeding over a three-year transition phase.The Commission will utilize the same IRP-based GHG adder for the natural gas sector as for the electric sector. Additionally, the natural gas ACC will use the cap-and-trade value so that the total per ton value of GHG reductions is that same as that used for electricity. Previously, the Natural Gas Greenhouse Gas Adder relied on the Interim GHG adder from 2017.Minor Changes to the Avoided Cost CalculatorsThe following minor changes will be made to the ACCs:Expand the outputs used for demand responseRemove separate outputs related Permanent Load ShiftingInclude historical year(s) in the ACCsCorrect minor errors in the 2019 Natural Gas ACCFollowing the version-control nomenclature ordered in D. 19-05-019, the new Avoided Cost Calculator posted on May 1, 2020 is ACC_2020_v1a, which replaced the previous version, ACC_2019_v1h. In accordance with D.20-04-010 Energy Division held Webinars on May 6, 7, and 8, 2020 to present the methodology related to the transmission and distribution avoided costs, high GWP avoided cost, generation capacity and energy avoided cost and the GHG adder. Following the Webinars, stakeholders had an opportunity to submit informal questions to Energy Division staff, in addition to formal comments as parties to this resolution. On May 22, 2020, Energy Division staff posted version 1b of the 2020 ACC and provided responses to the post-Webinar questions. Version 1b includes corrections to the calculator based on further internal review, party feedback during the Webinars, and the post-workshop questions. All corrections made to version 1b are listed in the change log of the 2020 ACC spreadsheet at the link in Appendix A. A new version 1c of the 2020 ACC will be posted consistent with this MENTSPublic Utilities Code section 311(g)(1) provides that this resolution must be served on all parties and subject to at least 30 days public review and comment prior to a vote of the Commission. Section 311(g)(2) provides that this 30-day period may be reduced or waived upon the stipulation of all parties in the proceeding. The comment period for this resolution was neither waived nor reduced.On May 28, 2020, five parties submitted comments – The Utility Reform Network (TURN); Solar Energy Industries Association and Vote Solar (SEIA/VS); California Large Energy Consumers Association (CLECA); Southern California Edison Company, Pacific Gas and Electric Company, and San Diego Gas & Electric Company (Joint Utilities); and Southern California Gas Company (SoCalGas). On June 2, 2020, three parties submitted reply comments – SEIA/VS, CLECA, and the Joint Utilities. We respond to each comment below, but first we would like to clarify one aspect of the new Avoided Cost Calculator. Some of the comments, as discussed below, speculate that there are errors in the 2020 ACC due to the IRP scenario used as the basis for the modeling – the “No New DER” scenario, which eliminates all new incremental energy efficiency, demand response, and customer generation & storage as baseload or candidate resources in the IRP RESOLVE model.The No New DER scenario is not meant to be a realistic planning scenario. It is a “what if” scenario – for example, what if the IOUs’ service territories didn’t have energy efficiency, new rooftop solar, or the ability to call demand response? The results of the No New DER scenario tell us that, without DERs, utilities would have to purchase more supply-side resources, import more electricity, and use natural gas plants more often. This means that, as compared with the Reference System Plan, the average heat rate would be higher, since gas generation would be higher. As SEIA/VS point out in their reply comments (p. 2), “The removal of EE raises loads in all hours, and these added loads increase steadily over the 2020-2030 decade. In addition, the removal of BTM PV increases midday loads. Thus, it makes sense that the No New DER results from SERVM show higher market heat rates…as less efficient units are dispatched to serve the higher loads in this scenario.” In addition, mid-day prices would be higher, because there would be less energy efficiency, less solar, and no ability to call demand response. The No New DER scenario results, in and of themselves, point to the value that DERs provide to the electric grid. However, they do not reflect a likely future electric grid, nor were they intended to.Use of the No New DER represents a departure in both modeling methods and conceptualization of avoided costs from previous versions of the ACC. In the past, we have assumed that all DERs avoided the marginal cost of providing electric service, based on the costs of providing an additional kilowatt hour of electricity with the existing electric grid. The new approach looks at the impact of our programs as a whole, and measures the marginal cost of each kilowatt hour of electricity that would have to be provided by a theoretical, “supply-side-only” electric grid. Undoubtedly, there are pros and cons to both these approaches. But, overall we believe that the latter approach more closely reflects the costs that ratepayers would incur in the absence of demand-side programs. The Joint IOUs comment that historical 2019 energy prices listed in the ACC appear to match those provided by the California Independent System Operator (“CAISO”) but are out of order, and that this will affect the scarcity adjustments calculation and forecasted energy and capacity prices. The model uses 2019 CAISO prices. They are out of order to match the CTZ22 weather year for consistent application of avoided costs to weather-dependent DER load shapes. This is why the prices did not match the historical calendar year of 2019 for prices downloaded directly from CAISO. To clarify, the scarcity scaling function is only applied to hourly energy market prices; it does not flow into the hourly marginal grid emissions factor for GHG emissions. The Joint IOUs note that historical gas prices and GHG costs used to develop 2019 historical heat rates are constant in each month. In the 2020 ACC, monthly gas prices are used to calculate historical heat rates, not daily. This is consistent with past ACC calculators, which have always used monthly gas prices for marginal heat rate and price calculations. The Commission understands that the majority of natural gas used for power production is contracted on a monthly basis. We recognize that there is an active daily spot market for natural gas, and generators could buy or sell gas at the daily price, possibly making the daily spot gas price the true marginal cost seen by electric generation. A methodology change to account for this would require further development and review. As this approach has not been taken in past ACC versions, we find it prudent to keep the monthly calculation in the 2020 update. Staff will engage in further discussion to consider changing this in future versions, if warranted. Joint Utilities believe that SERVM runs are resulting in unrealistic operating profiles and energy prices. They argue that almost all of the gas generators, and most other dispatchable units consistently operate in a “forbidden zone” between zero and their listed minimum generation level, and that minimum up and down times for combined cycle units are not always reflected in SERVM outputs. The Joint Utilities believe that these are modeling inaccuracies that are resulting in too-flat energy prices. We believe that this is the result of dispatching over multiple iterations. SERVM respects the operating constraints of power plants, including minimum down times and up times and minimum generation levels. What occurs in the modeling is that in some iterations a specific plant may start in one particular hour and in other iterations (due to outage events or other dispatch differences) the plant starts up an hour later or earlier. Hence, when averaging iterations together, the average of the plant’s dispatch may appear to be less than minimum generating levels. However, in the SERVM model, no individual iteration operates a plant outside its constraints.The Joint Utilities state that during many hours in which the SERVM-modeled energy price is below -$100 per megawatt-hour, all of the batteries with capacities greater than 100 megawatts are neither charging nor discharging, and most of the pumped hydro units are only slowly charging, whereas both batteries and pumped hydro are charging significantly more a few hours earlier when the energy price is low but positive. Energy Division staff reviewed the dispatch data in depth, and found that it appears that storage fills up and empties out quickly. Being only 4 hours in duration, the battery storage charges and discharges before the ramp and evening peak are over. The needs filled by storage partially mitigate pricing. It appears that negative pricing is strongly correlated with storage running out of charging capacity, as it appears storage gets full. The Commission does not find this to be an error.The Joint Utilities state that SERVM outputs show small shortages of regulation down (Reg down) in most hours and the same amount for many hours in a row, but the “supplied Reg down” is often much more than the listed Reg down requirement. We find this observation to be correct. Upon examination, Energy Division staff discovered a reporting error in the SERVM results related to Reg Down Shortages. The Reg Down Shortages are erroneous. We clarify that this error did not affect the calculation of Reg Prices, and only affected the results for Reg Down shortages. Reg Up and other types of services were unaffected by this error.Almost all biogas units have very low “Dispatch price at Max” but are rarely operated near maximum capacity, even when energy prices are very high. Energy Division staff is still examining this aspect of the model to determine why this is occurring and whether there may be an error. The Commission finds that this does not have a significant impact on the results.The storage dispatch model is used to calculate the Energy Gross Margin and marginal capacity prices, supposedly using a four-hour lithium-ion battery. However, the outputs reveal that the battery modeled has a duration of approximately 3.7 hours, not 4.0 required by the Decision. The storage dispatch results, which have previously been shared publicly, clearly show some hours when the state of charge is zero, and some when the state of charge is 4 MWh, indicating a 4 hour duration. However, the battery rarely goes all the way to zero or all the way to 4 MWh. In most hours, the revenue maximizing strategy does not completely empty or charge the battery. Rather, it only does so when energy prices are very low or very high. The battery does this to offer frequency regulation and spinning reserves in most hours. The Commission does not find this to be an error in the model.In their comments, the Joint Utilities propose an algorithm to improve the ACC’s regulation prices. The Commission finds that this methodology provides a more accurate landscape of ancillary services value streams, especially in the diurnal change dimension. We accept this comment and version 1c of the 2020 ACC incorporates updates to the regulation price calculation to reflect the Joint Utilities’ comments. This update includes using a piecewise linear function to predict the fraction of reg up price in the total reg price, based on the current market heat rates. The Joint Utilities also claim that the ACC's GHG avoided cost does not align with the method described in D.20-04-010. They criticized two aspects of this avoided cost, saying that the marginal GHG emissions rates do not accurately reflect a decarbonized generation fleet from the No New DER portfolio modeled in RESOLVE, and that the GHG price implemented in the draft ACC does not align with D.20-04-010, which instructs us to continue to use the straight-line greenhouse gas adder referred to in an earlier IDER Decision. The Joint Utilities argue that the Decision requires the values in the GHG adder to start at the current Cap-and-Trade Price Containment Reserve amount, and increase to the 2030 GHG abatement cost, as was done in the 2019 ACC, rather than use the average utility weighted average cost of capital (WACC) to discount the 2030 GHG abatement cost back to 2020, as recommended in the Staff Proposal.In answer to the first claim, the Commission clarifies that declining average GHG intensity does not directly imply declining marginal GHG intensity. Natural gas would still be on the margin most hours, and could be operating with more variability outside optimal ranges with more ramping and stop/starts which could contribute to higher marginal emissions. We agree with SEIA/VS’s reply comment that “there is no reason that the trajectory over time of marginal GHG emissions (i.e. the emissions associated with the last MWh of electric use) must be the same as the path of system average emissions (i.e. emissions averaged over all MWh used).” SEIA/VS also point out that even with large amounts of solar and storage in the grid, the marginal resource for hourly dispatch can still be an existing natural gas plant, and that “California will continue to rely on significant amounts of in-state gas-fired generation (or imported power that also is likely to be gas-fired) as the marginal resource in most hours.” (SEIA/VS reply comments, p. 3). We reject the Joint Utilities’ argument on this point.As for the latter claim, the Commission finds that this is a mis-interpretation of D.20-04-010, which states (at Ordering Paragraph 2e) that the ACC “shall continue to use the straightline greenhouse gas adder, as adopted in Decision 18-02-016.” The Joint Utilities believe that D.20-04-010 requires continued use of the previous, identical methodology to create the adder, but that is neither stated nor implied. The ordering paragraph quoted by the Joint Utility comments goes on to say “The values in the adder may be modified based on post-2030 data.” D.20-04-010, Section 7.1.4 states “We find that the Staff Proposal is consistent with the Commission’s prior policy adopted in D.18-02-016 [sic], whereby a consistent trajectory to the greenhouse gas adder, without a sharp spike in 2030, will provide a steady market signal…” and “While maintaining the straight-line greenhouse gas adder, as used in the current Avoided Cost Calculator, based on party comment, we authorize staff to consider modifying the adder such that it is based on post-2030 values to better reflect average long-term greenhouse gas abatement costs.”D.20-04-010 clearly intends adoption of a straight line adder – in other words, an adder without a sharp spike, which was deemed appropriate for distributed energy resources by D.18-02-018, used in the 2019 ACC, and as was recommended in the Staff Proposal considered in D.20-04-010. The Decision orders only that staff modify the proposed adder to reflect post-2030 values. Nowhere in D.20-04-010 is it stated or implied that the WACC methodology recommended in the Staff Proposal should be rejected, or that the trajectory or values used to determine the straight-line adder be identical with the previous adder.The Joint Utilities also believe that the No New DER scenario should be altered, stating that the Commission erred in eliminating all energy efficiency and behind-the-meter solar from the No New DER scenario, and that they believe that “naturally-occurring” and Title 24-mandated installations should have been retained in the scenario. D.20-04-010 adopted the No New DER scenario, with the energy efficiency and behind-the-meter solar eliminated. This resolution is not the procedurally appropriate venue to revisit this determination. We reject the Joint IOUs’ request to review this issue here. Joint Utilities believe that marginal transmission cost calculations should use forecast peak load growth from peak demand forecasts made several years prior to the year of each transmission expenditure. We recognize that it is true that the commitment to proceed with transmission projects is made years in advance of the actual construction schedule, based on load forecasts that exist at that time. However, the use of current load forecasts provides a better representation of the magnitude of the peak load reductions that would be required to actually defer such projects going forward. Although the issue of the impact of the lag in the transmission expansion process could be explored further in subsequent avoided cost updates, the Commission finds that no change is warranted at this time.The Joint Utilities state that it is inappropriate to include the Alberhill project when assessing potential transmission costs that may be avoidable in SCE’s territory, because this project is designed to address distribution system load growth and reliability needs in a specific local area, and that including this project in transmission avoided costs will lead to double-counting.The data used to develop the avoided costs of transmission for each utility’s territory were obtained via Energy Division data requests to each utility. It is troubling that SCE is indicating that the data provided specifically to use for estimating transmission marginal costs is incorrect. However, we are also concerned that there may be double counting of the same large investment in both the transmission and distribution data. The SCE distribution data does not provide cost information for sub-transmission projects like Alberhill, therefore the Commission cannot confirm that the investments in the distribution plan are the same as the investments in the transmission plan. Further, the sub-transmission component of distribution capacity costs are derived from SCE’s GRC estimates of marginal costs. Those GRC capacity costs are based on a regression analysis of largely historical data, and it is unclear if Alberhill cost information was used in the distribution analysis. Moreover, even if Alberhill cost information were included in the distribution analysis, it is not clear how much those costs would increase the estimate of SCE’s total distribution sub-transmission marginal costs, given that Alberhill represents only one datapoint in the regression analysis. Absent the confirmation of double counting of the same Alberhill investments in both the transmission and distribution marginal costs we reject the suggested change.The Joint Utilities believe that the calculation of short-term, unspecified distribution avoided costs requires more transparency.The avoided cost calculations for distribution capacity have been provided with extensive detail. However, given the risk of disadvantaging the utility customers via revelation of commercially sensitive information and potential infrastructure security issues, the Commission finds that the current method and process are appropriate, for the time being.The Joint Utilities recommend that the marginal distribution costs calculated in Phase 2 of the utilities’ GRC be de-rated to account for the fact that, over the long-run, utilities are likely to find low- and no-cost solutions to many distribution needs.The Commission rejects the Joint Utilities’ recommendation on this point. We find that a de-rate is not necessary for SCE and SDG&E, as their regression methods already reflect low- and no-cost solutions. The regression method uses ten years of historical distribution expenditures that reflect the utilities’ actual costs. To the extent there are low- and no-cost solutions, such cost savings are already reflected in the regression data and therefore reflected in long-run results. Turning to PG&E’s marginal distribution costs, we find that these forecasts should be reflective of expected projects, not simply a build up from overload amounts. A large portion of the PG&E distribution marginal costs are for projects under $1M. For those projects, historical spending amounts are used to construct the forecasts. In this way, the results already reflect the impact of low- and no-cost solutions. For the larger projects, we understand that PG&E’s marginal costs are based on distribution engineer project estimates and distribution plans, and low- and no-cost solutions should have already eliminated such projects from the provided estimates.SoCalGas objects to the valuation of methane leakage that is part of the new avoided cost of high global warming potential gases. Their main objections are: The Resolution presumes the only viable path to building decarbonization is electrification and suggests that legislation requires replacement of natural gas appliances with electric appliances.Methane leakage on the gas transmission and distribution system is not a function of throughput, so DERs will result in little or no reduction in the reported methane leakage from the natural gas pipeline system.The Resolution rationalizes that eventual building electrification and complete shutdown of the gas distribution system will take place and DER projects should reflect reduced methane leakage based on this hypothetical shutdownIf a DER project only seeks to replace a single piece of gas equipment in a home, the natural gas distribution company will still need to maintain its full system to meet remaining demand in the home and neighborhood.The inclusion of methane leakage associated with the category “Oil & Gas Production and Processing” from the CARB inventory grossly overstates the relationship and impact DER projects could have on this type of methane leakage because the vast majority of natural gas produced in California is associated with oil production.First, the Commission finds that this has already been litigated, and that D.20-04-010 adopted the avoided cost of high global warming potential gases, including adders for methane and refrigerant leakage. We offer the following additional information to clarify the modeling assumptions that were made to develop this new avoided cost. We make no assumptions about the path of the state’s building decarbonization policies. However, it is clear that electrification is, and will continue to be, an important focus of decarbonization efforts, and it is imperative to update the ACC so that it can be used to value those electrification efforts. We find that it is reasonable to assume decreases in natural gas consumption, infrastructure and appliance use will be substantial over the next 30 years, therefore we base the methane leakage values in the ACC on that assumption. However, the upstream methane leakage adder does not assume a complete shutdown of the gas distribution system. Rather, it assumes a partial reduction in the size of the gas distribution system due to avoided new gas system infrastructure and a partial shutdown of the gas distribution system in electrifying neighborhoods. We find it reasonable to assume that an increasing portion of new construction will be all electric, as evidenced by the many “reach codes” already enacted in California cities prohibiting new gas infrastructure, and that electrification programs will result in substantial decreases in natural gas usage.The reduction in natural gas usage, combined with declining levels of new natural gas infrastructure, will impact methane leakage in a number of ways. While the relationship between natural gas throughput and methane leakage is as yet unclear, there is some evidence that reducing throughput in the gas T&D system will lead to a reduction in leakage even if parts of the system are not shut down. In addition, our estimate of the upstream methane adder does not rely solely on that relationship. Rather, it reflects many factors, including possible reduced leakage due to decreased throughput, as well as in-state production decreases, and primarily, the reduced infrastructure discussed above. We understand that there is considerable uncertainty about the path, impacts, and level of decreased natural gas, and therefore methane leakage, reductions, and that our estimates are therefore uncertain. Nevertheless, we are certain that state policy demands considerable reduction in natural gas consumption and many strategies to reduce methane leakage. The Commission does not believe that we should delay valuing efforts to reduce a significant source of GHG emissions simply because of the uncertainty around methane leakage reduction.It is certainly true that, at an individual appliance level, the natural gas distribution company will still need to maintain its full system to meet remaining demand. However, the methane leakage adder is designed to reflect the long-run avoided methane leakage as many buildings are electrified, not just the avoided leakage from electrifying one appliance. This is the same justification as for the Avoided Generation Capacity avoided cost, wherein one efficient lightbulb is not assumed to result in less generation capacity being online, but rather the aggregate effect over many programs is “averaged out” over time, location, and installation, and assigned to each individual DER.It is true that much of the natural gas in California is associated with oil production, and thus it is not possible to say how much leakage at these wells will be avoided by reducing oil consumption as opposed to by reducing gas consumption. However, as the state seeks to reach its decarbonization goals, both oil and gas demand in the state will decline, meaning leakage at these wells will be reduced, particularly as a result of avoided new wells. We note also, as did SEIA/VS in their reply comments (p. 5), that electric vehicles are also a distributed energy resource. Therefore, in developing the upstream methane leakage adder, we only assumed that reduced natural gas consumption as a result of DERs will result in some avoided methane leakage at oil and gas wells in the state, not that leakage will go to zero or that all oil and gas extraction in the state will stop.TURN offers similar objections to the upstream methane leakage adder, stating that including any upstream leakage benefit due to a planned retirement of the gas distribution system is premature, because that would necessitate a fundamental transformation of the “obligation to serve” model. TURN also argues that the comparison to generation capacity is flawedThe upstream methane leakage adder does not depend on a fundamental transformation of the “obligation to serve” model in that it does not assume a complete shutdown of the natural gas distribution grid or a refusal to serve customers who would like to connect to the gas grid. Rather, as discussed above, it reflects the reduced methane leakage that will result from avoided new gas distribution infrastructure and a partial shutdown of the gas distribution system in neighborhoods that have electrified. Gas utilities do not have an obligation to serve all-electric customers, so a partial shutdown of the gas distribution grid in neighborhoods that have fully electrified is not in conflict with the existing “obligation to serve” legal framework. TURN argues that the parallel with avoided generation is flawed because there “are no regulatory or policy barriers for demand reductions caused by DERs to flow through as actual avoided generation capacity costs, even if the actual avoided costs are lumpy and not continuous,” (TURN comments, p. 4) whereas the natural gas utilities’ obligation to serve is a regulatory barrier that prevents the realization of any potential avoided costs related to decreased natural gas consumption. The Commission is unpersuaded by this argument. First, this ignores the value of avoided new infrastructure, as new construction will increasingly be all electric. Second, as stated above, there is no obligation to serve all-electric customers, so a partial shutdown of the gas distribution grid in neighborhoods that have fully electrified is reasonable to assume.SEIA/VS argue that the trajectory of the GHG Adder should be set such that it has the same 30-year net present value as the RESOLVE GHG Adder from the IRP. The Commission finds this idea intriguing and one that should be further discussed in future. Since the post-2030 RESOLVE GHG value is currently a single point showing an extremely high value, we are hesitant to use that value as anything other than an indicator that avoiding GHG will be increasingly expensive in later years. SEIA argues that the proposed portfolio re-balancing adjustment to hourly GHG emissions is appropriate only for DERs that change end-use loads and the GHG intensity of the electric sector. The Commission finds that it is appropriate for all DERs. All DERs could change loads and GHG intensity, and future capacity needs, since modeling is based on current forecasts, not actual load. Portfolio rebalancing is targeted to the average grid intensity, which we calculate with the denominator of retail sales (not end-use loads), per SB100. All customer behind-the-meter resources affect retail sales. SEIA believes that PG&E’s avoided transmission costs should be $30 per kW-year and that SDG&E’s avoided transmission costs should be $22 per kW-year. While the GRC is for a three year period, the transmission marginal costs calculated by PG&E are for a six year period, not the three-year period SEIA assumes. Of those six years, the investment costs in the first year and the last year are zero, which does raise the questions of whether it is appropriate for PG&E to include the first year of load growth. If the investment cost was excluded because 2020 is no longer applicable, then the load growth for 2020 should also be excluded. This would increase the marginal cost. Similarly, if the last year 2026 investment is zero because the investment plan does not extend that far, then the load growth for that year should also be excluded which would also increase the marginal cost. Given that the comments and data responses do not detail the reasons for the exclusions, the Commission does not find it reasonable to change PG&E’s marginal transmission costs at this time. SEIA/VS argues that some additional transmission projects should be included for SDG&E in the calculation of the transmission avoided cost. Determination of whether such projects should be added to what the utility considers to be deferrable peak load driven projects can be explored in the next avoided cost update, but there is insufficient information to make such a change at this time. SEIA also argues that the cumulative peak load growth used to estimate the transmission marginal costs should include the negative growth in 2020. We reject this recommendation because including the negative 2020 value would result in an underestimation of the growth that is forecast for the years that have investments (2021 through 2024). This would artificially increase the transmission marginal cost.SEIA believes that transmission peak capacity allocation factors (PCAFs) are the most direct measure of peak transmission system loads and should be used to allocate avoided transmission costs to hours of the year, similar to the use of distribution PCAFs to allocate avoided distribution costs. We agree with SEIA, and adopt this change. In version 1b of the 2020 ACC, transmission capacity costs were allocated using the same EUE-based allocation factors as generation capacity. While those allocation factors reflect probabilities of peak loadings, they also reflect generator output patterns that, while impacting the need for generation additions, are not the driver of the need for transmission capacity to serve load. The PCAF method proposed by SEIA derives transmission allocation factors using IOU system level loads. The use of PCAFs is well established in the ACC for allocating distribution capacity costs, and we find it reasonable to extend that method to transmission capacity costs. Version 1c of the 2020 ACC incorporates this change.SEIA/VS states that the significant GHG impacts of out-of-state methane leakage associated with natural gas burned in California should be considered as a societal benefit when DER programs are evaluated. The Commission find this comment to be out of scope here, as the ACC is not a cost-effectiveness tool.SEIA/VS states that the 2.2 % per year escalation of natural gas transportation rates in the Natural Gas ACC has not been demonstrated to reflect recent trends for the increases in these rates and that the Resolution therefore requires modification. We agree that this issue deserves more scrutiny. However, this modification would make the Natural Gas and Electric ACCs different, which would result in inconsistent valuations of program benefits. In addition, because the values used in the Electric ACC come from the California Energy Commission (CEC) IEPR, we believe this issue should be explored in that proceeding.CLECA argues that the capital costs of the four-hour storage battery used as the proxy unit for avoided generation capacity costs are under-estimated. They believe that it is unlikely that the cost of batteries will decline in the future and that important costs necessary for battery storage operation are missing The Commission relied on the IRP proceeding for battery cost inputs, which assumes declining costs over time. There are many reputable forecasts that have battery technology costs declining over 30 years, although at a lower rate in the future. Public price forecasts for battery technology (and indeed for most generation technologies) do most often omit some site-specific costs, but we have nevertheless relied on them for proxy generation costs. We find no reason to think the IRP costs are missing significant costs relative to what has been used in the past for natural gas turbine proxy costs. CLECA believes that the “energy rents” are over-estimated because the model assumes perfect foresight of the energy market, and has not been calibrated to current observed battery behavior. “Energy rents,” or “gross margins,” are the estimates of revenues that a generator can earn. The energy rents are subtracted from the capital costs of the generator proxy unit – a storage battery, in this case – to determine the residual capacity value to be used to determine the avoided cost of generation capacity. If energy rents are overestimated, then the avoided cost of generation capacity will be underestimated.It is true that the ACC assumes perfect foresight of the energy market in the model. Modeling in past ACC cycles of gas turbines similarly did not model expected economic losses from imperfect dispatch. Revenue modeling for proxy generation capacity resources has always been done at a relatively high level, as more refined modeling is assumed to provide a false level of precision. The Commission does not find it appropriate to calibrate results from the No New DER case to observed historical results. Revenue modeling done for the counterfactual No New DER case using SERVM prices, and as discussed above, is expected to produce different results than actual historical prices. We note that historic frequency regulation prices are much higher than those determined by SERVM (even with the adjustments suggested by PG&E). Based on Energy Division staff’s review, we observe historical dispatch to earn more than 70% of revenues in frequency regulation. This is much lower using SERVM results based on No New DER prices, where less than 40% of revenue is from frequency regulation. CLECA believes that there are errors in the IRP modeling that are affecting the energy and generation capacity avoided costs. CLECA argues that IRP’s system dispatch results violate resource operating restrictions, and that the energy prices are inaccurate because they increase over time, despite the transition to resources with near-zero variable costs. The Commission finds this to be a misunderstanding of how the SERVM model works, as addressed above in response to the Joint Utilities’ similar comment,CLECA believes that the annual values for avoided generation capacity have been incorrectly allocated to hours of the year, partially because of the use of a single weather year. The weather methodology was adopted in D.20-04-010 and cannot be changed by this Resolution. CLECA is welcome to offer further discussion of this in future ACC updates. In addition, we clarify that the allocation method we used – the RECAP model – is the same model that has been used in the past, as previously adopted. CLECA objects to use of the Discounted Total Investment Method (DTIM) to determine avoided transmission capacity costs, because it is not a universally-accepted marginal cost approach. However, this issue has also already been litigated, as use of DTIM was adopted in D.20-04-010. In addition, we note that nowhere is it claimed that the DTIM is universally-accepted. Universal or exclusionary adoption is not necessary for a method to be valid.In summary, the following changes were made to the ACC based on party comments. First, avoided transmission costs will be allocated to hours of the year using transmission PCAFs. Second, the algorithm proposed by the Joint Utilities will be used to improve the ACC’s regulation prices.FINDINGSD.20-04-010 directs Commission staff to update the Avoided Cost Calculator within 30 days of its issuance on April 16, 2020.D.20-04-010 OP 7 directs Commission staff to make major changes to the Avoided Cost Calculator, as specified in that Decision.D.19-05-019 OP 11 directs Commission staff to make corrections, data updates, and minor changes.The updates to the Avoided Cost Calculator as described by staff’s consultant E3 in its Avoided Cost Calculator spreadsheet and documentation are reasonable for use in DER cost-effectiveness. It is reasonable to adopt this 2020 Avoided Cost Calculator, specifically referred to as ACC_2020_v1a.It is reasonable for ACC_2020_v1a to adjust the generation capacity value to reflect the Net Cost of New Entry of new battery storage.It is reasonable for ACC_2020_v1a to adopt the No New DER Scenario as the counterfactual and use the resulting data to model hourly avoided costs.It is reasonable for ACC_2020_v1a to estimate the hourly avoided energy and ancillary services costs using production cost modeling.It is reasonable for ACC_2020_v1a to adopt the system-average approach for modeling unspecified distribution avoided costs.It is reasonable for ACC_2020_v1a to adjust the current approach for calculating transmission avoided costs to include costs for SCE and SDG&E based on the PG&E’s GRC method. It is reasonable for ACC_2020_v1a to adjust the straight-line GHG adder to utilize post-2030 values.It is reasonable for ACC_2020_v1a to adopt the short-and-long term greenhouse gas modeling approach described herein.It is reasonable for ACC_2020_v1a to adopt the simplified methodology for developing natural gas price forecasts described herein.It is reasonable for ACC_2020_v1a to utilize the IRP-based GHG adder used for the electric sector for the natural gas sector.It is reasonable for ACC_2020_v1a to make all corrections described in this resolution.THEREFORE IT IS ORDERED THAT:The updates to the Avoided Cost Calculator as specified herein and further enumerated in documents made available through Appendix A of this Resolution are adopted for use in demand-side distributed energy resource cost-effectiveness analyses.This Resolution is effective today.I certify that the foregoing resolution was duly introduced, passed and adopted at a conference of the Public Utilities Commission of the State of California held on June 25, 2020; the following Commissioners voting favorably thereon: _______________ALICE STEBBINSExecutive DirectorAppendix AAvoided Cost Calculator 2020 Update documents available online:2020 Avoided Cost Calculator ACC_2020_v1c, the 2020 Natural Gas Avoided Cost Calculator, the Avoided Cost Calculator 2020 Documentation version 1c, and related data files are all available for download on this site: (scroll down to Avoided Cost Calculator section)As a backup, these documents are also temporarily available here: ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download