Application No.: A Exhibit No.: SCE-1, Vol. 3. S. DiBernardo S. Samiullah D. Hopper G. Golden (U 338-E) Before the

Transcription

1 Application No.: A Exhibit No.: SCE-1, Vol. Witnesses: R. Thomas S. DiBernardo S. Samiullah D. Hopper G. Golden (U -E) SOUTHERN CALIFORNIA EDISON COMPANY S (U - E) TESTIMONY IN SUPPORT OF ITS APPLICATION FOR APPROVAL OF ITS 01 0 DEMAND RESPONSE PROGRAMS: VOLUME PROGRAM INCENTIVE DEVELOPMENT / COST-EFFECTIVENESS ANALYSIS / PROGRAM ENROLLMENT AND LOAD IMPACT FORECASTS / REVENUE REQUIREMENT AND COST RECOVERY Before the Public Utilities Commission of the State of California Rosemead, California January 1, 01

2 SCE s Testimony in Support of its Application for Approval of its 01 0 Demand Response Programs: Volume Program Incentive Development / Cost-Effectiveness Analysis / Program Enrollment and Load Impact Forecasts / Revenue Requirement and Cost Recovery Table Of Contents Section Page Witness I. INTRODUCTION...1 II. III. CONSOLIDATION OF DR PROGRAMS AND TARIFFS WITH DR INCENTIVES... A. Inputs to DR Program Incentive Calculation A-Factor.... B-Factor.... Program Valuation... B. Incentive Rate Design... C. Results of DR Program Incentive Calculations Schedule TOU-BIP.... Schedule TOU-AP-I.... Schedule SDP DR PROGRAM REVENUE REQUIREMENTS AND COST RECOVERY...1 A. Purpose and Overview...1 B. Cost Reasonableness...1 C. Ratemaking of DRP Funding...1 D. Balancing Accounts Associated with this Application DRPBA...1. BRRBA...1 R. Thomas S. DiBernardo i

3 SCE s Testimony in Support of its Application for Approval of its 01 0 Demand Response Programs: Volume Program Incentive Development / Cost-Effectiveness Analysis / Program Enrollment and Load Impact Forecasts / Revenue Requirement and Cost Recovery Table Of Contents (Continued) Section Page Witness. PAACBA...1 E DR Program Revenue Requirements...1 IV. ANTICIPATED 01 0 DEMAND RESPONSE PROGRAMS ENROLLMENT AND LOAD IMPACT FORECASTS...1 A. Background...1 B. Demand Response Enrollment Forecasts Program Enrollment Forecasts...1 a) Agricultural & Pumping Interruptible...1 b) Base Interruptible Program...1 c) Summer Discount Plan Program...1 d) Capacity Bidding Program...1 e) Save Power Day (PTR)...1 f) Permanent Load Shifting...1 C. Demand Response Load Impacts General Approach...1. Overview of Evaluation Methods...1 a) Individual Customer Time Series Regressions...0 b) Aggregate Time Series...1 c) Panel Regressions...1 S. Samiullah ii

4 SCE s Testimony in Support of its Application for Approval of its 01 0 Demand Response Programs: Volume Program Incentive Development / Cost-Effectiveness Analysis / Program Enrollment and Load Impact Forecasts / Revenue Requirement and Cost Recovery Table Of Contents (Continued) Section Page Witness D. Ex Post Load Impacts Event Averages by Program... E. Ex Ante Load Impacts DR Portfolio Resources by Forecast Year... V. COST-EFFECTIVENESS... D. Hopper/G. Golden A. Overview DR Protocols.... Avoided Cost... B. Net Present Value and Benefit-Cost Ratio Results by SPM Test... C. A-G Factors and Results A-Factor.... B-Factor.... C-Factor.... D-Factor... a) Right Certainty... b) Right Time...0 c) Right Availability...0 d) Right Place...1. E-Factor... iii

5 SCE s Testimony in Support of its Application for Approval of its 01 0 Demand Response Programs: Volume Program Incentive Development / Cost-Effectiveness Analysis / Program Enrollment and Load Impact Forecasts / Revenue Requirement and Cost Recovery Table Of Contents (Continued) Section Page Witness. F-Factor.... G-Factor... D. Administrative Costs... E. Incentive Costs... F. Net Bill / Revenue Reductions... G. Capital Cost to LSE and Amortization Period... H. Capital Cost to Participant and Amortization Period... I. CAISO Market Participation Revenue... J. Costs from Other Proceedings... K. Qualitative Analysis Social Non-Energy Benefits and Costs... a) Job Creation... b) Environmental Impact.... Utility Non-Energy Benefits and Costs.... Participant Non-Energy Benefits and Costs... a) Reasons for Participating... b) Spillover Effects... c) Snapback Impact...0 d) Loss of Service Transaction Patterns...0. Market Non-Energy Benefits and Costs...1 iv

6 SCE s Testimony in Support of its Application for Approval of its 01 0 Demand Response Programs: Volume Program Incentive Development / Cost-Effectiveness Analysis / Program Enrollment and Load Impact Forecasts / Revenue Requirement and Cost Recovery List Of Tables Table Page Table II-1 SCE Proposed TOU Periods... Table II- Current and Proposed BIP Credits... Table II- Current and Proposed AP-I Credits... Table III- Proposed Demand Response Funding for Table IV DR Portfolio Enrollment Forecast (August)...1 Table IV Ex Ante Load Impacts by DR Program August 1-in- SCE System Peak Conditions... Table V- Avoided Capacity Cost By Year... Table V- Cost-effectiveness Ratios from E s DR Reporting Template... Table V- A-Factor Results... Table V- B-Factor Criteria... Table V- B Factor Results... Table V-1 Right Certainty Criteria... Table V-1 Right Certainty Results...0 Table V-1 Right Place Results...1 Table V-1 D-factor Results... Table V-1 G Factor... 1 v

7 SCE s Testimony in Support of its Application for Approval of its 01 0 Demand Response Programs: Volume Program Incentive Development / Cost-Effectiveness Analysis / Program Enrollment and Load Impact Forecasts / Revenue Requirement and Cost Recovery List Of Figures Figure Page Figure IV-1 Summary of Ex Post and Ex Ante Analysis, Process, and Connections vi

8 I. INTRODUCTION Exhibit SCE-01, Volume is comprised of five chapters and sets forth SCE s 1 approach to DR program incentive development, cost-effectiveness analysis, DR program enrollment and load impact forecasts and revenue requirement and cost recovery proposals. Chapter I provides an overview of Volume and describes its organization. Chapter II describes SCE s approach to determine the DR program incentives consistent with the Guidance Decision. Chapter III identifies SCE s 01 through 0 revenue requirement based on the DR program costs forecast in Volume and describes SCE s cost recovery proposal for these costs utilizing existing balancing account mechanisms. Chapter IV identifies SCE s forecast DR program enrollments and load impact forecasts and Chapter V sets forth SCE s cost-effectiveness analysis demonstrating that SCE is proposing a cost-effective DR portfolio. 1 Acronyms and abbreviations are defined only the first time they are used in SCE-01, and are not re-defined in each volume. For a complete list of all acronyms and abbreviations used throughout all volumes of SCE s direct testimony, please see SCE-01, Vol., Appendix B. 1

9 II. CONSOLIDATION OF DR PROGRAMS AND TARIFFS WITH DR INCENTIVES The purpose of this Chapter is to set forth SCE s proposal to consolidate the development of the DR program incentives with the development of the DR program costs. The Guidance Decision instructed SCE to propose its DR program incentives in this application, rather than in the GRC rate design phase (Phase ) as has been the practice up to now. Specifically, this Chapter describes how SCE has consolidated the DR programs (BIP, AP-I, and SDP) that have been the subject of previous DR funding applications with the data and process used to determine DR incentives that have been developed in the GRC rate design proceedings. A key principle for incentive development that SCE has used in the rate design proceeding is that accurate incentive design should convey the proposed benefit a resource may provide in avoided cost, while accounting for the constraints with which such a resource may be deployed. The avoided cost of generation capacity is a key determinant of value when pricing incentives for DR programs. As further discussed below, SCE proposes to use the Avoided Cost Methodology ( A times B methodology) to determine program incentives for the 01 through 0 program cycle. The initial input for this methodology is an avoided generation capacity value of $ per-kw-yr., as adopted in SCE s 01 GRC Phase Settlement agreement. Using the A times B methodology, the avoided generation capacity value is adjusted through the use of A and B factors that compare the program resource value to a proxy combustion turbine (CT) resource. The proposed DR incentives will be implemented at the expiration of DR incentives agreed to in SCE s 01 GRC Phase Settlement Agreements. See D.1-0-0, OP. A , Medium and Large Rate Group Rate Design Settlement Agreement, dated October, 01, p. A-1. The Settlement Agreement was approved by the Commission in D

10 A. Inputs to DR Program Incentive Calculation 1. A-Factor The A-factor adjusts for event duration, call frequency and total callable program hours. This method has been consistently applied to incentive design for DR programs since SCE s 00 GRC Phase proceeding, and appropriately values the efficacy of the different DR programs in comparison to a hypothetical marginal generation resource, which, for SCE has typically been defined as a CT proxy. For example, a DR program with unlimited calls and unlimited frequency of occurrence would have an A-factor of 1, reflecting that a program with these parameters should have a comparable resource value as the CT Proxy. SCE uses a Loss of Load Expectation (LOLE) model to determine which hours in the year SCE s grid is most likely to face a 1-in- capacity shortfall event. SCE uses the 1-in- standard as it is a common reliability metric that strikes a balance between procuring enough capacity that shortfall events are uncommon, but not so much that too much money is spent ensuring reliability. The LOLE model performs a Monte Carlo simulation on load, wind, and solar and compares the resulting net load to the available resources with their associated outages to determine the likelihood of a capacity shortfall event on any given hour in a month. This is performed for both peak and ramp capacity since SCE system shortfall could be a result of a lack of generation serving peak load or ramp in the future. The final output of the LOLE model is a dataset that represents the hours in the year with the highest probability of a 1-in- shortfall event occurring. SCE has adopted the LOLE model to account for shortfall events caused by simulated peak and ramp system constraints. The increased proliferation of renewable resources in California is expected to gradually exacerbate the effect of the duck curve phenomenon on the system, with the potential of capacity constraints at the time of peak system load and peak ramp load. The LOLE framework described above is an attempt to allocate the joint capacity cost of a single flexible resource (in this case the CT proxy) to expected hours of both peak and ramp Adopted in D A proxy resource is defined here as an SCE-owned CT in the Southern California region.

11 system needs. SCE plans to propose a similar framework in the 01 GRC Phase application that will be filed on June 1, 01. In its Rate Design Window (RDW) proposal for new TOU periods, SCE included a more comprehensive discussion that describes the need to account for both peak and ramp constraints on the system. The resulting LOLE is an input when determining the individual A-Factors assigned to specific DR programs. The A-Factor represents the ability of DR programs to cover the expected 1-in- capacity shortfall events resulting from the LOLE simulation. In the rate setting application, a higher A-Factor will occur if a DR program can mitigate more events based on the program s duration and frequency of availability. For example, if Programs X and Y can be dispatched twice in a year, but Program X is available for three hours longer, then Program X will likely have a higher A-Factor than Program Y.. B-Factor The B-factor represents the value of a DR program to be dispatched with day-of versus day-ahead notification. A DR program with day-of notification has a B factor value of 1, and DR programs with day-ahead notification would be assigned a value of less than 1. In addition to the above two factors, the avoided capacity cost is adjusted for the fact that DR programs are counted toward SCE s Planning Reserve Margin (PRM) in SCE s RA filings. SCE applies a 1 percent reserve valuation adder to the avoided cost to accurately reflect the PRM value. For SDP, a distribution reliability adder of $.1per-kW-yr. is also included in the avoided cost calculation to account for the distribution-related right place, right time capability of the program. The methodology described above can be formulaically expressed as listed below. Equation (1): A , filed September 1, 01. A , SCE s 01 GRC Phase proceeding, filed June 0, 01, SCE-0, p.. Errata filing September 1, 01.

12 DR Program Avoided Cost = Avoided Generation Capacity Cost (A B + PRM) + Distribution Reliability Adder. Program Valuation In Chapter IV of this Volume SCE describes the method and criteria used when estimating ex ante expected demand for DR programs. For the purposes of incentive valuation, SCE is using ex ante estimates that are forward-looking and reflect the potential load reduction given a 1-in- peak capability scenario of DR program dispatch. SCE s A-factor used in the incentive design values the efficacy of DR programs for a potential 1-in- system reliability event. Aligning the estimate of expected demand with the same 1-in- criteria allows a consistent application of the valuation criteria used when determining overall incentives at the program level. Annual DR Program incentives are calculated using the following formula. Equation (): DR Program Annual Incentive Value = DR Program Avoided Cost Expected Demand B. Incentive Rate Design When designing the incentive rate structure for DR programs, where possible, SCE attempts to align the structure of the incentive with the applicable retail tariffs specific to the rate group. Such alignment serves two purposes: (1) eases the customer understanding of the incentive as it mimics the tariff structure, and () promotes equity on a per-customer basis in a manner that aligns incentive payouts with the recovery of generation costs in rates. Line Losses and TOU periods form essential components when designing incentive structure for DR programs, and aligning these structures with the retail rates. Line Losses For customers taking service at different service voltages, the avoided generation-level cost value is adjusted to reflect the estimate of line losses for each service voltage. In this proceeding, SCE Because estimated load reductions are measured at line end voltages, the actual avoidable generation level capacity is somewhat higher.

13 1 is using line losses adopted in SCE s 01 GRC Phase proceeding. The resulting DR program avoided costs, adjusted for service voltage line loses, are derived using the following formula. Equation (): Adjusted DR Program Avoided CostService Voltage = DR Program Avoided Cost (1 + Line LossesService Voltage) Time of Use Periods TOU definitions used in the design of the incentive rate structure for each DR program are based on SCE s proposed TOU period studies submitted in the RDW proceeding. If needed, SCE intends to submit updated designs for the proposed incentive structures in the 00 mid-cycle review, in order to align incentive rate structures with new TOU period rates scheduled for implementation as part of SCE s 01 GRC Phase proceeding. Table II-1 shows the TOU periods SCE proposed in the 01 RDW proceeding. Table II-1 SCE Proposed TOU Periods Proposed TOU Periods Description Summer On Peak Summer Mid Peak Summer Off Peak Period Clock Time Weekdays: pm to pm Weekends: pm to pm Weekends and Weekdays: All hours excluding pm to pm Winter Mid Peak Winter Off Peak Winter Super Off Peak Weekdays and Weekends: pm to pm Weekdays and Weekends: pm to am Weekdays: am to pm A , SCE s 01 GRC Phase proceeding, filed June 0, 01, SCE-0. Errata filing September 1, 01. A , filed September 01, 01.

14 In summary, when designing the rate structure for DR incentives, SCE follows a methodical framework of first valuing annual program incentive levels based on DR program avoided capacity cost and subsequently conveying this value to program participants based on a structure that is aligned with the default applicable tariff and TOU periods. Because all customers benefit from the interruptible and SDP programs, the cost of credits paid to participating customers is recovered through an energy surcharge imposed on all customers. Pursuant to D , the surcharges paid by all customers for the recovery of the costs of the interruptible and SDP programs are to be reflected in the distribution component of customer rates. In order to make sure that participating customers continue to pay for delivery-related costs, SCE will continue to limit the overall program credits provided to the distribution energy component of the customer s otherwise applicable tariff, by applying overall credit caps associated with SCE s currently tariffed dual enrollment rules. C. Results of DR Program Incentive Calculations 1. Schedule TOU-BIP BIP is a voluntary reliability DR program for commercial and industrial customers with demand above 00 kw. BIP customers receive year-round monthly bill credits in exchange for allowing SCE to temporarily interrupt service to their elected minimum operational requirement, or FSL, during program events. The BIP program has two options, a 1-minute and a 0-minute response time. SCE proposes to adjust BIP incentives to account for the avoided cost of having to procure Local RA for the BIP-1 option. As previously discussed in Volume 1 of this Exhibit, the CAISO has specific rules that govern how resources can be counted toward meeting local capacity needs. The characteristics required by CAISO for a resource to be counted towards local capacity requirements include: 1) ability to dispatch within 0 minutes, or ) sufficient availability for predispatch. There is still uncertainty as to what each qualification option means. For example, it is unclear if the 0 minute response time is based on nominal program terms (e.g. BIP-1 fully qualifies, D , OP 1 at p..

15 BIP-0 does not at all), or on actual program performance (i.e. MW expected to be delivered within 0 minutes which might include the ramp-down of BIP-0 customers who drop the majority of their load within 0 minutes as measured by meter data). While there is still uncertainty regarding each qualification, SCE is preparing for strict enforcement of the requirements. Assuming the 0-minute requirement is based on nominal program terms, the BIP-0 option may not count towards CAISO local capacity requirements depending on how the pre-dispatch criteria are finalized. Therefore, if BIP-1 provides local capacity and BIP-0 does not, then the incremental value provided by BIP-1 should be reflected in the incentive determination. To reflect the CAISO rules for local capacity resources, SCE herein proposes to promote the BIP 1-minute option by increasing its incentive by percent while decreasing the BIP 0-minute option incentive by percent to maintain a constant overall program incentive value. The asymmetric adjustment between the programs is due to the larger amount of expected demand available under the BIP 0 program in comparison to the BIP 1 program. The percent estimate, applied as an adder to the BIP 1 incentive, accounts for the program s eligibility in meeting CAISO s local capacity requirements. This incentive differential can be scaled down if BIP-0 received partial or full local capacity credit. Table II- summarizes the current and proposed BIP incentive level and structure by Option and Service Voltage. If necessary, SCE intends to submit updated designs of proposed incentive structures in its 00 mid-cycle review in order to align incentive rate structures with the new TOU period rates ultimately adopted in SCE s 01 GRC Phase proceeding.

16 Table II- Current and Proposed BIP Credits Program Option/Service Voltage Level Summer On Peak Summer Mid Peak Winter Mid Peak 1 minute Option Secondary Service (below kv) (.1) (.) (1.0) Primary Service ( to 0 kv) (.) (.) (1.0) Sub transmission Service (above 0 kw) (1.) (.) (1.0) 0 minute Option Secondary Service (below kv) (1.) (.) (1.) Primary Service ( to 0 kv) (1.0) (.) (1.) Sub transmission Service (above 0 kw) (1.) (.) (1.1) * Current TOU periods 1) On Peak: Summer: Noon to :00 p.m. weekday except holiday ) Mid Peak: Summer: :00 a.m. to Noon and :00 p.m. to :00 p.m. weekday except holiday Winter: :00 a.m. to :00 p.m. weekday except holiday ) Off Peak: Summer and Winter: All other hours * Current BIP Credit $/Average Demand ($/kw a ) Program Option/Service Voltage Level Summer On Peak Summer Mid Peak Summer Off Peak Winter Mid Peak 1 minute Option Secondary Service (below kv) (1.) (1.0) (0.1) (.1) Primary Service ( to 0 kv) (1.0) (1.) (0.1) (.) Sub transmission Service (above 0 kw) (1.) (0.) (0.) (.) 0 minute Option Secondary Service (below kv) (1.) (1.) (0.1) (.) Primary Service ( to 0 kv) (1.1) (1.) (0.1) (.) Sub transmission Service (above 0 kw) (1.1) (0.1) (0.) (.) * Proposed TOU periods 1) On Peak: Summer: p.m. to p.m. Workday ) Mid Peak: Summer: p.m. to p.m. Weekends and Holiday Winter: p.m. to p.m. Workday ) Super Off Peak: Winter: a.m. to p.m. Weekend, Workday and Holiday ) Off Peak: Summer and Winter: All other hours * Proposed BIP Credit $/Average Demand ($/kw a ) 1. Schedule TOU-AP-I In Volume of this Exhibit, SCE describes the introduction of a FSL level option within the AP-I program. Incentive levels being proposed herein include AP-I participants enrolled under either the CLC option or the traditional means of AP-I participation where customers utilize a DLC device to participate in the program. Similar to the current BIP incentive structure, participants enrolled in the AP-I CLC option using their own devices to control load drop during events (e.g., CLC devices

17 that operate in larger customer controlled SCADA systems), would pay an excess energy charge for consumption above their elected FSL during an event. Table II- below illustrates current and proposed AP-I incentives. If necessary, SCE intends to submit updated designs of proposed incentive structures in the 00 mid-cycle review in order to align incentive rate structures with the new TOU period rates ultimately adopted in SCE s 01 GRC Phase proceeding. Table II- Current and Proposed AP-I Credits * Current AP I Credit $/Average Demand ($/kw a ) All Service Voltage Levels Summer On Peak Summer Mid Peak Winter Mid Peak TOU AP I (1.) (.1) (1.) * Current TOU periods 1) On Peak: Summer: Noon to :00 p.m. weekday except holiday ) Mid Peak: Summer: :00 a.m. to Noon and :00 p.m. to :00 p.m. weekday except holiday Winter: :00 a.m. to :00 p.m. weekday except holiday ) Off Peak: Summer and Winter: All other hours * Proposed AP I Credit $/Average Demand ($/kw a ) All Service Voltage Levels Summer On Peak Summer Mid Peak Summer Off Peak Winter Mid Peak TOU AP I (1.1) (0.) (0.) (.) * Proposed TOU periods 1) On Peak: Summer: p.m. to p.m. Workday ) Mid Peak: Summer: p.m. to p.m. Weekends and Holiday Winter: p.m. to p.m. Workday ) Super Off Peak: Winter: a.m. to p.m. Weekend, Workday and Holiday ) Off Peak: Summer and Winter: All other hours 1. Schedule SDP SDP is a voluntary program that operates as both a reliability and price responsive program. The SDP program offers credits to customers who allow their A/C units to cycle off during curtailment events. Participating customers receive a credit on their electric bills each year from the first of June to the first of October. Program credits are based on the customer s air conditioner tonnage and the program cycling option. The 01 ex ante demand for a 1-in- load scenario is used for the

18 program s expected demand. There is an observable difference between the estimated annual incentive level for the proposed SDP program ($. million) and the estimated level of incentives currently being paid to program participants ($.1 million). The difference in overall incentives is primarily due to the significant reduction in the program s expected demand, as explained in Volume of this Exhibit. In order to mitigate the impact to participating customers, SCE is proposing a five year transitional period during which annual incentive payouts to participants will incrementally converge toward a more appropriately-valued incentive level for the program. Based on this -year transition period, annual incentive levels for the program are expected to reduce by $. million per year for residential participants and $ million per year for commercial participants. Table II- illustrates the current SDP incentive and the proposed annual incentive for residential and commercial customers. The incentives are expressed in dollars per-ton-per-summer day for each of the five transitional years. Table II- Current and Proposed SDP Incentives Current $ per tonper Summer Proposed $ per ton per Summer day SDP Credit day Residential SDP Credit Year 1 Year Year Year Year 0% Cycling with Override Option (0.00) (0.0) (0.0) (0.0) (0.0) (0.0) 0% Cycling with Override Option (0.) (0.1) (0.1) (0.) (0.) (0.0) 0% Cycling with No Override Option (0.) (0.1) (0.1) (0.) (0.) (0.0) 0% Cycling with No Override Option (0.0) (0.) (0.) (0.) (0.0) (0.1) Current $ per tonper Summer Proposed $ per ton per Summer day SDP Credit Commercial day SDP Credit Year 1 Year Year Year Year 0% Cycling Option (0.0) (0.0) (0.0) (0.01) (0.01) (0.01) 0% Cycling Option (0.1) (0.1) (0.) (0.0) (0.0) (0.00) 0% Cycling Option (0.1) (0.) (0.1) (0.0) (0.1) (0.1)

19 III DR PROGRAM REVENUE REQUIREMENTS AND COST RECOVERY A. Purpose and Overview The purpose of this Chapter is to set forth SCE s revenue requirement for the period 01 through 0 for the DR portfolio described in Volume of this Exhibit. This Chapter also sets forth SCE s cost recovery proposal for that revenue requirement. In this application, SCE is requesting DR program funding in the amount of $1. million over the 01 0 funding cycle. 1 As such, SCE proposes to decrease customer rate levels to fund the DR Program budgets requested in this application. In OP 1 of D , the Commission capped SCE s annual DR portfolio expenditures for 01-0 at its 01 authorized amount of $.1 million (excludes authorized DRAM funding). SCE s annual average funding request is $. million over the 01 0 period. SCE is not proposing any change in its approved DR ratemaking, and will utilize existing balancing accounts to recover only actual DR Program costs. B. Cost Reasonableness The reasonableness of the requested costs is discussed in Volume of this Exhibit, with each program s costs discussed in the related section. In addition, the operation of the balancing accounts associated with this application, as discussed below, are reviewed in SCE s Energy Resources Recovery Account (ERRA) Review of Operations Application filed on April 1 of each year. This Commission review verifies that the entries are stated correctly and are consistent with relevant Commission decision(s). C. Ratemaking of DRP Funding SCE proposes no change to the approved DR Program ratemaking mechanisms. SCE s current ratemaking associated with the DR Program includes: (1) the recovery of the authorized DR Program annualized funding through the operation of the Base Revenue Requirement Balancing Account 1 This does not include funding that would be required for a DRAM program, if approved by the Commission. 1

20 (BRRBA); and () recording the difference between the authorized DR Program annualized funding and incurred DR Program expenses in the existing DRPBA. Through this process, customers will ultimately only pay for the incurred DR Program costs. 1 Through the operation of the BRRBA, SCE records, on a monthly basis, the difference between recorded distribution and generation revenues with authorized distribution and generation costs, including authorized DR Program funding amounts. The BRRBA includes a Distribution sub-account and a Generation sub-account because it is necessary to track over- and under-collections that are refunded to or recovered from bundled service customers only (i.e., Generation sub-account) and overand under-collections refunded to or recovered from both bundled service and direct access customers (i.e., Distribution sub-account). Year-end over-collections recorded in the BRRBA are refunded to customers in the subsequent year. Likewise year-end under-collections are recovered from customers in the subsequent year. Also, on a monthly basis, SCE records in the DRPBA the difference between actual DR Program expenses and authorized DR Program funding. Like the BRRBA, the DRPBA includes a Distribution sub-account and a Generation sub-account because it is necessary to track for recovery those DR Programs offered to, and whose costs are recovered from, bundled service customers only (i.e., Generation sub-account), and the DR Programs offered to, and whose costs are recovered from both bundled service and direct access customers (i.e., Distribution sub-account). SCE will continue base allocation of program balancing account revenues based on customer eligibility. Programs that are available to bundled and departing load customers will have revenues allocated on the basis of distribution, while programs that are available only to bundled customers will have revenues allocated on the basis of generation. The allocation factors used for each function will initially be as determined in SCE s 01 GRC Phase proceeding. These factors will be superseded by 1 The Commission adopted similar ratemaking for the capacity contract costs authorized in D For example, the difference between the revenue resulting from the authorized funding for the capacity contracts included in distribution rates and the actually incurred capacity contract costs are recorded in the BRRBA. 1

21 the functional allocation factors in SCE s 01 GRC Phase proceeding once they have been approved by the Commission. D. Balancing Accounts Associated with this Application On March 1, 00, the Commission issued D which authorized the creation of a oneway balancing account, the DRPBA, to track DR program costs and authorized revenues. In addition, D established the funding levels for In D and D.0-0-0, the Commission authorized the funding levels for D.--00 authorized the funding levels for 01 and D authorized the funding levels for Further, D authorized the funding levels for the DR Integrated Demand Side Management Program portfolio and D.1--0 authorized the funding for years The Commission authorized 01 bridge funding levels in D SCE proposes to retain the DRPBA and DRAM recovery through the BRRBA. However, SCE proposes to eliminate the Purchase Agreement Administrative Costs Balancing Account (PAACBA) with the elimination of the AMP contracts in DRPBA Pursuant to D.0-0-0, SCE filed Advice Letter 1-E to establish SCE s Preliminary Statement Part Y concerning the DRPBA. The DRPBA records the difference between the actual DR program costs incurred by SCE and the authorized DR funding level approved by the Commission. SCE proposes to retain this balancing account for the 01 0 time period.. BRRBA In accordance with D.1-0-0, SCE records DRAM-related costs to the BRRBA for cost recovery. The operation of the BRRBA is reviewed in SCE s ERRA Review of Operations Application filed on April 1 of each year. SCE proposes to retain the same DRAM ratemaking for the 01 0 time period.. PAACBA Under D , SCE filed Advice Letter -E to establish SCE s Preliminary Statement Part L concerning the PAACBA. The PAACBA records the difference between SCE s actual and authorized administrative costs associated with the AMP Program authorized by the Commission in 1

22 Decisions D , D.0-0-0, D , D.1-0-0, and D Given that the AMP Program will be discontinued at year-end 01, SCE proposes to eliminate the PAACBA. E DR Program Revenue Requirements Table III- shows SCE s total proposed 01 0 DR funding as it will be recorded in the DRPBA. The total amount will be split between distribution and generation sub-accounts as described above. Table III- Proposed Demand Response Funding for 01-0 Proposed Annualized Cost Recovery of 01-0 DR Application Total Demand Response Program Balancing Account (DRPBA) Distribution Sub-Account $,, $,,1 $ 1,, $,,0 $ 1,,1 $,1, Generation Sub-Account $ 1,, $,,1 $,, $,,00 $,1,1 $,, Subtotal DRPBA $,, $,,01 $,1,1 $,1,01 $,, $ 1,, Base Revenue Requirement Balancing Account (BRRBA) DRAM Distribution Sub-Account $ - $ - $ - $ - $ - $ - Total Proposed Annualized DR Funding in Rates $,, $,,01 $,1,1 $,1,01 $,, $ 1,, 1

23 IV. ANTICIPATED 01 0 DEMAND RESPONSE PROGRAMS ENROLLMENT AND LOAD IMPACT FORECASTS This Chapter presents 01 0 enrollment projections and load impacts for SCE s DR portfolio. A. Background Since the last quarter of 01, SCE has made advances towards bifurcation of its DR resources and has implemented numerous program changes to improve participation and load impacts. SCE anticipates changes approved in its 01 DR Bridge funding proposal and changes provided in this application will have a positive impact on the amount of DR resources available in These changes include, but are not limited to: Implementation and oversight of compliance requirements related to prohibited resources; Introduction of a FSL for AP-I customers; Simplification and improvements to customer enrollment; Caps to utility DR programs (reliability cap and budget cap); Potential conversion of the DRAM from pilot to program; Technology incentives that support and provided under AB ; The expiration of DR resource contracts with third-party aggregators and SCE s DBP. B. Demand Response Enrollment Forecasts In this section, SCE provides a summary of each program s enrollment forecast that will be used in the load impacts for SCE s DR programs requested in this Application. 1. Program Enrollment Forecasts a) Agricultural & Pumping Interruptible AP-I has experienced steady growth over the last several years even though customer outreach and marketing efforts have been minimal. SCE anticipates that the implementation of the ban on prohibited resources, starting in 01, may cause a majority of dairy farms to remove 1

24 themselves from the AP-I program due to generator use. Currently dairy farms make up 0 percent of the AP-I program enrollment population. b) Base Interruptible Program BIP has had steady enrollment in the program. Any reductions or de-enrollments have been offset with new program participants. SCE cannot determine how the ban on prohibited resources will impact program enrollments, therefore, SCE expects little to no changes to enrollments. c) Summer Discount Plan Program SDP has experienced considerable attrition since the program transitioned to a price-responsive program. SCE plans to maintain current device and program operation and to continue to provide support to customers who request the override option, but persistent and mandatory dispatches of the resource may continue to drive down program enrollment. d) Capacity Bidding Program SCE anticipates increased aggregator and customer enrollment and participation in CBP due to the elimination of the AMP contracts in 01 and third-party aggregators that do not receive a DRAM award. CBP continues to be a viable option that provides participants the flexibility to participate from month-to-month. e) Save Power Day (PTR) SCE anticipates continued growth and enrollment in PTR. Growth and program enrollment is primarily attributed to the continued incentive for programmable communicating thermostats and residential customers that will switch to PTR from the SDP program. f) Permanent Load Shifting PLS expects one enrollment annually for the 01 0 funding cycle. PLS projects are costly and can take several years to install. SCE does not have any plans to launch any large marketing initiatives or campaigns to create greater interest in this program. Table IV- presents the forecast annual August enrollment for SCE s DR programs requested in this Application. 1

25 Line No. Table IV DR Portfolio Enrollment Forecast (August) Demand Response Program Enrollment Enrollment Enrollment Enrollment Enrollment 1 Agricultural & Pumping Interruptible (API) 1,1 1,1 1, 1, 1,0 Base Interruptible Program 1 Minute (BIP1) Base Interruptible Program 0 Minute (BIP0) Summer Discount Program Residential (SDP-Res),,, 1, 0,00 Summer Discount Program Commercial (SDP-Com),1,,,,1 Capacity Bidding Program Day Of(CBP-DO) 1,0 1,0 1,0 1,0 1,0 Capacity Bidding Program Day Ahead (CBP-DA) Save Power Days (SPD),00,00,00,00,00 Permanent Load Shifting (PLS) Total,,1,, 1, C. Demand Response Load Impacts This section summarizes the approach used to develop ex ante load impacts for the DR programs for which the Load Impact Protocols were applied, and the estimates that stem from that analysis. The load impacts estimates provided in the following section are generally based on the 01 ex ante load impacts that were filed with the Commission on April 1, 01. SCE s DR portfolio is bifurcated into their supply-side DR programs or load modifying DR programs. 1. General Approach The protocols governing the development of ex ante load impacts align DR impact estimates with other resource alternatives (i.e., other DR resources, EE, renewables, and DG). Figure IV-1 below shows how the ex ante load impact estimates can be integrated with cost-effectiveness analysis and resource planning. As shown, the protocols require that the ex ante load impact estimates be based on analysis of historical data whenever the existing data and characteristics of the program allow for such an approach. SCE analyzes historical program data to produce ex ante load impact estimates used for RA, cost-effectiveness assessment and resource planning. Ex ante load impacts provide a comparison with alternative resources under the same planning paradigm reflecting that DR load impacts vary as a function of weather, participant 1

26 characteristics, changes in the number of program participants and other factors (e.g., switch failure rates). Put differently, ex post load impacts for any year may differ from the load impacts that could be achieved during the low probability, extreme conditions under which many DR resources are likely to be used and for which they provide insurance value. Figure IV-1 Summary of Ex Post and Ex Ante Analysis, Process, and Connections. Overview of Evaluation Methods The methodologies used to estimate ex post and ex ante load impacts for each of the DR programs in the SCE portfolio are conceptually similar. Each of the 01 evaluations (filed April 1, 01) relied on regression analysis reflecting the relationship between customer or end-use load and key determinants of the variation in energy use over time, such as weather and time-of-day, day-of-week and seasonal patterns that reflect the normal pattern of business or household operations. These models are based on historical hourly or sub-hourly electricity use data for customers that have participated in the 1

27 DR programs. Each model or set of models is used to estimate the reference load for an average customer enrolled in a program, which represents what customers would be expected to use in the absence of an event on days on which program events either were called (for ex post impact estimation) or have a high probability of being called (for ex ante impact estimation). In most instances, ex post impacts were estimated by comparing the reference level energy use in each hour with actual energy use in the hour on each event day. For ex ante estimation, predicted energy use in each hour was estimated under the assumption that an event occurred and also under the assumption that it did not occur, while everything else (e.g., weather, day of week effects) was held constant at values representative of a typical event day or monthly system peak day. When enrollment in a DR program is expected to change significantly over time and, importantly, if the mix of customers is expected to change in ways that would likely affect average customer impact for the program, this fact must be taken into consideration when developing ex ante impact estimates. Load reduction resource amounts and customer mix are not expected to change substantially for most of SCE s DR resources. At a more technical level, three general approaches were used to estimate the regression models: a) Individual Customer Time Series Regressions This method works well for event-based programs with numerous events and for programs with substantial variation in the drivers of load response or load shifting. This approach is also useful for programs with substantial differences in the magnitude and load patterns of customers, which is more typical among large customers. The coefficients vary at the customer level. While the regressions do not necessarily explain individual customer behavior perfectly, in aggregate, they explain most of the program level variation in loads. Individual customer regressions can, however, be employed to describe the distribution of customer load reductions as well as the distribution of percent load reductions. They can also be used to describe impacts for segments of the participant population. The key limitation to individual customer regressions is their inability to make use of control groups. 0

28 b) Aggregate Time Series This method aggregates across customers in a program or segment (e.g., by industry), within a program. It works effectively for event-based programs where enrollment is stable and event periods are the same for all participants. By averaging or aggregating across customers, individual customer idiosyncrasies are smoothed out and the load patterns are more predictable. With some foresight, the approach can be used to produce estimates for specific customer segments. Like individual regressions, aggregate time series cannot employ control groups. This approach is inappropriate when the participant population is changing substantially. c) Panel Regressions This method is particularly suitable when equivalent control groups are available or sample sizes are sufficient for the territory but inadequate for smaller segments such as local capacity areas. A key strength of panel regressions is the ability to control for certain omitted or unobservable variables. While panel regressions can increase the accuracy of impact estimates for the average customer, they cannot be employed to describe the distribution of impacts among the participant population. Nor can panel regressions control for customer characteristics that interact with occupancy or weather unless those variables are explicitly included. The regression models used to predict the reference load were developed with the goal of accurately predicting average customer load given the time of day, day of week, temperature, and location of the customer and predicting load reductions under different temperature conditions. The focus was on the accuracy of the prediction and the validity of load impact estimates. D. Ex Post Load Impacts This section summarizes the load impacts in 01 for event based programs. Ex post load impacts are based on modeling electricity use patterns and load impacts over a historical load period. They estimate what happened based on the conditions that were in effect during that time. While historical load patterns and impacts are critical to understanding the magnitude of load reduction resources, they have limitations. Because historical performance is tied to past conditions such as weather, price levels and dispatch strategy (e.g., localized dispatches), ex post load impacts may not 1

29 reflect the full option value of a DR resource. For example, a test event of a highly weather-sensitive program such as SCE s commercial SDP program may yield lower impacts than the actual resources the program can provide because future events might occur at hotter temperatures when air conditioner loads are larger. Likewise, resources such as CBP may be dispatched partially one product line is called in which case ex post events do not necessarily reflect the program load reduction capability. 1. Event Averages by Program Interpreting the aggregate average impact across events can be difficult because multiple factors vary across days, including temperature, the normal pattern of energy use, enrollment, number of customers called, dispatch strategy, and event hours. For programs such as PTR, with stable participation, fixed event windows, less weather sensitive customers, and universal dispatch for all events, the average event impacts can provide meaningful and insightful data about program performance. However, for resources that do not have those characteristics, the average aggregate impacts provide limited insight and can be misleading. A better metric for average ex post impacts is the average and percent impacts per event hour for customers that were dispatched. Still, this metric has flaws in that it does not necessarily reflect the load reduction capability for weather sensitive resources, or for resources that are dispatched on a localized basis where the dispatched participants may have different characteristics than the average participant. In short, ex post load impacts may not reflect the full option value of a DR resource and should be interpreted with caution. E. Ex Ante Load Impacts The portfolio ex ante load impact estimates summarize the load reduction that can be expected from all of SCE s DR programs if called simultaneously. The impact estimates are based on a common event window of 1 p.m. p.m., and the weather conditions underlying 1-in- and 1-in- monthly system peak days. In other words, they provide estimates of the resources available under conditions that are linked to the need for additional capacity. To the extent that growth in programs is expected, enrollment forecasts from SCE were factored into the ex ante estimates.

30 1. DR Portfolio Resources by Forecast Year Table IV- below summarizes the projected load reduction resources by program for 01 0 under 1-in- annual system peak conditions. Table IV Ex Ante Load Impacts by DR Program August 1-in- SCE System Peak Conditions Line MW Demand Response Program No Agricultural & Pumping Interruptible (API) 1 Base Interruptible Program 1 Minute (BIP1) Base Interruptible Program 0 Minute (BIP0) 1 0 Summer Discount Program Residential (SDP-Res) Summer Discount Program Commercial (SDP-Com) 0 Capacity Bidding Program Day Of(CBP-DO) Capacity Bidding Program Day Ahead (CBP-DA) Save Power Days (SPD) 1 1 Permanent Load Shifting (PLS) Total (MW) 1,0 1,0 1,0 1,00 1

31 V. COST-EFFECTIVENESS A. Overview The purpose of this Chapter is to describe SCE s cost-effectiveness analysis model and describe inputs to the model. As shown in Table V- below, SCE s DR portfolio is cost-effective, having a TRC of 1.. In D.-0-0 and D.1--0, the Commission adopted a methodology for estimating the cost-effectiveness of DR programs and directed the utilities to perform DR cost-effectiveness analysis using the approved methodology and the DR Reporting Template developed by Energy and Environmental Economics, Inc. (E). This cost-effectiveness methodology is also referred to as the cost-effectiveness protocols (01 DR Protocols). A detailed explanation of the cost-effectiveness analysis that was performed under the 01 DR Protocols and E s model is provided below DR Protocols SCE conducted its DR cost-effectiveness analysis as ordered by D in compliance with the 01 DR Protocols adopted in D Most model assumptions and avoided cost inputs are provided in the DR Reporting Template developed by E. Below is a list of utility inputs required by the model: Load impacts (MW) Energy savings (MWh) Adjustment factors to E avoided cost inputs Administrative costs Incentive costs Net bill/revenue reductions Amortized equipment costs and amortization periods CAISO market participation revenues

32 . Avoided Cost SCE used the public avoided cost numbers provided by E in the 01 avoided cost calculator. 1 The components of avoided cost include: generation energy, generation capacity, ancillary services, transmission and distribution capacity, and emissions. Compared to other avoided cost components, generation capacity typically has the largest effect on cost-effectiveness results. Generation capacity uses the long-run capacity cost for all years. In the previous filing, the generation capacity cost transitioned from a near-term capacity cost based on RA costs, to the long-run capacity cost based on the resource balance year (first year in which new capacity resources may be needed to meet the growth of peak loads and reliability requirements). Using the longrun capacity cost for all years produces higher avoided generation capacity costs. Table V- below shows the generation capacity values used to calculate avoided costs. Table V- Avoided Capacity Cost By Year The values for the remaining components are found in the cost-effectiveness calculator in a tab called, Inputs. B. Net Present Value and Benefit-Cost Ratio Results The Commission has ruled that programs with TRC test results higher than 0. are considered to be cost-effective. 1 As shown in Table V- below, the TRC score for SCE s total 01 0 DR portfolio is 1., and is therefore considered to be cost-effective. 1 The E avoided cost model is available at [as of January 1, 01]. 1 D.1-0-0, p..