2018 General Rate Case Rebuttal Testimony

Transcription

1 Application No.: A Exhibit No.: SCE-18, Vol. 08 Witnesses: M. Flores J. Goizueta (U 338-E) 2018 General Rate Case Rebuttal Testimony Transmission & Distribution (T&D) Volume 8 - Infrastructure Replacement Before the Public Utilities Commission of the State of California Rosemead, California June 16, 2017

2 SCE-18: Transmission & Distribution Volume 08 - Infrastructure Replacement Table Of Contents Section Page Witness I. INFRASTRUCTURE REPLACEMENT REBUTTAL...1 M. Flores/ J. Goizueta A. Distribution Infrastructure Replacement (DIR) Program...3 J. Goizueta 1. Worst Circuit Rehabilitation Program...3 a) b) TURN s Position...4 SCE s Rebuttal to TURN s Position...4 (1) (2) (3) (4) SCE s modeling assumptions do not bias the model results...4 The Reliability Model s exclusion of Grid Modernization benefits is a model feature, not a flaw...5 TURN s Argument for Including Reliability Benefits from Grid Modernization is Contradictory to their Grid Modernization Proposal...6 Cost-benefit analysis of mainline injection should be performed prior to initiating a mainline injection pilot Overhead Conductor Program...7 a) b) ORA s Position...7 SCE s Rebuttal to ORA s Position...9 (1) (2) (3) (4) The primary driver of OCP is safety, not reliability...9 SCE s SAIDI/SAIFI Simulations Do Not Understate the Reliability Impact of OCP...10 SCE's requested increase is not "skyrocketing" nor "uncommon"...11 OCP Ramp Up i-

3 SCE-18: Transmission & Distribution Volume 08 - Infrastructure Replacement Table Of Contents (Continued) Section Page Witness c) d) TURN s Position...13 SCE s Rebuttal to TURN s Position...13 (1) (2) (3) TURN s Use of PRISM Results is Premature...13 TURN Does Not Accurately Describe the Differences between Reactive and Proactive Projects...14 TURN s Characterization of Oversized BLFs as Resulting in self-inflicted damage to the System is Simply Wrong...15 e) f) Consumer Federation of California s (CFC) Position...17 SCE s Rebuttal to CFC s Position...17 (1) (2) (3) The primary driver of OCP is safety, not reliability...17 The Rate of SCD Increase in Years Past Should Not be Confused with the Current Level of Small-wire Risk...18 CFC Inappropriately Proposes a Rate Rider for OCP Capacitor Bank Replacement Program...19 a) b) TURN s Position...20 SCE s Rebuttal to TURN s Position...20 Appendix A IR Appendix A Data Request Responses... -ii-

4 I. INFRASTRUCTURE REPLACEMENT REBUTTAL In SCE-02, Volume 8, SCE presented evidence that large parts of its substation and distribution infrastructure are beyond their useful service life. We explained why it makes sense to replace that aging infrastructure before it fails in service and causes safety and reliability problems. We helped justify our proposed infrastructure replacement work and costs with detailed workpapers. In response to SCE s forecast of capital work needed to replace aging infrastructure, ORA presented testimony that proposes reductions in SCE s Overhead Conductor Program (OCP), which is aimed at improving public and employee safety. ORA s forecast for all other Distribution Infrastructure Replacement Programs include minor adjustments due to rounding. During the discovery process, SCE made a number of corrections to the forecasts for SCE s Substation Infrastructure Replacement programs, which ORA found reasonable and accepted. 1 TURN offers testimony on SCE s Worst Circuit Rehabilitation (WCR) Program, OCP, and Capacitor Bank Replacement Program and recommends lower alternative forecasts for each program. CFC also proposes an alternative forecast for OCP. Below, SCE responds to the ORA, TURN, and CFC recommendations. Table I-1 summarizes SCE s recorded and forecast spending in the various infrastructure replacement accounts and the 2017 and 2018 forecasts proposed by ORA, TURN, and CFC. 1 See responses to ORA-SCE-131-GAW Q.02.a Supplemental, TURN-SCE-016 Q.11.a Supplemental, TURN- SCE-016 Q.14.a Supplemental, TURN-SCE-016 Q.14.d Supplemental on Appendix pp. A-1 to A-5. 1

5 Table I-1 Summary of Infrastructure Replacement Capital Expenditures 2 Total Company Nominal $000 3 Recorded Activity Worst Circuit Rehabilitation $ 77,010 $ 66,942 $ 121,016 $ 153,013 $ 117,935 $ 143,162 Cable Life Extension $ - $ 1,361 $ 7,943 $ 13,244 $ 11,665 $ 22,858 CIC Replacement $ 5,582 $ 4,674 $ 13,343 $ 23,042 $ 54,077 $ 33,468 Overhead Conductor Program $ - $ - $ - $ - $ 58,126 $ 97,330 Underground Oil Switch Replacement $ 15,580 $ 9,799 $ 18,842 $ 19,580 $ 25,938 $ 17,566 Capacitor Bank Replacement $ 7,957 $ 6,583 $ 8,917 $ 7,814 $ 8,679 $ 7,765 Automatic Reclosure Replacement $ 1,471 $ 905 $ 1,240 $ 1,523 $ 2,488 $ 2,136 PCB Transformer Replacement $ 719 $ 514 $ 937 $ 1,368 $ 1,326 $ 1,579 Substation Transformer Bank Replacement $ 68,021 $ 67,729 $ 63,497 $ 70,271 $ 72,981 $ 104,905 Substation Circuit Breaker Replacement $ 28,246 $ 24,239 $ 35,944 $ 60,161 $ 47,711 $ 49,849 Substation Switchrack Rebuilds $ - $ - $ - $ - $ 1,304 $ 3,937 Total Capital SCE-02 Volume 8 $ 204,584 $ 182,746 $ 271,678 $ 350,017 $ 402,229 $ 484,555 SCE Forecast ORA Forecast Total Total Activity Variance Worst Circuit Rehabilitation $ 123,106 $ 126,207 $ 249,313 $ 123,098 $ 126,175 $ 249,273 $ (40) Cable Life Extension $ 23,402 $ 23,991 $ 47,393 $ 23,340 $ 23,923 $ 47,263 $ (130) CIC Replacement $ 31,142 $ 41,643 $ 72,785 $ 31,140 $ 41,633 $ 72,773 $ (12) Overhead Conductor Program $ 136,087 $ 139,514 $ 275,601 $ 90,724 $ 116,262 $ 206,986 $ (68,615) Underground Oil Switch Replacement $ 11,150 $ 12,701 $ 23,851 $ 11,150 $ 12,701 $ 23,851 $ 0 Capacitor Bank Replacement $ 17,156 $ 17,588 $ 34,744 $ 17,155 $ 17,583 $ 34,738 $ (6) Automatic Reclosure Replacement $ 2,310 $ 2,368 $ 4,678 $ 2,310 $ 2,368 $ 4,678 $ (0) PCB Transformer Replacement $ 1,413 $ 1,449 $ 2,862 $ 1,413 $ 1,449 $ 2,862 $ (0) Substation Transformer Bank Replacement $ 66,349 $ 68,003 $ 134,352 $ 66,349 $ 68,003 $ 134,352 $ - Substation Circuit Breaker Replacement $ 43,875 $ 44,943 $ 88,818 $ 43,875 $ 44,943 $ 88,818 $ - Substation Switchrack Rebuilds $ 18,362 $ 18,825 $ 37,187 $ 18,362 $ 18,825 $ 37,187 $ (0) Total Capital SCE-02 Volume 8 $ 474,352 $ 497,232 $ 971,584 $ 428,916 $ 473,865 $ 902,781 $ (68,803) SCE Forecast TURN Forecast Total Total Activity Variance Worst Circuit Rehabilitation $ 123,106 $ 126,207 $ 249,313 $ 103,761 $ 103,761 $ 207,522 $ (41,791) Cable Life Extension $ 23,402 $ 23,991 $ 47,393 n/a n/a n/a n/a CIC Replacement $ 31,142 $ 41,643 $ 72,785 n/a n/a n/a n/a Overhead Conductor Program $ 136,087 $ 139,514 $ 275,601 $ 73,230 $ 75,075 $ 148,305 $ (127,296) Underground Oil Switch Replacement $ 11,150 $ 12,701 $ 23,851 n/a n/a n/a n/a Capacitor Bank Replacement $ 17,156 $ 17,588 $ 34,744 $ 9,024 $ 9,250 $ 18,274 $ (16,470) Automatic Reclosure Replacement $ 2,310 $ 2,368 $ 4,678 n/a n/a n/a n/a PCB Transformer Replacement $ 1,413 $ 1,449 $ 2,862 n/a n/a n/a n/a Substation Transformer Bank Replacement $ 66,349 $ 68,003 $ 134,352 n/a n/a n/a n/a Substation Circuit Breaker Replacement $ 43,875 $ 44,943 $ 88,818 n/a n/a n/a n/a Substation Switchrack Rebuilds $ 18,362 $ 18,825 $ 37,187 n/a n/a n/a n/a Total Capital - Infrastructure Replacement $ 474,352 $ 497,232 $ 971,584 $ (185,557) SCE Forecast CFC Forecast Total Total Activity Variance Overhead Conductor Program $ 142,203 $ 136,087 $ 139,514 $ 275,601 $ 136,087 $ 116,300 $ 252,387 $ (23,214) 2 SCE forecast based on Errata served on June 16, For Substation Transformer Banks and Circuit Breakers, ORA s testimony showed CPUC-jurisdictional costs only. 2

6 A. Distribution Infrastructure Replacement (DIR) Program SCE accepts ORA s proposals for all Infrastructure Replacement Programs, except for the OCP. Additionally, based on ORA s analysis, some rounding errors occurred when ORA applied escalation to 2015 dollars for Worst Circuit Rehabilitation, Cable Life Extension, CIC Replacement, and Capacitor Bank Replacements. 4 When final escalation rates applicable to this GRC are approved, SCE will apply those rates to the constant values for these programs, which will eliminate these rounding discrepancies. 1. Worst Circuit Rehabilitation Program The objective of the WCR program is to improve system reliability by replacing distribution circuit infrastructure before it fails, thereby avoiding unplanned outages to our customers, and making circuits more resilient to future failures. The WCR program focuses on those circuits that disproportionately contribute to system SAIDI and SAIFI, and those circuits where customers are receiving relatively lower service reliability. Because cable failure is the largest equipment contributor to poor system reliability, circuit rehabilitation typically involves replacing each circuit s most risksignificant mainline cable. This program also replaces infrastructure that has a lower reliability record and adds circuit enhancements such as automation, automatic reclosers, branch line fuses, and fault indicators wherever determined to be cost-effective. To develop the forecast for the WCR program, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. Table I-2 Worst Circuit Rehabilitation Program Capital Expenditures 100% CPUC Jurisdictional Nominal $000 Recorded Activity Worst Circuit Rehabilitation $ 77,010 $ 66,942 $ 121,016 $ 153,013 $ 117,935 $ 143,162 SCE Forecast TURN Forecast Activity Total Total Variance Worst Circuit Rehabilitation $ 123,106 $ 126,207 $ 249,313 $ 103,761 $ 103,761 $ 207,522 $ (41,791) 4 See Exhibit ORA-08 footnote 5 on page 12. 3

7 a) TURN s Position TURN proposes reducing SCE s WCR forecast by $ million in 2017 and 2018, based on a 55 conductor-mile reduction in the cable replacement forecast in each year with no adjustment to the unit cost. TURN agrees that SCE must continue to replace mainline cable, but argues that the reliability modeling that supports SCE s forecast is possibly flawed, that SCE should expect to realize reliability improvement in the coming years from the Grid Modernization program, and that SCE could determine that cable injection is more cost-effective than replacing mainline cable. TURN also makes three additional recommendations related to the WCR program. First, TURN claims that SCE should begin recording cable failures by cable type (e.g., XLPE, PILC, etc.). Second, TURN urges the Commission to direct SCE to change the minimum age used to select mainline-cable replacements. Third, TURN posits that SCE should pilot cable injection on mainline cable, using funds from the forecast that the Commission adopts for the WCR program, and that SCE should report on quantitative and qualitative findings from the pilot in the next GRC. b) SCE s Rebuttal to TURN s Position (1) SCE s modeling assumptions do not bias the model results TURN points out that SCE uses equipment failure rates, the number of customers affected, and restoration time to determine circuit performance in the Reliability Model. TURN then asserts that if there is a problem with the estimates of any of these items it will bias the model s results. SCE disagrees with this statement. What TURN characterizes as problems in SCE s models are really modeling assumptions that SCE uses in lieu of perfect data. SCE has validated key modeling assumptions through the process of comparing model results to data that is available. This validation/review process helps ensure that modeling assumptions do not bias the model results. For example, TURN criticizes modeling assumptions that SCE has made in developing the Weibull model for underground cable. As discussed in TURN-SCE a, 5 SCE made a modeling assumption about the future failure rate of cables removed from inventory due to infrastructure replacement. TURN takes issue with this, stating that SCE has not provided any basis for this specific modeling assumption, and asserts that this assumption results in an upward bias to the resulting failure rate curve and downward bias to the mean time to failure (MTTF). 5 See response to TURN-SCE-108-Q-02.a-c. on Appendix pp. A-6 to A-7. 4

8 SCE disagrees with TURN s conclusions for a simple reason SCE has compared the model results to available data as a means of validating the reasonableness of the underlying assumptions. In this case, SCE validated the reasonableness of the assumptions by comparing the actual total cable failures from 2014 Outage Data and Reliability Metrics (ODRM) system data with the predicted total failures from the model developed by 2014 inventory data. Even TURN points out that this match was favorable, with less than a 1% difference. This is clear evidence that SCE s modeling assumptions do not bias the model results. TURN criticizes two specific details of the validation namely, that the validation was performed only on total failures and not on specific cable types, and that the validation was performed with only one year of data (2014). The first criticism is itself a form of circular reasoning lack of data to make such a comparison is the very reason for the need for the modeling assumption in the first place. In response to the second criticism, SCE points out that 2014 ODRM data is the most appropriate year for comparison to failures predicted from 2014 inventory data; selecting any other year would not be an apples-to-apples comparison for model validation purposes. In short, SCE s model validation process was reasonable and appropriate based on available data, and the assumptions that SCE has made do not bias the model results. (2) The Reliability Model s exclusion of Grid Modernization benefits is a model feature, not a flaw TURN criticizes the reliability model for not simulating the potential impacts of the proposed Grid Modernization program. But this exclusion is by design. The Reliability Model is intended to isolate the future impacts of aging cable from other factors. Because the model focuses on aging cable, SCE can confidently use the model results to identify a pre-emptive cable replacement rate (i.e., WCR forecast) proportional to the targeted problem (i.e., increased failures of aging cable). This is why programs such as Grid Modernization and OCP are not included in the simulation model; this exclusion is a model feature, not a model flaw. In addition, Grid Modernization is not a reason to reduce the WCR forecast. Grid Modernization will help isolate failed cables faster, but will not address the underlying problem of aging cable. SCE s cable inventory continues to age, and will continue to fail at an increasing rate absent robust WCR funding. Grid Modernization will also not change the fact that underground cable cannot be visually inspected, and without a deliberate preemptive replacement program, cable would be removed from the system only because of in-service failure. Due to the 5

9 problems inherent in any unscheduled in-service failure of cable, the rationale supporting SCE s WCR forecast is unaffected by additional SAIDI/SAIFI benefits that may be achieved through Grid Modernization. (3) TURN s Argument for Including Reliability Benefits from Grid Modernization is Contradictory to their Grid Modernization Proposal The primary objective of SCE s WCR program is to proactively replace deteriorated underground cable to avoid catastrophic in-service failures. In-service failures to mainline cable can have major impacts on a large number of customers. While benefits show up in many areas including SAIDI measurements, and TURN argues that SCE should have included Grid Modernization SAIDI benefits in our forecast for WCR, TURN fails to acknowledge that SCE is pursuing Grid Modernization for different reasons than WCR. In addition, TURN cites anticipated SAIDI improvements from a Grid Modernization analysis that assumes full funding. But TURN is proposing draconian cuts to SCE s Grid Modernization funding proposal. 6 It is fundamentally inconsistent for TURN to argue that for the purposes of WCR the Commission should take into account anticipated SAIDA benefits from a different program that TURN is simultaneously proposing to slash. (4) Cost-benefit analysis of mainline injection should be performed prior to initiating a mainline injection pilot SCE agrees with TURN that it is prudent to explore if cable injection would be beneficial for mainline cable. However, instead of going directly to a pilot as TURN suggests, SCE recommends a cost-benefit analysis be performed first to determine if a pilot is necessary. SCE has taken this approach with both CIC testing and WCR testing in recent years. In the case of CIC testing, the cost-benefit analysis was favorable, and this provided justification to further develop and implement CIC injection testing as a program. In the case of mainline testing, the cost-benefit analysis was not favorable, as discussed in SCE s direct testimony. SCE recommends that similar cost-benefit analysis be performed for mainline cable injection as a first step, to determine if a pilot should be implemented. If the cost-benefit analysis is positive, then SCE would evaluate a pilot. The pilot would then test the assumption of the cost-benefit analysis and also provide insight into operational considerations for 6 See Exhibit TURN-04, p

10 potential mainline injection activities. The Commission should adopt TURN s recommendation with SCE s proposed modification, i.e. to perform a cost-benefit analysis before undertaking a potential pilot. 2. Overhead Conductor Program SCE s Overhead Conductor Program (OCP) is a new program designed to address public safety risks. The goals of OCP are to reduce the frequency and impact of wire down events by executing proactive overhead conductor replacement projects and reactive planned conductor work after wire down events. Similar to the WCR program that focuses on the worst performing circuits to address reliability risks, OCP ranks overhead circuits based on risk-based criteria such as small wire. The objective of OCP is to reduce wire down events in order to address safety and reliability risks. To develop the cost forecast for this program, SCE analyzed 2015 historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. Table I-3 Overhead Conductor Program Capital Expenditures 100% CPUC Jurisdictional Nominal $000 Recorded Activity Overhead Conductor Program $ - $ - $ - $ - $ 58,126 $ 97,330 SCE Forecast ORA Forecast Activity Total Total Variance Overhead Conductor Program $ 136,087 $ 139,514 $ 275,601 $ 90,724 $ 116,262 $ 206,986 $ (68,615) SCE Forecast TURN Forecast Activity Total Total Variance Overhead Conductor Program $ 136,087 $ 139,514 $ 275,601 $ 73,230 $ 75,075 $ 148,305 $ (127,296) SCE Forecast CFC Forecast Activity Total Total Variance Overhead Conductor Program $ 142,203 $ 136,087 $ 139,514 $ 275,601 $ 136,087 $ 116,300 $ 252,387 $ (23,214) a) ORA s Position ORA s position regarding OCP is based on the totality of programs that make up cable or conductor replacements; SCE will refer to these as linear assets for purpose of this discussion. 7

11 ORA does not agree that the SAIDI reliability projections shown in Figure III-9 7 accurately represent how reliability will change over the years. ORA feels that these projections exclude the impacts of CIC Replacement expenditures, the impacts due to the growing CIC Injection program, and the impacts of the newly created OCP. ORA contends that instead of only considering how one preemptive replacement program impacts SAIDI, all four programs should be incorporated into such an analysis. 8 ORA has calculated total replacement quantities of linear assets among the four programs (i.e. WCR, CIC injections, CIC replacements, and OCP) for both historical and forecast years. ORA has calculated that the highest historical total replacements was 637 conductor-miles in 2015, and that SCE s request would result in total replacements of 1,315 conductor-miles in 2017 and 1,350 conductor-miles in ORA states that an increase of this magnitude is uncommon and must be closely examined. 10 ORA has stated that it does not oppose SCE s proposed replacement levels for CIC Injections and CIC Replacements, but does take issue with SCE s replacement forecast for the OCP. Citing SCE s response to data request ORA-SCE-110-GAW, question 5c, 11 ORA states that SCE held overhead conductor reliability constant when it derived its SAIDI and SAIFI graphs. ORA further states that the creation of the OCP should mean that overhead conductor reliability should be improving, not held constant. As a result, ORA concludes that the SAIDI/SAIFI analysis performed by SCE will automatically understate the level of reliability improvements that the distribution system will experience. 12 ORA recommends that 200 circuit-miles be replaced in 2017 and 250 circuitmiles be replaced in 2018 for OCP, which is lower than SCE s forecast by 100 circuit-miles in 2017 and 50 circuit-miles in These adjustments are based on two separate analyses by ORA. First, ORA states that the level of preemptive cable replacements has skyrocketed in recent years, resulting in 7 See Exhibit ORA-08, p. 15, Figure III-4. 8 See Exhibit ORA-08, p. 16, lines See Exhibit ORA-08, p. 17, Table See Exhibit ORA-08, p. 22, lines See response to ORA-SCE-110-GAW Q.5c on Appendix pp. A-8 to A See Exhibit ORA-08, pp

12 reliability improvements that already exceed the levels developed by SCE. Second, ORA states that its recommendation is a large OCP replacement ramp up over 2015 recorded levels for 2017 and 2018, and there is no immediate need to undertake the even higher replacement levels proposed by SCE. 13 b) SCE s Rebuttal to ORA s Position (1) The primary driver of OCP is safety, not reliability ORA s argument to reduce the forecast of overhead conductor replacements is made on the basis of a SAIDI/SAIFI reliability argument. However, the primary driver of OCP is safety, not reliability. 14 Safety and reliability risk analysis has demonstrated that the safety risk caused by energized wire down events is considerable relative to other system risks, and OCP was created to address these safety risks. 15 OCP addresses the problem of overhead wire installations being undersized relative to existing levels of available short circuit duty (SCD). This is mitigated through a combination of changes to protection settings, installation of new protective devices, and the reconductor of small-gauge wire to a larger size that reduces the risk of conductor damage during fault conditions. 16 Based on a replacement rate of 300 miles per year, SCE estimates that it will take approximately 53 years to replace all small wire on the SCE system. 17 In fact, if design and construction resources were not a consideration, SCE would prefer to replace small wire at an even faster pace, as doing so would likely lead to increased mitigation of public safety risks. 18 ORA s argument in favor of reducing the OCP forecast based on SAIDI/SAIFI factors overlooks the fundamental safety risk driver of the program. ORA never addressed these safety concerns in its testimony. 13 See Exhibit ORA-08, p. 22, lines See response to CFC-SCE-002-Q.6e on Appendix p. A SCE-02, Vol. 8, p. 47, lines SCE-02, Vol. 8, p. 48, lines See response to TURN-SCE-016Q.4a on Appendix p. A See response to TURN-SCE-042Q.4a on Appendix p. A-12. 9

13 (2) SCE s SAIDI/SAIFI Simulations Do Not Understate the Reliability Impact of OCP In addition, SCE disagrees with ORA s conclusion that reliability simulations which hold overhead conductor failure rates constant will automatically understate the expected reliability impacts of OCP. In its testimony, ORA cites a portion of SCE s data request response to ORA-SCE-110-GAW question 5c. But ORA neglects to cite or address other key elements of that same response. 19 As indicated in the entirety of SCE s response to ORA-SCE-110-GAW question 5c, SCE clarifies that an equivalent model of age-based probability of failure of overhead conductor does not yet exist. As a simplifying assumption, SCE uses a uniform, time-independent failure rate for all overhead conductors in reliability simulations. It is important to recognize that this simplifying assumption applies to baseline reliability simulations (i.e. the top curves in Figures III-9 and III-10) 20 as well as WCR reliability simulations (i.e. the bottom curves in Figures III-9 and III-10). 21 In other words, reliability simulations show neither the reliability degradation associated with undersized overhead conductors over years of fault current exposure, nor the offsetting reliability benefits of OCP. Net system reliability may improve if the rate of small-wire overhead conductor degradation is exceeded by the rate of OCP benefit; net reliability may get worse if the rate of degradation exceeds the rate of OCP benefit. The effect over time SAIDI/SAIFI improvement or SAIDI/SAIFI degradation is not currently known based on the underlying modeling limitations and are therefore assumed to have an offsetting impact. The SAIDI/SAIFI impacts of OCP are expected to be better understood by SCE as more post-project operational experience is gained over time See Exhibit ORA-08, p. 18, lines The concluding paragraph in SCE s response to ORA-SCE-110-GAW question 5c states: in fact, the no WCR program curves (i.e., Figures III-4 and III-5) do not show any changes in long-term SAIDI/SAIFI trends due to overhead conductors. These curves were made with a simplified assumption of a uniform, timeindependent failure rate for all overhead conductors. In other words, overhead conductor reliability performance was held constant for all simulations in this rate case in order to isolate anticipated reliability impacts of the WCR program alone. For clarification, the curves in Figures III-4 and III-5 are repeated as the top curves in Figures III-9 and III-10 respectively. 21 See response to ORA-SCE-110-GAW Q.5c on Appendix pp. A-8 to A See response to CFC-SCE-005-Q. 9b on Appendix p. A

14 The study assumption identified by ORA is related to overhead conductor failure rates, but in fact the assumption helps support simulation-based conclusions regarding the WCR program, not the OCP program. The assumption serves to isolate simulated reliability impacts of underground cable (i.e. with established time-dependent failure rate models) from reliability impacts of overhead conductor by holding overhead failure rates constant in all simulations. No OCP conclusions pro or con can be supported by the reliability simulations shown in Figures III-9 or III (3) SCE's requested increase is not "skyrocketing" nor "uncommon" SCE presented its forecast quantities of overhead replacements in terms of circuit-miles and not conductor-miles. ORA has made the assumption that two-phase and three-phase circuits exist in equal quantities, resulting in a factor of 2.5 to convert circuit-miles to conductor miles. With this conversion, ORA calculates that SCE has replaced approximately 185 conductor-miles in OCP in 2015, forecasts 800 conductor-miles in 2016, and forecasts 750 conductor-miles in each year 2017 and While SCE s OCP forecast and unit costs are based on circuit-miles, SCE agrees that ORA s assumed factor of 2.5 is reasonable for purposes of comparison of quantities between overhead and underground linear asset programs. SCE clarifies that this 2.5 factor should only be used for comparing total conductor-mile replacements across multiple programs. For GRC forecast or unit cost purposes, OCP by itself should continue to be based on circuit-miles, not on conductor-miles with an assumed 2.5 conversion factor. SCE disagrees with ORA s characterization of increases from 2015 (637 conductor-miles based on the 2.5 factor) and 2017 (1,315 conductor-miles based on the 2.5 factor) as a skyrocket increase or uncommon. The proposed increase is modest relative to the total inventory of linear assets in the SCE system. ORA s table 8-2 shows this total increase of 678 conductor-miles includes 113 conductor-miles of underground assets (column g) and 565 conductor-miles of overhead linear assets (column i). SCE s primary overhead distribution system consists of approximately 105,800 conductor miles, and SCE s primary underground distribution system consists of approximately 54,800 conductor-miles. Increases of 565 and 113 conductor-miles, respectively, represent only approximately 0.5% and 0.2% of SCE s total overhead and underground primary distribution inventory, respectively. 23 See response to ORA-SCE-110-GAW Q.5c on Appendix pp. A-8 to A-9. 11

15 SCE s proposed increased remediation rates do not represent "skyrocketing" increases as ORA characterizes, but instead represent modest increases in light of the drivers for these forecast increases and the inventories that are subject to these replacement rates. Increases of this magnitude are also not uncommon, as ORA states. As new problems are identified on the electric distribution system, prudent utility operators should pursue new activities to address those problems, and the rate of increase should be proportionally sized to address the problem. For example, CIC replacement programs in the timeframe replaced CIC at an average replacement rate of 9.9 conductor-miles per year in However, as SCE experienced problems associated with CIC failure rate increases, these replacement rates increased to present levels (150 conductor-miles replacement and 100 conductor-miles of rejuvenation). Likewise, in the case of OCP, the increase of 565 conductor-miles per year cited by ORA is entirely justified by the current degree of risk associated with small overhead wire, SCE s understanding of those risks, and SCE s urgent need to upgrade the overhead system to present-day wire size standards. Any new/emerging problem on the distribution network naturally results in a new/emerging increase in work activities to address these problems. This is a prudent way to manage assets within such a large, diverse, and evolving electric system. (4) OCP Ramp Up In 2016, SCE completed approximately 202 miles of its planned scope of 204 miles. 24 This is up from 74 miles in 2015 and 0 miles in This demonstrates that SCE is able to ramp up the new OCP effort and that SCE will be able continue to execute, at a minimum, this level of work. In 2016, SCE completed all forecast work, but at a much lower total cost. 25 This lower cost was achieved due to SCE s continued efforts to look for ways to improve processes and lower costs for our customers. Due to the importance of mitigating safety issues related to wire down events, SCE believes it is reasonable to increase the annual scope of our initial OCP forecast. The forecast SCE originally 24 See responses to TURN-SCE-059-Q1.b.iii for scope, TURN-SCE-059-Q12a for completed miles, and TURN- SCE-059-Q12c for differences between forecasting methodology that uses on miles vs. actual scope that includes miles and BLFs resulting in less equivalent miles between recorded and forecast. See Appendix pp. A-14 to A SCE recorded $ million for OCP compared to the 2016 forecast of $ million, which is approximately 31% lower cost. 12

16 provided would require approximately 53 years to mitigate the current population of small wire. 26 Based on 2016 results, SCE believes that for the same amount of money SCE requested in its original GRC capital forecast, SCE can replace approximately 434 miles of small wire versus the originally-forecast 300 miles in each of 2017 and All other things being equal, this funding level would allow SCE to replace all small wire on its system in 37 years instead of 53 years. SCE has proven that it can ramp up the OCP program to meet operational and safety needs, and respectfully requests that the Commission approve our total capital request. As mentioned briefly above, one of the reasons why the unit costs for OCP were reduced in 2016 was the effectiveness of our Operational Excellence (OpX) efforts. These efforts are detailed in SCE-02, Volume 1. If the Commission adopts a lower forecast for OCP in years based on the realization of lower unit costs, then a portion of the OpX benefits presented in SCE- 02, Volume 1 will be double-counted. SCE requests that the Commission consider this overlap in adopting forecasts for OCP, and in examining the OpX efforts described in SCE-02, Volume c) TURN s Position TURN proposes a much more limited program with funding of $ million in 2017 and $ million in TURN s main reason for proposing lower funding is that SCE is currently evaluating additional mitigations which may provide lower cost mitigation options in the future. Detailed support for TURN s recommendation is based on their premature use of SCE s PRISM analysis and a misunderstanding of the differences between Reactive and Proactive projects. In addition, TURN proposes a 10% shareholder penalty based on a factually-incorrect characterization of oversized branch line fuses (BLFs) impact on overhead conductor. d) SCE s Rebuttal to TURN s Position (1) TURN s Use of PRISM Results is Premature In SCE-18, Volume 1, SCE rebuts TURN s incorrect use of preliminary PRISM results to propose alternative mitigations for OCP. In this discussion, SCE shows that TURN takes scoring results from a preliminary set of mitigations, some of which have not been tested or 26 See responses to TURN-SCE-016-Q4a and TURN-SCE-042-Q.4a on Appendix pp. A-11 to A See SCE-18, Volume 1 for additional information regarding OpX benefits. 13

17 deployed on SCE s systems, and inappropriately attempts to redesign SCE s Overhead Conductor Program. (2) TURN Does Not Accurately Describe the Differences between Reactive and Proactive Projects TURN provides summary information about quantitative differences between reactive and proactive projects such as miles of reconductor per project (i.e., per circuit). TURN claims that it is less likely that all of the wire being reconductored in proactive projects would be problematic, and that replacing the entire length of a circuit is potentially less effective at reducing safety-related risks than more targeted reconductoring projects. Use of a simplified miles per project metric is an arbitrary method of evaluation, and provides no meaningful measure of comparing effectiveness of proactive and reactive projects. However, more meaningful comparisons can be used to illustrate how the effectiveness of proactive and reactive OCP projects are comparable. As provided in response to CFC data requests, 28 the unit cost of proactive projects are comparable and in fact appear to be slightly lower than the unit cost of reactive projects. This means that on a per-mile basis, there is no significant difference between the two types of projects. Both types of projects replace overhead conductor that is smaller than current standard and subject to damage during fault conditions. Furthermore, both proactive and reactive projects are reactive in a certain sense they both react to historical wire-down events. Reactive projects are initiated by individual wire-down events on a single circuit; proactive projects are initiated on circuits with repeated instances of multiple wire-down events in recent history, in addition to other relevant risk factors such as history of Circuit Breaker (CB) operations, Short Circuit Duty (SCD), and customer density. It is reasonable for SCE to identify more scope on circuits with more problems i.e., more targeted miles in the presence of a history of repeat wire-down events and less scope on circuits with less problems i.e., less targeted miles in the presence of one recent wire-down event. 28 See responses to CFC-SCE-002-Q.23d Supplemental and CFC-SCE-006 Q.7a on Appendix pp. A-17 to A

18 Finally, TURN argues that because proactive projects are more likely to have more conductor miles per project, they are less cost-effective than reactive projects. That is incorrect and unrelated to project effectiveness. 29 One simple numerical illustration can be found by looking at BLF installations per project, also provided (but not discussed) by TURN in Table 5. BLFs have been identified as a mitigation alternative with both a relatively high risk spend efficiency and a high confidence in the assumptions underlying that score. Based on the values provided by TURN in Table 5, proactive projects install more BLFs per project than reactive projects, as well as more BLFs per mile than reactive projects. Therefore, TURN s conclusion could just as logically been the opposite: that proactive projects are more effective not less effective than reactive projects. This illustrates how arbitrary measures such as TURN s suggested amount-of-scope-per-project lead to arbitrary conclusions. TURN s assertion that proactive projects are less effective than reactive projects, simply because proactive projects are larger on a conductor-miles-per-project basis, is simply not supported by the evidence. (3) TURN s Characterization of Oversized BLFs as Resulting in selfinflicted damage to the System is Incorrect TURN characterizes the condition of oversize BLFs as evidence of SCE causing self-inflicted damage to its overhead conductor. TURN uses that characterization to justify their recommendation of a 10% disallowance of any capital spending for the OCP. SCE disagrees; oversized BLFs do not result in self-inflicted damage to the SCE system. TURN has singled out one line in SCE s whitepaper, which stated that oversizing branch line fuses increases the risk of conductor failures. In testimony, TURN describes SCE as stating unequivocally that the common practice of oversizing fuses increases the risk of conductor failure. TURN is misinterpreting SCE s statements; in fact BLFs whether oversized or not do not result in self-inflicted damage to overhead conductor. For clarity, the use of the term oversized in this context is in relation to BLF coordination with upstream protective devices. In response to a TURN data request, SCE listed three potential consequences of such BLF mis-coordination: 1) delayed or increased outage time in locating faults; 2) increased number of customers impacted by an outage; and 3) increased strain on 29 See Exhibit TURN-04, p

19 SCE s system due to additional testing. TURN failed to consider whether these consequences are greater in the oversized BLF case than they would be in the unfused case. Each of these three consequences would exist in equal (or greater) measure in the unfused case. On a system with unfused radials, the installation of BLFs even if oversized as described above still results in system improvements. Typically, such BLFs coordinate with downstream protective devices such as transformer fuses. Also, various types of faults (phase faults, ground faults, etc.) can occur on the system. In SCE s experience it is common for these BLFs to still be coordinated for some types of faults even if not for all types of faults. Finally, BLFs regardless of size provide a convenient means of isolation for trouble crews when performing work such as repairs to transformers downstream of the BLF. Installation of BLFs improve the system; individual BLF mis-coordination with upstream devices may result in certain location-specific sub-optimal benefits but does not cause self-inflicted damage to the system. To illustrate the logical inconsistency in TURN s argument, had SCE not installed any BLFs on unfused radials in years past, and instead relied entirely on only one layer of protection (i.e. circuit breaker or automatic recloser), there would have been no opportunities for BLF mis-coordination and no opportunity for TURN to argue for their OCP disallowance. However, SCE made the decision to install BLFs over the years because of the aggregate benefits that BLFs afford. Two consequences of this decision are clear: first, overall system risks have gone down. Second, opportunities have been created for individual BLF to be resized when conditions change to maintain the optimal coordination with upstream devices. The latter is an inevitable consequence of the former. Yet TURN seeks to penalize SCE for the latter while ignoring the far greater benefits of the former. This is illogical. SCE also points out that TURN demonstrates a lack of understanding of the history of protective relaying in its proposal. Historically speaking, protection coordination has been viewed as both art and science in the industry 30. Advancements in protective technology development and application will, over time, cause protection coordination to become somewhat more 30 For example, see the frequently referenced and highly influential text The Art and Science of Protective Relaying by C. Russell Mason, available at 16

20 science and less art. 31 As such, it is unreasonable for TURN to demand of SCE perfect coordination in previous years; no utility can meet that performance standard. SCE acknowledges that it has improved its BLF sizing process in recent years. However, the previous process for sizing BLFs has resulted in less risk to overhead conductor relative to the unfused alternative. It is unfair to penalize SCE, through TURN s proposed disallowance, for improving its process. e) Consumer Federation of California s (CFC) Position CFC generally supports SCE s OCP but proposes a reduced spending level in 2018 and beyond. CFC s proposed reductions are based on: (1) asserting that the program is still part of a pioneering phase; (2) reducing the program to 250 miles replaced each year, starting in 2018, with a 2.5 percent increase in miles per year through 2029; (3) invoking the need for a strategic spending plan which is not reflected in SCE s proposal; and (4) suggesting alternative rate recovery in the form of a rate rider. f) SCE s Rebuttal to CFC s Position (1) The primary driver of OCP is safety, not reliability CFC has based its recommendation on its flawed assumption that OCP benefits can be divided between safety benefits and reliability benefits. As stated in response to CFC-SCE-002 Q6e, SCE is not able to track what portion of OCP expenditures address safety versus reliability, as the same work addresses both. CFC s analysis approach based on X% for safety and (100-X)% for reliability does not make sense, particularly in light of SCE s response to this data request and in light of the primary driver for OCP. As SCE has stated, the primary driver of OCP is safety, not reliability. 32 Safety and reliability risk analysis has demonstrated that the safety risk caused by energized wire down events is considerable relative to other system risks, and OCP was created to address these safety risks. 33 OCP addresses the problem of overhead wire installations being undersized relative to existing levels of short circuit duty (SCD) through a combination of changes to protection settings, installation of new 31 For example, see the presentation made at November 2013 IEEE Southern Alberta Section PES/IAS Joint Chapter Technical Seminar: Power System Protection Coordination by Mr. Rasheek Rifaat and Dr. Peter Sutherland, available at 32 SCE s response to CFC-SCE-002-Q.6e on Appendix p. A SCE-02, Vol. 8, p. 47, lines

21 protective devices, and the reconductor of small-gauge wire to a larger size that reduces the risk of conductor damage during fault conditions. 34 As previously stated, based on a replacement rate of 300 miles per year, SCE estimates that it will take approximately 53 years to replace all small wire on the SCE system. 35 In fact, if design and construction resources were not a consideration, SCE would prefer to replace small wire at an even faster pace, due to the significant safety risks associated with small wire and wire down. 36 On a related note, CFC has incorrectly characterized SCE s response to the CFC data request regarding performance measures for OCP. CFC states that the lone metrics the company apparently intends to apply for program performance evaluation are SAIDI and SAIFI. In fact, the complete response to CFC-SCE-002 question 9b is quite different. In this response, SCE states that SCE plans to compare wire down data prior to the project and after the project. The impacts will be measured at the circuit level because the projects and other data available are at the circuit level. In addition, SCE plans to evaluate the SAIDI and SAIFI impacts on the overall system. SCE intends to evaluate OCP first and foremost on the primary program driver safety. SCE s interest in also evaluating the SAIDI/SAIFI impacts of OCP is because, at this time, the program impacts in these reliability metrics are not currently known based on underlying simulation model limitations. By extension, since SAIDI/SAIFI is not the driver for OCP, then CFC s statements related to value-of-service considerations are also not relevant to SCE s forecast for OCP. SCE s OCP is intended to address the risks associated with wire-down; value-of-service differentiation among customer classes is not relevant for this purpose. (2) The Rate of SCD Increase in Years Past Should Not be Confused with the Current Level of Small-wire Risk CFC states that increased loads are not in and of themselves sufficient reason for conductor replacement. SCE reiterates that OCP is not driven by increased loads on the system, but by the fact that available SCD in today s system is in excess of the withstand capability of small overhead wire. 34 SCE-02, Vol. 8, p. 48, lines SCE s response to TURN-SCE-016-Q.4a on Appendix p. A SCE s response to TURN-SCE-042-Q.4a on Appendix p. A

22 CFC s use of the phrase circuit loading can create a false implication of a lower sense of urgency because of historical rates of system load growth. The rate of system growth in years past is simply not relevant to the sense of urgency that is needed to address small wire problems today. SCE disagrees with CFC s characterization of OCP as a surprising and counter-intuitive sudden need to address the problem of small wire risks. SCE studies have only recently been able to definitively identify the true nature of the small wire problem that is to say, inconsistencies between small-wire conductor damage curves, SCD levels, and characteristics of available protective devices are what contribute to wire-down events. This new understanding of the extent of existing risks is what has led SCE to introduce OCP at its current forecast levels. It is only counter-intuitive because CFC appears to be making an incorrect correlation between historical rates of SCD increase and levels of present risk related to small wire on the SCE system. No such correlation should be drawn; and it is entirely reasonable that the system can have both a historically low rate of SCD increase and a very high present level of existing small wire risk. (3) CFC Inappropriately Proposes a Rate Rider for OCP SCE rejects the notion that a rate rider is appropriate or practical for OCP. Also, any consideration of rate design should be resolved in Phase 2 of the GRC rather than Phase Capacitor Bank Replacement Program SCE s Capacitor Bank Replacement program replaces failed and obsolete capacitor banks. Capacitor banks are used in our distribution system to regulate the voltage to usable levels by compensating for load inductance. Without adequate numbers of properly operating capacitor banks, the voltage of electricity supplied to many of our customers could drop below required levels, cause damage to customer equipment, and lead to grid reliability issues. To develop the cost forecast for this program, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. 19

23 Table I-4 Capacitor Bank Replacement Program Capital Expenditures 100% CPUC Jurisdictional Nominal $000 Recorded Activity Capacitor Bank Replacement $ 7,957 $ 6,583 $ 8,917 $ 7,814 $ 8,679 $ 7,765 SCE Forecast TURN Forecast Activity Total Total Variance Capacitor Bank Replacement $ 17,156 $ 17,588 $ 34,744 $ 9,024 $ 9,250 $ 18,274 $ (16,470) a) TURN s Position TURN does not object to the overall justification of this program, but does provide alternative forecasts based different unit costs. Also, TURN does not agree with SCE s proposal to increase the annual replacements to a level closer to steady state and proposes an alternative forecast based on the annual average of SCE replacements from b) SCE s Rebuttal to TURN s Position SCE concedes to TURN s proposal to use SCE s 2014 recorded unit cost to forecast this program, since the ratio of overhead to underground replacements in 2016 is similar to that experienced in However, TURN s recommendation to replace Capacitor Banks at a rate equal to the average will not allow SCE to address the oldest capacitor banks in our inventory. SCE calculates the long-term steady state replacement rate for Capacitor Banks to be 450 per year. 37 As such, SCE s forecast of 350 units per year 38 strikes a reasonable balance between current inventory, historical replacement rates, and the need to advance to the long-term steady state replacement rate. Considering this discussion, SCE proposes an alternative forecast that adopts TURN s unit costs and SCE s units of replacements. This is shown below in Table I-5. This new forecast is more than $7 million lower than SCE s original forecast presented in Exhibit SCE-02, Volume See response to TURN-SCE-031-Question 37 a-d on Appendix pp. A-21 to A See responses to TURN-SCE-031-Question 40a and 40b on Appendix pp. A-23 to A24. 20

24 Table I-5 Revised Capacitor Bank Replacement Program Capital Forecast 100% CPUC Jurisdictional Nominal $ Total Unit Rate $ 39 $ 40 Units Annual Forecast $ 13,674 $ 14,018 $ 27,692 21

25 Appendix A IR Appendix A Data Request Responses

26 Question 02.a Supplemental: Originated by: Greg Wilson Exhibit Reference: SCE-02, Vol. 08 SCE Witness: J. R. Goizueta Subject: Substation Switchrack Rebuild Please provide the following: Southern California Edison 2018 GRC A DATA REQUEST SET ORA-SCE-131-GAW To: ORA Prepared by: Matt Stumpf Title: Project Manager Dated: 01/10/2017 Beginning on page 92 of Exhibit SCE-02, Vol. 08, SCE presents its testimony regarding its Substation Switchrack Rebuild program. ORA is in the process of reviewing this testimony, and has the following questions regarding certain portions of SCE s analysis. 2. Since SCE has classified the Substation Switchrack Rebuild program as a new program, ORA understands that little recorded data will exist. Based on SCE s testimony, it appears that 2015 is the only recorded year available. However, it does not look as if the recorded data in SCE s testimony (see Table III-26 on page 95) agrees with the recorded data in the workpapers (see page 209). In fact, while the testimony shows only one year of recorded expenditures, the workpapers show two. In addition, the recorded costs shown in the testimony and workpapers do not agree with one another. a. Please explain why there is a discrepancy between the testimony and the workpapers for recorded switchrack rebuild data. Response to Question 02.a Supplemental: To develop the Switchrack Rebuild unit costs (shown on page 209 of workpapers for SCE-02, Volume 8), SCE analyzed completed Circuit Breaker and Transformer Bank projects that included Switchrack Rebuild scope, but recorded to other WBS Elements. As the Switchrack unit cost workpaper shows, SCE completed 5 Switchrack Rebuilds in 2014 and 7 more in SCE does not adjust recorded capital costs, so previous costs for these projects are included in the historic costs for transformers and circuit breakers. The costs shown in Table III-26 on page 95 of testimony are for projects that started under the new WBS element for Switchrack Rebuilds. Given this approach, the historical costs used in the workpaper for the unit costs do not and should not match the amounts shown in Table III-26. When developing the unit costs for transformer banks and circuit breakers, we excluded the A-1

27 closed work orders that had Switchrack Rebuilds. This was done to help ensure our transformer bank and circuit breakers unit costs are not overstated. SCE has identified an error for Circuit Breaker and Transformer Bank workpapers (see supplemental to TURN-SCE-016 Q14a and Q14d). SCE discovered that the unit cost for Circuit Breakers and Transformer Banks included Switchrack Rebuild projects. These Switchrack Rebuild projects have been removed from the Circuit Breaker and Transformer Bank Unit Cost and have been added to the Switchrack Rebuild unit cost in workpapers for SCE-02, Vol 8 on page 209. Please see attached workpaper for the updated unit cost for Switchrack Rebuild. Additionally this update has impacted the forecast dollars for Table III-26 on page 95 and for Figure III-36 on page 96. Please see attachments for updates to Table III-26 and Figure III-36. SCE apologizes for the inconvenience. A-2

28 Question 11.a Supplemental: Southern California Edison 2018 GRC A DATA REQUEST SET TURN-SCE-016 To: TURN Prepared by: Matt Stumpf Title: Project Manager Dated: 11/30/ Table III-19 references footnote 53, which indicates that pages of the workpapers for SCE-02, Vol. 8 contain information regarding A & B Bank Unit Costs. For the tables on page 128 of the workpapers: a. Please explain why the counts for each year are different for many of the kv levels as compared to the counts set forth in Table III-19 of the testimony for the same levels for the same year. Please also reconcile the counts as between the 2018 testimony and workpapers. Response to Question 11.a Supplemental: SCE discovered that there was an error when developing the unit counts for Table III-19, which should have matched the workpaper unit counts. Also, SCE identified two errors in the unit cost workpaper on page 128. The first error was incorrect labeling in some of the tables that had titles that showed 4kV which should have been 12kV, and 12/16kV which should have been 16kV. The second error occurred when recorded nominal costs for 220kV banks were escalated to 2015 constant dollars, using the escalation factors for distribution equipment instead of the transmission escalation factor. This second error impacts the forecast amounts for the 220kV transformer banks. SCE has identified an additional error in the Transformer Bank workpapers for SCE-02, Vol 8 on page 128. In response to another party s data request, SCE discovered that the unit cost for Transformer Banks included some work orders for Switchrack Rebuild projects. These Switchrack Rebuild work orders have now been removed from the Transformer Bank unit cost analysis. Additionally, this update has impacted the forecast dollars and historical counts for for Table III-19 on page 73 of testimony and has also impacted the forecast dollars for Figure III-27 on page 72 of testimony. Updates for Table III-19 and Figure III-27. Attached are the updated files. SCE apologizes for the inconvenience. A-3

29 Question 14.a Supplemental: Southern California Edison 2018 GRC A DATA REQUEST SET TURN-SCE-016 To: TURN Prepared by: Matt Stumpf Title: Project Manager Dated: 11/30/ Table III-23 on page 86 of SCE-02, Vol. 8 references footnote 61, which indicates that pages of the workpapers for SCE-02, Vol. 8 contain information regarding Circuit Breaker Unit Costs. For the tables on page 166 of the workpapers: a. Please explain why the counts for each year are different for many of the kv levels as compared to the counts set forth in Table III-23 of the testimony for the same levels for the same year. Please also reconcile the counts as between the 2018 testimony and workpapers Response to Question 14.a Supplemental: SCE discovered that there was an error when developing the unit counts for Table III-23, which should have matched the workpaper unit counts. SCE has identified an additional error for Table III-23 in SCE-02, Vol 8. In response to another party s data request, SCE discovered that the unit cost for Circuit Breakers included Switchrack Rebuild projects. These Switchrack Rebuild projects have been removed from the Circuit Breaker unit cost analysis. Please see supplemental response to TURN-SCE-016 Q14d for the updated unit cost analysis. This update impacts the forecast dollars and historical counts for for Table III-23 on page 86 of testimony and has also impacted the forecast dollars for Figure III-32 on page 84 of testimony. Please see attachments for updates to Table III-23 and Figure III-32. SCE apologizes for the inconvenience. A-4

30 Question 14.d Supplemental: Southern California Edison 2018 GRC A DATA REQUEST SET TURN-SCE-016 To: TURN Prepared by: Matt Stumpf Title: Project Manager Dated: 11/30/ Table III-23 on page 86 of SCE-02, Vol. 8 references footnote 61, which indicates that pages of the workpapers for SCE-02, Vol. 8 contain information regarding Circuit Breaker Unit Costs. For the tables on page 166 of the workpapers: d. Please provide the text for footnote 6 on page 166 of the workpapers. Response to Question 14.d Supplemental: The use of a footnote numbered 6 is an error in labeling the footnotes in this workpaper. The correct references are described below and included on a revised workpaper. SCE apologizes for this error. The title for the Nominal Unit Cost table was mislabeled with footnote references to both 4 and 5. The footnote reference for this should have only been to footnote 4. The title for the Unit Cost Used for Forecasting table was mislabeled with a footnote reference to footnote 6. The footnote reference for this should have been to footnote 5. SCE has identified an additional error for the Circuit Breaker workpaper in SCE-02, Vol 8, Page 166. In response to another party s data request, SCE discovered that the unit cost for Circuit Breakers included Switchrack Rebuild projects. These Switchrack Rebuild projects have been removed from the Circuit Breaker unit cost analysis. Please see attached workpaper for the updated unit cost for circuit breakers. SCE apologizes for the inconvenience. A-5

31 Southern California Edison 2018 GRC A DATA REQUEST SET TURN-SCE-108 To: TURN Prepared by: Shan Chien Title: Senior Engineer Dated: 04/03/2017 Question 02: SCE-2, Vol. 8 Infrastructure Replacement Worst Circuit Rehabilitation (WCR) Program 2. At p. 25 of the workpapers, SCE states: For completeness of the analysis, this failure rate model also credits effects of the cable infrastructure replacement (IR) program implemented since a. Please describe in detail how the model credits effects of the cable infrastructure replacement (IR) program implemented since b. Please describe how it is possible that the model credits effects of the cable infrastructure replacement (IR) program implemented since 2005, given that SCE does not know for how much cable of each type (e.g., XLPE, etc.) was replaced because of failure (vs. pre-emptive replacement, etc.), as indicated in SCE s response to TURN-SCE b, which states that SCE does not know the number of failures by insulation type. c. Please provide any workpaper that is available outside of the Weibull++9 software program that shows the credit of the cable IR program to the reliability model. Response to Question 02: a. The Weibull analysis described in pages of the workpapers was based on cable retirement/replacement records based on accounting inventory data, not ODRM failure data. This accounting data includes information related to cable inventory based on removal types (i.e. XLPE versus PILC etc.) but does not include information related to cable removal reason (i.e. IR versus failure etc.). Therefore, this removal data needs to be adjusted based on quantities of cable removals for IR. The "credit" referred to on page 25 of the workpapers refers to this adjustment process. This adjustment process makes two assumptions. The first assumption is related to the age profile of cable removed from inventory due to IR; it is assumed that such cable is from older cable with higher inventory. The second assumption is related to the future failures of cable removed from inventory due to IR; it is assumed that such cable will have failed within four years of removal. A-6

32 b. As discussed in SCE's response to TURN-SCE-031 question 4b, DTOM/ODRM data does not classify individual cable outage events (i.e. "failures") by cable type. As discussed in SCE's response to part a, the accounting data used for cable retirement/replacement records does not have information related to cable removal reason (i.e. IR versus failure etc.). The model described in workpapers pages uses assumptions in order to adjust (i.e. "credit") cable retirement/replacement records for the effect of cable retirement/replacements through the IR program. For additional details see response to part a. c. SCE has not prepared such a workpaper. For additional details of the "credit" of the cable IR program to the reliability model, see response to part a. A-7

33 Southern California Edison 2018 GRC A DATA REQUEST SET ORA-SCE-110-GAW To: ORA Prepared by: Jamal Cherradi Title: Manager Dated: 01/02/2017 Question 05.c: Originated by: Greg Wilson Exhibit Reference: SCE-02, Vol. 08 SCE Witness: J. R. Goizueta Subject: Historical and Forecast WCR Cable Replacements Please provide the following: Beginning on page 13 of Exhibit SCE-02, Vol. 08, SCE presents its testimony regarding its Worst Circuit Rehabilitation (WCR) program. ORA is in the process of reviewing this testimony, and has the following questions regarding certain portions of SCE s analysis. 5. On pages 24 and 25 of Exhibit SCE-02, Vol. 08, SCE provides two graphs (Figures III-9 and III-10) that show how SAIDI and SAIFI levels would be impacted by various levels of preemptive cable replacements. The testimony accompanying these two graphs, as well as the descriptions contained in their legends, indicate that only WCR replacements were considered in the development of these graphs. In prior GRCs, SCE has provided similar graphs that show how SAIDI and SAIFI would fluctuate over time for various replacement levels. However, in these prior GRCs, it is ORA s understanding that the graphs also reflected the impact of Cable-In-Conduit (CIC) replacements. Stated another way, prior graphs showed how SAIDI and SAIFI were impacted by total cable replacements (i.e., the sum of both WCR and CIC replacements). c. In this GRC, SCE has introduced what appears to be a new capital program, the Overhead Conductor Program. How will the implementation of this new program impact SAIDI and SAIFI levels in the future? Response to Question 05.c: At this time, SCE is not able to perform analysis similar to Figures III-9 and III-10 to predict long-term SAIDI/SAIFI impacts of OCP. Overhead conductors, when compared to underground cable, have a much wider variety of factors that impact failure rates such as physical exposure (e.g., 3 rd party contact, mylar balloons, vegetation), environmental exposure (weather, fire), and electrical exposure (wire size, fault A-8

34 history, and short circuit duty). As a result, SCE does not have an equivalent Weibull model for age-based probability of failure of overhead conductor. Since simulations to project impacts of equipment aging on future reliability depend on such models, SCE is not able to simulate the SAIDI/SAIFI benefit of OCP in the future using these techniques. In fact, the no WCR program curves (i.e., Figures III-4 and III-5) do not show any changes in long-term SAIDI/SAIFI trends due to overhead conductors. These curves were made with a simplified assumption of a uniform, time-independent failure rate for all overhead conductors. In other words, overhead conductor reliability performance was held constant for all simulations in this rate case in order to isolate anticipated reliability impacts of the WCR program alone. Therefore, no conclusions can be reached from any of the curves in Figures III-9 and III-10 regarding SAIDI/SAIFI trends related to overhead conductors. A-9

35 Southern California Edison 2018 GRC A DATA REQUEST SET CFC-SCE-002 To: CFC Prepared by: Bryon Watkins Title: Engineer Dated: 12/20/2016 Question 06.e: 6. Table III-12 provides historical and forecast figures for the Overhead Conductor Program: e. What proportion of proposed OCP expenditures are intended primarily to address safety, and what proportion is meant primarily to address reliability? Please explain. Response to Question 06.e: SCE is not able to track what portion of OCP expenditures address safety vs. reliability risk as the the same work addresses both. The primary driver of OCP is to address the safety risk of wire-down events. While not the primary driver of OCP, the OCP work also results in reliability benefits. See response to CFC-SCE-002-Q9.b for additional details regarding how SCE will evaluate reliability impacts of OCP. A-10

36 Southern California Edison 2018 GRC A DATA REQUEST SET TURN-SCE-016 To: TURN Prepared by: Matt Stumpf Title: Project Manager Dated: 11/30/2016 Question 04.a: 4. In Table III-12 on page 49 of SCE-02, Vol. 8, SCE has the historical and forecast spending for the Overhead Conductor Program. a. Please identify by page and line number where in SCE s testimony and workpapers the utility explains the reasonableness of its forecast of 320 circuit-miles in 2016, and 300 circuit-miles in each year from Response to Question 04.a: As of 2014, SCE had approximately 16,000 circuit miles of small conductor (which met standards at the time of installation) that does not meet current design standards. Small conductor is at higher risk of being damaged during faults. At a replacement rate of approximately 300 miles per year, it would take approximately 53 years to replace all small conduction on SCE s system. Given other necessary work that utilizes the same resources and can occur in the same geographic area, SCE believes that our request of 320 miles in 2016 and 300 annually from is reasonable and justified. This annual level of work balances costs, resources, and impacts to customers, while SCE s prioritization of scope selection maximizes the impact of annual OCP work. SCE s OCP testimony in SCE02, Volume 8 on pages discusses the justification of the program and how scope is selected and prioritized. A-11

37 Southern California Edison 2018 GRC A DATA REQUEST SET TURN-SCE-042 To: TURN Prepared by: Rob Tucker Title: Senior Manager Dated: 02/01/2017 Question 04.a: Infrastructure Replacement (SCE-2, Vol. 8) 4. SCE s response to DR TURN 4.a. states, in part, that the proposed annual level of work for the Overhead Conductor Program balances costs, resources, and impacts to customers, while SCE s prioritization of scope selection maximizes the impact of annual OCP work. a. Please provide documentation of SCE s analysis that led the utility to conclude that its proposed forecast of 320 circuit-miles in 2016 and 300 circuit-miles in each year from best balances costs, resources, and impacts to customers. In particular, provide the analysis SCE performed before submitting its testimony that led the utility to conclude that its proposed figures achieve that balance better than, for example, forecasts of 20 miles less each year, or forecasts of 20 miles more each year. Response to Question 04.a: SCE s response to DR TURN-SCE a. states that the annual level of [Overhead Conductor Program] work balances costs, resources and impacts to customers and does not qualify that statement with the term best as stated in DR TURN-SCE a. SCE has proposed 300 miles per year from 2017 to 2020 because it is a rate of work that SCE can steadily execute without over-burdening current design and construction resources, while addressing safety risk. In fact, it will take SCE approximately 53 years to reconductor all small wire on the SCE system at a replacement rate of 300 miles per year (see response to TURN-SCE a.). If design and construction resource constraints were not an issue, SCE would prefer to replace small wire at a faster pace, due to the significant problems associated with small wire and wire down (see for example: SCE-02 Volume 08 beginning on page 47 line 23; SCE-02 Volume 08 beginning on page 49 line 1; SCE-02 Volume 01 beginning on page 43 line 24; and SCE-02 Volume 01 Appendix beginning on page 6 line 7 including Table I-2). If the proposed forecast is not best, it can only be made better through increases not decreases in the OCP forecast. A-12

38 Southern California Edison 2018 GRC A DATA REQUEST SET CFC-SCE-005 To: CFC Prepared by: Bryon Watkins Title: Engineer Dated: 03/03/2017 Question 09.b: 9. SCE-14, page 87, describes expected risk reduction effectiveness for the OCP program: Since we do not have field data on the impact of the various mitigation options, SMEs collaborated to evaluate the mitigation options and estimated the expected impact of each mitigation option in reducing TEF, CP and/or CI based on their professional judgment. Most of the mitigation alternatives considered for overhead conductor risks are primarily expected to reduce the probability of wire down events, or TEF. The response to CFC-SCE-002, Question 9b, described how SCE will measure OCP project impacts: SCE plans to compare wire down data prior to the project and after the project. The impacts will be measured at the circuit level because the projects and other data available are at a circuit level. In addition, SCE plans to evaluate the SAIDI and SAIFI impacts on the overall system. b. Considering the proposed means of evaluating the accuracy of the estimates versus the actual impact of the mitgations, at what future time would SCE be able to use observed mitigation impacts, rather than SME estimated impacts? Please explain, including how soon after a project s completion the company will be able to evaluate its mitigation impacts. Response to Question 09.b: Because circuit-level performance metrics are typically quantified in annual terms, SCE wants a minimum of one year of post-project operational experience before beginning to compare pre-project and post-project performance. Note that SCE's methodology for ranking Worst Performing Circuits (WPCs) is based on 3-year circuit performance data. Therefore, SCE expects that trends in observed mitigation impacts will be most clear after three years. A-13

39 Southern California Edison 2018 GRC A DATA REQUEST SET TURN-SCE-059 To: TURN Prepared by: Bryon Watkins Title: Engineer Dated: 02/24/2017 Received Date: 02/24/2017 Question 01.b.iii: Infrastructure Replacement (SCE-2, Vol. 8) Overhead Conductor Program (OCP) 1. Referencing the Excel file, TURN-SCE-031 Q.29.d.i-vi Att-DR-31-Q29d that SCE attached b. For each of the completed projects (as indicated within the table in the Completed Project column) please identify the following: iii. The cost of the mitigation. If available, please disaggregate the total into reconductoring and BLF activities. Response to Question 01.b.iii: Please see attachment A _TURN_DR-59-Q1biii.xlsx. Disaggregated totals for reconductoring and BLF activities are not available. A _TURN_DR-59-Q1biii.xlsx A-14

40 Southern California Edison 2018 GRC A DATA REQUEST SET TURN-SCE-059 To: TURN Prepared by: Matt Stumpf Title: Project Manager Dated: 02/24/2017 Question 12.a: Infrastructure Replacement (SCE-2, Vol. 8) Overhead Conductor Program (OCP) 12. In TURN DR 16-4.d, SCE identifies an expenditure and mile count through October 2016 of $76.4 million and miles. a. Please update the recorded cost and miles that SCE identified in TURN 16-4.d, but for through the end of 2016 with the understanding that they might change, once SCE finalizes them in the first quarter of Response to Question 12.a: Please see response to ORA-SCE-108-TXB-Supplemental -Revised Question 2 for 2016 recorded expenditures of $ million. In 2016, the overhead conductor program replaced circuit miles. A-15

41 Southern California Edison 2018 GRC A DATA REQUEST SET TURN-SCE-059 To: TURN Prepared by: Matt Stumpf Title: Project Manager Dated: 02/24/2017 Question 12.c: Infrastructure Replacement (SCE-2, Vol. 8) Overhead Conductor Program (OCP) 12. In TURN DR 16-4.d, SCE identifies an expenditure and mile count through October 2016 of $76.4 million and miles. c. TURN calculates on the basis of TURN-SCE-031 Q.29.d.i-vi Att-DR-31-Q29d that the cost and unit-count estimates for projects completed in 2016 were $101.5MM and 141, respectively. This represents a unit-cost estimate of about $722,000 (i.e., $101.5MM divided by 141 units). (Please note: this calculation is for all projects completed in 2016, regardless of whether they were originally scoped for 2016 or 2017.) Please identify and explain each factor that contributed to the difference between the estimated unit cost of the work that was completed in in 2016 and the unit cost that identified in Part (b) above. (TURN understands that there may be minor differences between the two numbers that owe to minor differences related to when projects might close and the like; we would like the response to address the magnitude of the relative difference.) Response to Question 12.c: The attachment referenced is based on the planned work and is not completed. Those forecasts included specific scope that includes costs for conductor replaced and BLFs. Based on the references, SCE is unable to match TURN s figures in this question. Based on the attachment, the 2016 scope was approximately $142M for miles and over 2,763 BLFs. SCE forecasting methodology utilizes an overall unit cost that takes total costs for projects that have both conductor replacement and BLFs and forecasts a total number of conductor miles to achieve steady state replacement levels. As specific scope is identified, the individual project costs will vary compared to the forecast unit cost. See response to TURN-SCE-031-Q26.e for additional detail. A-16

42 Question 23.d.Supplemental: Southern California Edison 2018 GRC A DATA REQUEST SET CFC-SCE-002 Q.23.d Supplemental To: CFC Prepared by: Matt Stumpf Title: Project Manager Dated: 01/26/ SCE02V08 page 48 describes how the OCP program identifies iscuss replacement decision analytics. Page 12 states: SCE has approximately 106,000 conductor miles of primary overhead conductor in the service territory. On this system, in 2015, SCE experienced 1,039 wire down events associated with distribution primary overhead conductor. To address these challenges, the OCP program includes preemptive mitigations like conductor replacement and inspection driven mitigations, and reactive in-service failure mitigations to address safety and reliability needs at a reasonable cost... Historically, SCE has performed work to address these risks. This work was conducted in accordance with SCE s Distribution Inspection and Maintenance Program (DIMP), which conducts overhead detailed inspections. The OCP program was launched as a concentrated effort to specifically focus on causes and mitigations leading to a specific strategy to address wire down risks. The following is an excerpt from the table provided in WPSCE02V08 Workpaper OCP Scope for 2016 and 2017 : d. Please provide a version of the OCP Scope table, but with columns added showing (for each circuit listed) the circuit-miles to be replaced, the method of candidate identification (i.e., inspections, engineering calculation), the ages of conductor being replaced, and the A-17

43 number of splices that will be eliminated. Response to Question 23.d.Supplemental: Please find the attached spreadsheet CFC-SCE-002-Q23.d which shows the requested information to the extent it is available. The age of overhead conductor being replaced is not available as discussed in SCE's response to CFC-SCE-002-Q6d. Splice data is also not included in the spreadsheet because SCE does not currently have splice data that is readily available for scoping of OCP projects. SCE began tracking splice data in 2015, however the information has not been mapped to current projects. Some circuits did not have any conductor replacement work, but did have other scope, such as installation of protective devices, or reconfiguring of the circuit. The table below is a summary of the detail included in the attachment. The reactive work refers to the follow-up work that is scoped after the initial repairs are performed to restore service (see page 51 of SCE-02, Volume 8 for more detail regarding reactive work). Proactive work is identified based on engineering analysis of historical circuit performance and other circuit specific information as discussed in testimony. As discussed in testimony, the scope of OCP includes more than just replacing conductor, such as installing protective devices. As a result, some OCP projects may not include any conductor replacements, in particular for the follow-up reactive work where the conductor may have been replaced during service restoration. Additionally, while the total forecast is based on a number of circuit miles and average cost per mile, the actual scope may include less miles replaced than forecast, based on specific scope identified during the engineering process. The specific scope can also lead to differences in cost per mile for each project compared to the average cost per mile used to forecast the 5 years of OCP capital expenditures. A-18

44 Southern California Edison 2018 GRC A DATA REQUEST SET CFC-SCE-006 To: CFC Prepared by: Bryon Watkins Title: Engineer Dated: 04/05/2017 Question 07.a: 7. The response to CFC-002 Question 23d, described the cost differences between proactive and reactive OCP projects: The reactive work refers to the follow-up work that is scoped after the initial repairs are performed to restore service... Proactive work is identified based on engineering analysis of historical circuit performance and other circuit specific information as discussed in testimony. As discussed in testimony, the scope of OCP includes more than just replacing conductor, such as installing protective devices. As a result, some OCP projects may not include any conductor replacements, in particular for the follow-up reactive work where the conductor may have been replaced during service restoration. Additionally, while the total forecast is based on a number of circuit miles and average cost per mile, the actual scope may include less miles replaced than forecast, based on specific scope identified during the engineering process. The specific scope can also lead to differences in cost per mile for each project compared to the average cost per mile used to forecast the 5 years of OCP capital expenditures. a. The table of figures provided in the response to CFC-002 Q23d indicates that proactive projects (at roughly $650K/mi) are less costly than reactive projects (at roughly $720K/mi). Is the differential in average cost due to there being more repair activities, involving added scope uncertainties, for reactive projects, as opposed to replacement? Please comment. Response to Question 07.a: Reactive OCP projects are driven by recent wire down events and resolve portions of circuits related to the wire down event. Proactive OCP projects do the same work, but address the entire circuit instead of portions of the circuit. SCE believes that by addressing the entire circuit, efficiencies are gained through set up and take down tasks, such as traffic control or use of cranes. A-19

45 Southern California Edison 2018 GRC A DATA REQUEST SET CFC-SCE-002 To: CFC Prepared by: Bryon Watkins Title: Engineer Dated: 12/20/2016 Question 06.e: 6. Table III-12 provides historical and forecast figures for the Overhead Conductor Program: e. What proportion of proposed OCP expenditures are intended primarily to address safety, and what proportion is meant primarily to address reliability? Please explain. Response to Question 06.e: SCE is not able to track what portion of OCP expenditures address safety vs. reliability risk as the the same work addresses both. The primary driver of OCP is to address the safety risk of wire-down events. While not the primary driver of OCP, the OCP work also results in reliability benefits. See response to CFC-SCE-002-Q9.b for additional details regarding how SCE will evaluate reliability impacts of OCP. A-20