(U 338-E) 2018 General Rate Case A Workpapers. T&D- Infrastructure Replacement SCE-02 Volume 08

Transcription

1 (U 338-E) 2018 General Rate Case A Workpapers T&D- Infrastructure Replacement SCE-02 Volume 08 September 2016

2 Workpaper Southern California Edison / 2018 GRC I. Infrastructure Replacement A. Content and Organization SCE s distribution and substation infrastructure includes major equipment, such as transformers, switches, circuit breakers, capacitors, automatic reclosers, cable, and conductors. Typically, these assets operate for many years, some over fifty, before they wear out. However, eventually every part of the infrastructure will wear out and need to be replaced. Every piece of equipment must be replaced repeatedly over the life of the system. Ongoing replacement of every component as it wears out is an inescapable part of maintaining a distribution and substation system. Assets are typically replaced on a preventive basis based on the results of visual inspections, or based on engineering analysis that predicts an in-service failure. The costs associated with replacing these assets are recorded as Distribution and Substation Maintenance 1 and Infrastructure Replacement (IR), depending on whether the need for replacement was identified through visual inspection or engineering analysis. Programs that replace equipment based on engineering analysis are described in this volume. This volume includes an overview of how our capital requests address risks related to operating SCE s distribution and substation system. 2 A summary of the regulatory background and policies that influence how we implement Infrastructure Replacement is also included, as well as requirements identified in the 2015 GRC decision for implementation in this 2018 GRC. The Infrastructure Replacement capital activities requested in this volume are shown in Table I-1 below. 1 Please see Mr. Stark s testimony in Exhibit SCE-02, Vol. 04 for Distribution Capital Maintenance, and Mr. Flores s testimony in Exhibit SCE-02, Vol. 06 for Substation Breakdown Maintenance and Substation Preventive Capital Maintenance. 2 Please see Mr. Woods testimony in Exhibit SCE-02, Vol. 1 for detailed risk analysis on several of the capital requests presented in this Volume. 1

3 2 Workpaper Southern California Edison / 2018 GRC B. Summary of Capital Request The tables below summarize the Distribution and Substation capital activities discussed in this volume, and their associated expenditures. Table I-1 Infrastructure Replacement Capital Activities Forecast (Total Company Nominal $000) Description Worst Circuit Rehabilitation 137, , , , ,305 Cable Life Extension 21,570 23,402 23,991 24,744 25,530 CIC Replacement 26,528 31,142 41,643 42,949 44,315 Overhead Conductor Program 142, , , , ,466 Underground Oil Switch Replacement 10,923 11,150 12,701 13,099 13,516 Capacitor Bank Replacement 12,005 17,156 17,588 18,140 18,717 Automatic Reclosure Replacement 2,565 2,310 2,368 2,442 2,520 PCB Transformer Replacement 1,385 1,413 1,449 1,494 1,542 Substation Transformer Bank Replacement 121,753 66,932 68,601 70,590 89,844 Substation Circuit Breaker Replacement 45,140 46,833 47,994 49,418 51,403 Substation Switchrack Rebuilds 5,244 16,060 16,464 16,981 17,521 Total $527,142 $475,590 $498,520 $513,915 $547,679 Table I-2 Infrastructure Replacement Capital Activities Forecast (CPUC-Jurisdictional Nominal $000) Description Worst Circuit Rehabilitation 137, , , , ,305 Cable Life Extension 21,570 23,402 23,991 24,744 25,530 CIC Replacement 26,528 31,142 41,643 42,949 44,315 Overhead Conductor Program 142, , , , ,466 Underground Oil Switch Replacement 10,923 11,150 12,701 13,099 13,516 Capacitor Bank Replacement 12,005 17,156 17,588 18,140 18,717 Automatic Reclosure Replacement 2,565 2,310 2,368 2,442 2,520 PCB Transformer Replacement 1,385 1,413 1,449 1,494 1,542 Substation Transformer Bank Replacement 90,582 50,999 52,276 53,835 55,455 Substation Circuit Breaker Replacement 39,861 40,957 41,974 43,239 44,544 Substation Switchrack Rebuilds 5,244 16,060 16,464 16,981 17,521 Total $490,692 $453,782 $476,174 $490,980 $506,432 2

4 Workpaper Southern California Edison / 2018 GRC C. Regulatory Background/Policies Driving SCE s Request Over the last several rate case cycles, SCE has demonstrated the need to preemptively replace infrastructure when our risk and reliability analysis determines that an in-service failure is likely to occur, the consequences are high, and the required mitigation justifies the cost. The Commission has agreed with the overall need for an IR program, and stated in Decision : [t]he key decision before the Commission is how rapidly to replace infrastructure considering safety, reliability, and cost, in addition to other factors. 3 As shown in Figure I-1, the Commission authorized $309 million in 2015, which was approximately 88% of SCE s request. In addition to D , the Commission has also established new reliability reporting requirements through D , which closed CPUC proceeding R (Order Instituting Rulemaking Regarding Policies, Procedures and Rules for Reliability Reporting Pursuant to Public Utilities Code Section ). As part of these new reporting requirements, the Commission directed SCE to file with the Commission its plans to improve the reliability performance of repeat Worst Performing Circuits (WPC) as part of its annual reliability report. This regulatory requirement adds another driver for SCE s Worst Circuit Rehabilitation (WCR) program, which is the program that SCE uses to identify and address the WPCs within its service territory. D. Compliance Requirements Decision The Commission s 2012 GRC Decision, D , identified requirements SCE must address when making its IR request in subsequent GRCs. Below, we list these requirements and explain where they are addressed in this Volume: SCE should carefully document the data collection from the Cable-in-Conduit (CIC) pilot program, and other efforts it undertakes to develop a best practice and most cost-effective method for replacements. This information shall be submitted 3 See page 62 of D

5 4 Workpaper Southern California Edison / 2018 GRC to support future GRC requests in this category to assist the Commission and to illustrate that ratepayers achieved value from SCE s lessons learned. 4 o SCE addresses this requirement in Section III. A. 2 Cable Life Extension Program. SCE should document the B-Bank transformer replacements performed in this rate cycle and submit the names, locations, and ages of the replaced transformers to support future GRC requests in this category. 5 o SCE addresses this requirement in Section III. B. 1 Transformer Bank Replacement Decision The Commission s 2015 GRC Decision identified new requirements for SCE to address when making its IR request in this GRC. Below, we list these requirements and explain where they are addressed in this Volume: To the extent that SCE proposes to replace untested cable (either mainline or CIC) in its next GRC, it must clearly explain why a testing-based replacement program is not more cost-effective 6 o SCE addresses this requirement in Section III. A. 2 Cable Life Extension Program. In the next GRC, SCE should analyze the Distribution Circuit Breaker preemptive replacements in combination with other types of replacements. 7 o SCE addresses this requirement in Section III. B. 2 Circuit Breaker Replacement. 4 D Conclusions of Law 106 and Section 5.5.2, page D Conclusions of Law See page 72 of D See page 80 of D

6 Workpaper Southern California Edison / 2018 GRC E. Comparison of 2015 Authorized to Recorded Figure I-1 summarizes SCE s Infrastructure Replacement Capital request in the 2015 GRC, the 2015 GRC authorized amounts, and variance between recorded and authorized amounts. The major drivers for the increase over the authorized amount is the new Overhead Conductor Program, Worst Circuit Remediation, Transformer Banks, and Circuit Breakers. These increases are partially offset by lower spending in Cable Life Extension and CIC replacements. Figure I-1 Infrastructure Replacement GRC Authorized Variance Summary 2015 Capital (CPUC-Jurisdictional Nominal $Millions) 8 Refer to WP SCE-02, Vol. 08, pp (Capital Authorized vs. Recorded). 5

7 6 Workpaper Southern California Edison / 2018 GRC II. Need for Infrastructure Replacement A. Infrastructure Replacement Balances Safety, Reliability, and Cost Every piece of equipment will eventually wear out and need to be replaced. Three options are available for dealing with equipment as it wears out: Run-to-failure, i.e., wait until the equipment fails in service. This subjects the affected customers to an unscheduled interruption of power, and, in some cases, could affect worker or public safety, and we must then replace the equipment on an unplanned basis; Inspection-driven replacement, i.e., replace the component prior to in-service failure after visual inspections identify observable indications of imminent failure; Preemptive replacement, i.e., replace the component prior to in-service failure based on engineering analyses. The optimum strategy for replacing aging equipment varies with equipment type and depends on: (1) how thoroughly the condition of the equipment can be assessed by inspection; (2) the consequences of an in-service failure in terms of cost, reliability, and safety; and (3) the ability to conduct a predictive engineering analysis to calculate the probability and consequence of failure. For equipment where in-service failures would have minimal consequences, a run-to-failure strategy can be the preferred approach. For such equipment, there may be little benefit derived from preemptive replacement prior to actual end of life. In-service failures of capacitor banks are rarely violent, rarely result in customer interruptions, and usually pose little threat to system reliability if replaced in a timely manner after they fail. Therefore, SCE s capacitor replacement program uses a runto-failure approach. For equipment where in-service failures have more significant consequences, visual analysis (i.e. inspections) can lead to replacement decisions. Most equipment can benefit from routine visual inspections. Such inspections can reveal external signs, such as external corrosion or leaking oil, of increased probability of failure and can result in equipment being replaced prior to an in-service failure. 6

8 Workpaper Southern California Edison / 2018 GRC However, visual inspections cannot determine the probability of failure of internal working components of equipment, and these failures make up the majority of in-service equipment failures that impact the reliability of the system. Inspections can also miss rapid-onset conditions where equipment deterioration occurs between inspection cycles. Therefore, while inspection programs can prevent some in-service equipment failures, they cannot prevent all of them. Relying on visual analysis alone would result in an increased number of equipment failures, leading to power outages and safety hazards. In combination with visual inspection, SCE performs engineering analyses to evaluate the internal condition of certain types of equipment. This predictive engineering analysis approach is the only method practicable for equipment where in-service failures cannot be fully identified through visual inspection programs alone. Predictive engineering analyses should identify equipment approaching the end of its service life, where failure could cause significant unnecessary expenses, prolonged and/or widespread power outages, and/or injury to our employees and the public. Our infrastructure replacement program is a necessary part of providing safe and reliable electric service at a reasonable cost. B. Equipment Failure Probabilities Increase as Equipment Ages The likelihood that a component will fail depends on multiple factors, including its age and the environmental and operational conditions it has been exposed to. If environmental and operational conditions are held constant, the reliability of most types of equipment is described by a timedependent failure rate curve, 9 as shown conceptually in Figure II-2. 9 Ref. Probabilistic Risk Assessment and Management for Engineers and Scientists, 2 nd Ed., Hiromitsu Kumamoto, Ernest J. Henley, p. 267, Figure

9 8 Workpaper Southern California Edison / 2018 GRC Figure II-2 Time-Dependent Failure Rate (A) Probability that an Individual Piece of Equipment will Fail or (B) Fraction of Components in a Large Population Reaching the End of their Service Lives (A) Age of Equipment or (B) Average Age of Population This curve may be read two ways. First, for an individual component, the curve shows that the probability of failure will remain low for a long period given constant operational and environmental conditions. Eventually, the component s materials weaken and its probability of failure increases. Second, for a large population of components, the fraction of components reaching the end of their service lives will be small if the average age of the population is young. 10 Then, as the average age of the population approaches its mean-time-to-failure, the volume of components wearing out and needing replacement will increase significantly. In a fixed population, the volume of equipment failing each year will not increase indefinitely. This is because the average age of a fixed population does not increase indefinitely as aged equipment fails and is replaced by brand new units. At steady state, the average age of a fixed population will 10 The average age must be measured on a fixed (non-growing) population in order for it to be useful in forecasting the number of failures in that population. The addition of new equipment to accommodate new customers reduces the average age of the total population but clearly does not reduce the number of failures expected in the original population. 8

10 Workpaper Southern California Edison / 2018 GRC plateau and, after that, the replacement rate will also plateau at a long-term steady-state replacement rate. Figure II-3 Long-Term Steady-State Replacement Rate 1.2 Fraction of Components Reaching End of Service Life "long-term steady-state replacement rate" Time While precise calculations of when the average age of the population will plateau and when the steady state replacement rate will plateau is complicated, the simple message of these curves should not be lost. As long as the average age of a population continues to increase, the number of components wearing out and needing replacement each year will also increase For further discussion of the use of steady-state replacement rates in infrastructure replacement planning, see the argument of Mark Joseph for CCUE during the oral argument on SCE s 2015 GRC Decision, October 14, 2015, Transcript Vol. 21, pages

11 10 Workpaper Southern California Edison / 2018 GRC C. SCE s Infrastructure Is Aging The average age of many types of equipment in SCE s distribution system is increasing. For example, the average age of primary distribution cable in the SCE system continues to rise, as shown in Table II-3 below. Table II-3 Average Age of Distribution Cable 12 Year End: Paper Insulated Lead Covered High Molecular Weight Polyethylene Cross-Linked Polyethylene Tree Retardant Cross-Linked Polyethylene As SCE infrastructure continues to age, the volume of infrastructure wearing out and needing to be replaced each year will continue to grow. D. Replacement Decision Analytics Equipment age is not the only driver of equipment failures. Many other factors besides age, such as operating history, maintenance history, and environment can drive equipment failures at the individual asset level. As a simple example, younger equipment may have a higher probability of failure when subject to additional operating stresses, such as repeated fault currents from failures of adjacent equipment or from external factors such as storms, vegetation, or Mylar balloons. Likewise, older equipment may have a lower probability of failure if that equipment has been in a fault-free environment for much of its operational life. Therefore, time-dependent asset failure rate curves are a useful starting point to forecast annual infrastructure replacement at the program level. However, additional tools must identify scope at the 12 Cable age data for 2006, 2009, and 2012 is based on previous rate case testimony (2009 GRC, 2012 GRC, and 2015 GRC respectively). 10

12 Workpaper Southern California Edison / 2018 GRC individual asset level. The analytic processes or techniques to select assets for replacement vary depending on asset type. Three illustrative examples are discussed below. As one example, the annual volume of B-bank transformers predicted to reach end-of-life is based on time-dependent failure rate curves. However, each year, selection of B-bank transformers for replacement is based on an algorithm-derived replacement prioritization. This prioritization is based on each individual asset s physical condition and the severity of consequences should that asset fail. This is discussed in greater detail in Section III. B. 1 Transformer Bank Replacement. As another example, distribution cable predicted to reach end-of-life is based on time-dependent failure rate curves, which are subsequently translated into reliability trajectories for the SCE system through circuit modeling and simulation techniques. However, each year, selection of cables starts with identification of SCE s worst performing circuits. These circuits, referred to as WPCs in D , have the greatest frequency of in-service failures and circuit faults. When reviewing WPCs, different analytic processes identify replacement candidates among mainline versus radial cable. Cable testing is incorporated into the analytic process for radial cable, but not for mainline cable due to cost effectiveness. The details of these processes are described in Section III. A. 1 Worst Circuit Rehabilitation Program for mainline cable and Section III. A. 2 Cable Life Extension Program for radial cable. As a third example, the volume of distribution transformers containing PCB oil targeted for replacement is not based on time-dependent failure rate curves, because the program objectives are to systematically remove from service all PCB-contaminated transformers. Each year, selection of transformers for replacement is based on an analysis that compares serial numbers of transformers still in service with serial numbers of transformers removed from service and known to contain PCBcontaminated oil. This serial number comparison has been a simple and effective way of identifying transformers of like vintage, manufacturer, and age of installation, enabling SCE to target PCB transformers effectively as part of this program. This process is described in Section III. A. 8 PCB Transformers Replacement Program. 11

13 12 Workpaper Southern California Edison / 2018 GRC Regardless of the asset, the selected analytic approach to the replacement decision produces infrastructure replacement scope intended to optimize safety, reliability, and affordability. The effectiveness of replacement decisions will improve as additional data becomes available and more sophisticated probabilistic models are developed. SCE s understanding of risk drivers and mitigation alternatives related to overhead conductor has significantly improved in recent years. This greater understanding has helped SCE improve its Overhead Conductor Program OCP scope selection processes, which is expected to increase the effectiveness of OCP in reducing identified risks. For all assets in the IR program, SCE is committed to the continuous improvement of replacement decision analytics to achieve desired safety and reliability goals with maximum customer benefit. III. Capital Work Activities A. Distribution Infrastructure Replacement (DIR) Program The Distribution Infrastructure Replacement program replaces aging or obsolete equipment to minimize the negative effect of in-service equipment failures on system reliability and associated safety risks. Seven individual programs make up the DIR program: Worst Circuit Rehabilitation Program, Cable-in-Conduit Replacement Program, Underground Oil Switch Replacement Program, Capacitor Bank Replacement Program, Distribution Automatic Recloser Replacement Program, Overhead Conductor Program, and PCB-contaminated Transformer Replacement Program An eighth program, the Cable Life Extension Program, is an additional program that does not directly replace infrastructure but provides information to target cable segments to be replaced by the Cable-in-Conduit Replacement Program. 12

14 Workpaper Southern California Edison / 2018 GRC Worst Circuit Rehabilitation Program The Worst Circuit Rehabilitation (WCR) program is an ongoing effort to manage system reliability by dealing with the challenge of infrastructure aging. The objective of the WCR program is to both 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 average customers are receiving relatively lower service reliability. Because cable failure is the largest equipment contributor to poor system reliability, circuit rehabilitation typically involves replacement of each circuit s most risk-significant 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. a) Program Necessity While age-related deterioration will affect every piece of equipment in SCE s distribution infrastructure, underground cable is unique. Significant percentages of the transformer, switch, and pole populations have, for decades, been replaced each year because of routine visual inspections. Underground cable is unique because it cannot be visually inspected without first digging it up. Without a deliberate preemptive replacement program, cable would be removed from the system only because of in-service failure. In 2015 alone, in-service failures of cable or cable-related equipment (elbows, junction bars, or cable splices) were associated with 23% of SCE customer average outage duration 13 and 21% of SCE customer average outage frequency System Average Interruption Duration Index (SAIDI) measures the total duration of interruption for the average customer during a given year. SCE s 2015 SAIDI, with Major Event Days (MEDs) excluded, was minutes of interruption. Outages recorded as cable, elbow/junction bar, or cable splice contributed 22.7 minutes, or approximately 23% of the system total. 14 System Average Interruption Frequency Index (SAIFI) measures the total frequency of sustained interruption for the average customer during a given year. SCE s 2015 SAIFI, with MEDs excluded, was 0.86 (Continued) 13

15 14 Workpaper Southern California Edison / 2018 GRC Table II-3, Average Age of Underground Cable, illustrates the degree at which SCE s underground cable is continuing to age. Figure II-3, Time-Dependent Failure Rate, illustrates that with increasing age will come an increasing volume of cable reaching the end of its service life. This increasing volume of cable reaching the end of its service life will increase the number of inservice failures and resultant circuit outages. These cable outages, which are typically protracted, will negatively affect system reliability and customer satisfaction. SCE performed a probabilistic analysis to forecast the impact of cable aging on system reliability. This probabilistic analysis simulated SCE system reliability if existing infrastructure replacement programs cease beginning in 2015, and cable is treated with a run-to-failure strategy for the next 20 years. The results are shown in Figure III-4 and Figure III-5. Continued from the previous page interruptions. Outages recorded as cable, elbow/junction bar, or cable splice contributed 0.18 interruptions, or approximately 21% of the system total. 14

16 Workpaper Southern California Edison / 2018 GRC 15 Figure III-4 System SAIDI Forecast with No Program of Preemptive Cable Replacement Figure III-5 System SAIFI Forecast with No Program of Preemptive Cable Replacement 15

17 16 Workpaper Southern California Edison / 2018 GRC These figures indicate that absent preemptive replacement, SAIDI would increase by approximately 35 minutes per year (a 34% increase) and system SAIFI by about 0.51 interruptions per year (a 56% increase) over the next 20 years. This would be a significant negative impact to reliable service to our customers. For illustrative purposes, a SAIDI increase of 35 minutes and a SAIFI increase of 0.51 is equivalent to increases of 175 million customer minutes of interruption (CMI) and 2.5 million customer interruptions (CI) respectively. This could represent between 1,250 and 1,450 additional circuit-level interruption events per year. 15 We consider this an unacceptable level of reliability. Our WCR program is designed to prevent this. The WCR program is also a key element of SCE s plan for compliance with new reliability reporting requirements implemented in D This decision updated existing electric reliability reporting requirements for California electric utilities, as part CPUC proceeding R Based on this decision, SCE is required to annually report specific reliability information regarding its list of the top 1% of WPCs for each reporting year in both SAIDI and SAIFI metrics. For any WPC that appears on the list in repeated years, the Commission has directed SCE to identify: An explanation of the metric used to indicate the circuit s performance; A historical record of that circuit s performance in that metric; An explanation of why the circuit is on the WPC list again; An explanation of what is being done to improve the circuit s future performance and the anticipated timeline for completing those activities; and A quantitative description of SCE s expectation for that circuit s future performance The WCR program is the program that executes reliability improvement projects for WPCs within the SCE service territory. 15 This simplified calculation assumes a typical circuit-level interruption event is 2,000 CI and 120,000 CMI. 16

18 Workpaper Southern California Edison / 2018 GRC b) Cost Forecast Our recorded and forecast spending for underground cable replacement under the Worst Circuit Rehabilitation Program is shown in Table III-4 and Figure III-6. To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. Table III-4 16 Historical and Forecast Spend under the WCR/Cable Replacement Programs 17 Year Cable Replacements Completed under the WCR / Cable Replacement Program (conductor-miles) Forecast Unit Cost of WCR / Cable Replacement Program (Nominal $ x 1,000) Recorded/Forecast Spend for WCR / Cable Replacement Program (Nominal $ x 1,000) $77, $66, $121, $153, $117, $345 $137, $352 $123, $361 $126, $372 $130, $384 $134, A separate Cable Replacement Program was started in 2000 in order to focus solely on underground cable replacement, however this program was absorbed into the WCR program in 2012 in order to improve work efficiencies. 17 Refer to WP SCE-02, Vol. 08, pp (Cost of Worst Circuit Rehabilitation). 17

19 18 Workpaper Southern California Edison / 2018 GRC Figure III-6 Worst Circuit Rehabilitation Program 18 Multiple WBSs 19 Recorded /Forecast (CPUC-Jurisdictional Constant 2015 and Nominal $000) c) Justification of Forecast Work (1) Background Table III-5 shows that as of year-end 2014, SCE had approximately 53,701 conductor-miles of primary underground cable in approximately 4,600 distribution circuits. This cable comprises five basic insulation types: (1) lead-covered cable (LCC) or paper insulated leadcovered (PILC) cable; (2) ethylene propylene rubber (EPR) insulated cable; (3) high molecular weight polyethylene (HMW) insulated cable; (4) cross linked polyethylene (XLPE); and (5) tree retardant cross linked polyethylene (TR-XLPE). 18 Refer to WP SCE-02, Vol. 08, pp These capital expenditures includes two WBS elements: CET-PD-IR-WC and CET-PD-IR-CR. 18

20 Workpaper Southern California Edison / 2018 GRC Lead-covered cable such as LCC or PILC cable is the oldest in our distribution system. PILC cable has an average expected time-to-wear-out of about 50 years, and LCC cable is assumed to have a comparable average expected time-to-wear-out. Both types of cable were used in the same general time period prior to 1968, but PILC cable proved to be a more successful design than LCC as shown by the quantities installed. While it is relatively long-lived and resistant to voltage spikes, PILC cable or LCC has many disadvantages. First, it cannot be moved once installed and therefore cannot be used with today s removable elbow connectors for which all modern switches, transformers, and junction bars are designed. Second, when PILC cable or LCC fails, it presents a significant repair challenge. Performing repairs in PILC cable or LCC is time-consuming, can only be done by a few specially trained Cable Splicers, and often results in splices which are prone to subsequent failure. EPR cable has been SCE s standard for PILC cable replacement since Due to PILC s small and compact design, SCE s EPR cable was specially designed to fit into small ducts originally sized for PILC cable. Using existing ducts reduces the cost of installing larger conduit through trenching and civil construction and reduces the circuit outage time when PILC cable is replaced. HMW cable was the industry s first effort at polymer insulation. HMW cable has an average expected time-to-wear-out of about 22 years, which is a significantly shorter life expectancy than cable manufactured today. Most of it has already been removed. We believe that what remains is at the end of its service life. Most of the underground cable in our system is XLPE. This cable has an average expected time-to-wear-out of about 41 years. For XLPE cable, breakdown of the insulation over time causes cable failure. Typically, moisture around the cable penetrates through the polyethylene insulation, causing electrical tracking along voids and contaminants in the insulation and forming patterns that look like trees. This phenomenon of water treeing is a common cause of underground cable failure. 19

21 20 Workpaper Southern California Edison / 2018 GRC TR-XLPE cable has been SCE s standard since TR-XLPE contains XLPE insulation material but with improved resistance to water-treeing. TR-XLPE cable has an expected time to wear out of about 46 years. When a cable fails, as all cable eventually will if not preemptively replaced, electricity breaks through the insulation and results in a ground fault. This condition causes an upstream protective device, such as a fuse, automatic recloser, or substation circuit breaker, to operate and cut off power to all customers downstream of the protective device. Figure III-7 below shows the current inventory of underground cable by year of installation. Figure III-7 Inventory of Underground Cable by Year of Installation average time to wear out, and the current inventory. For each of the cable types, Table III-5 below shows its average age, its 20

22 Workpaper Southern California Edison / 2018 GRC 21 Table III-5 Underground Cable Statistics Estimated dates of Installation Average Age (YE 2014) MTTF (as of 2015 Conductor analysis) miles Type of insulation Description PILC Paper Insulated Lead Covered Prior to ,943 HMW High Molecular Weight Polyethylene ,189 XLPE Cross-Linked Polyethylene ,852 TR-XLPE Tree Retardant Cross-Linked Polyethylene Since ,161 EPR Ethylene Propylene Rubber Unknown 231 LCC Lead-Covered-Cable Unknown 325 Total 53, For the three most dominant aging cable (PILC, HMW and XLPE), SCE has analyzed its cable failures and identified the relationship between the probability of failure and cable age. This relationship is shown in Figure III-8 below. 20 Figure III-8 Cable Failure Rates 8.0E E E-01 Cable failure/mile/yr 5.0E E E-01 XLPE PILC TR-XLPE 2.0E E E Cable Age 20 Refer to WP SCE-02, Vol. 08, pp (Underground Cable Reliability). 21

23 22 Workpaper Southern California Edison / 2018 GRC As one would expect, these cable failure curves resemble the timedependent failure rate curve of Figure II-2 and predict an increase in the number of in-service cable failures as our cable continues to age. And our cable will continue to age. The average age of our XLPE cable (which constitutes 56% of our cable population) is only 29 years, with a mean-time-to-wear-out of 41 years. The average age of our TR-XLPE (which constitutes 38% of our cable population) is only 9 years old, with a mean-time-to-wear-out of 46 years. (2) Program Approach The objective of the Worst Circuit Rehabilitation program is to invest infrastructure replacement dollars where they will achieve the most cost-effective benefit in reliability improvement. Each of our 4,600 circuits is ranked in industry-standard reliability metrics such as SAIDI and SAIFI. The highest ranked (i.e., worst) circuits in each ranking category are selected for further evaluation. This draft list of WCR circuits is then reviewed by SCE stakeholders to help ensure that factors such as local knowledge and needs have been considered in determining which circuits will be addressed that year. With the approved list of circuits, reliability engineers evaluate each circuit s outage history to determine the causes of past outages and what corrective measures could improve future performance. Only the work determined to be the most cost-effective will be performed. When cable is replaced, only the least reliable portion of the circuit, typically less than 10% of the cable in the circuit, is replaced. The installation of fuses on radial portions of the circuit (the single-most costeffective improvement possible), is always performed on any circuit worked whenever physically possible. Isolation devices, such as automatic reclosers, which are relatively inexpensive, are installed wherever determined to be both effective and cost-effective. Replacements of infrastructure where we would expect less than optimum reliability benefits are not performed under the WCR program. 22

24 Workpaper Southern California Edison / 2018 GRC (3) Quantitative Benefits To quantify the impact of cable aging and the resulting increase in cable failures on system reliability, SCE performed an engineering analysis 21 of 20 representative circuits. SCE selected these 20 circuits using cluster analysis methods so they would statistically represent all 4,600 circuits in SCE s distribution system. SCE then developed models of these 20 representative circuits, and assigned probabilities of failure. With these models, we then calculated the theoretical reliability of each circuit. An overall system level reliability was calculated by extrapolating the reliability of these 20 circuits to the entire distribution system. To estimate future system reliability, each cable segment in each of the modeled circuits was aged by increasing its probability of failure with each additional year. Then, SCE calculated each circuit s SAIDI and SAIFI for each year out to 20 years with the results again being extrapolated to reflect the entire distribution system. The results of our analysis are shown in Figure III-9 and Figure III-10 below. 21 Refer to WP SCE-02, Vol. 08, pp (Impact of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability). 23

25 24 Workpaper Southern California Edison / 2018 GRC Figure III-9 Impact of WCR Program on Future SAIDI 24

26 Workpaper Southern California Edison / 2018 GRC 25 Figure III-10 Impact of WCR Program on Future SAIFI The curves labeled No WCR Program on Figure III-9 and Figure III-10 reflect our 20-year forecasts of system SAIDI and SAIFI without the benefit of SCE s WCR program. What our analysis predicts is a SAIDI increase of approximately 35 minutes/year and a SAIFI increase of approximately 0.51 interruptions/year over the next 20 years. After analyzing the impact of infrastructure aging with no preemptive cable replacement, we re-performed the analysis assuming various levels of preemptive cable replacement coupled with the typical WCR circuit enhancements, such as installation of automatic reclosers, fuses, and fault indicators, wherever possible. To forecast the volume of work needed to be performed under the WCR program, our objective was to identify the level of infrastructure replacement which would keep pace with the adverse impacts of aging infrastructure on system reliability. The curves labeled 200 conductor-miles under WCR, 350 conductormiles under WCR and 500 conductor-miles under WCR on Figure III-9 and Figure III-10 reflect our 20-year forecasts of system SAIDI and SAIFI with the benefit of 200 conductor-miles, 350 conductor- 25

27 26 Workpaper Southern California Edison / 2018 GRC miles and 500 conductor-miles (respectively) of annual mainline cable replacement in the WCR program. The analysis concludes that approximately 350 conductor-miles of primary mainline underground cable must be preemptively replaced each year to achieve, in 20 years, approximately today s level of SAIDI. In general, SCE seeks to improve system reliability, rather than simply maintaining existing system reliability. While replacement of 350 conductor-miles of cable per year will generally maintain existing levels of system reliability, SCE fully anticipates that additional reliability benefits will be realized from the Distribution Automation program. SCE plans on leveraging the planning and construction efforts of the WCR Program to maximize the reliability benefits that can be realized in Distribution Automation. Consistent with SCE s distribution automation scheme installed since the late 1990s, the WCR program benefits shown in Figure III-9 and Figure III-10 assume we install automation up to a single mid-point switch and, where applicable, an additional circuit tie switch. In order to maximize the effectiveness of both programs, the WCR program will coordinate with the Distribution Automation program beginning in At that time, the Distribution Automation program will begin to augment each WCR project so that the remediated circuit will have all the reliability benefits of the modern automation scheme. This augmented scope will include up to three mid-point switches, three circuit ties and tie switches, as well as Remote Fault Indicators with modern telemetry. Therefore, the WCR program forecast in Table III-4 includes infrastructure replacement and initial levels of circuit automation. The Distribution Automation program forecast in SCE-02, Volume 10 includes the additional levels of circuit automation needed for the enablement and optimization of DER generation. These additional levels of automation will be used to augment the WCR program beginning in 2018 in order to maximize the reliability benefits of both programs. (4) Qualitative Benefits The WCR program is also designed to address equity in the quality of service provided to our customers. This program is directed specifically at circuits whose reliability 26

28 Workpaper Southern California Edison / 2018 GRC makes them outliers, i.e., worst of the worst. These programs use a variety of reliability metrics to track circuit reliability to ensure that no degradations in performance will go undetected. Finally, SCE monitors circuit performance monthly to help ensure that emergent reliability issues in individual circuits are identified promptly. The reliability performance of each worst-performing circuit in each of SCE s eight Regions is reviewed on a monthly basis. Performance outliers are identified and corrective actions initiated as indicated. The WCR program is SCE s first line of defense in pursuing our goal to provide an equivalent level of reliable and affordable service to all our customers, despite our forecast decline in average overall system reliability. d) Mainline Cable Testing SCE has historically based its underground cable Infrastructure Replacement decisions on risk factors, age, and the likelihood of failure without performing inspection based replacements. With recent advances in cable testing and diagnostic techniques, SCE has implemented cable testing techniques as an input to its Cable-in-Conduit (CIC) replacement program. However, cable testing is not incorporated into the analytic process for mainline cable in the WCR program, as it has been determined that it is not cost effective. Mainline cable testing is discussed in the Cable Life Extension Program section in this volume. 2. Cable Life Extension Program As elucidated below, Weibull analysis of CIC indicates the possibility of a 450% increase in CIC failures through The Cable Life Extension (CLE) Program is one of the mitigations that SCE has in place to avoid these failures and the resulting unplanned outages. The CLE Program performs two types of life-extension activities for radial distribution cable cable testing and cable injection. The first activity in the CLE Program is a partial discharge testing activity ( cable testing ) which provides life extension benefits by identifying those cable segments at greatest risk for imminent failure. Cable segments which test good are those cable segments shown to be partial discharge-free at expected operating voltage levels and therefore expected to have minimal risk of imminent failure. Cable segments which test bad are those that are found to exhibit partial discharge 27

29 28 Workpaper Southern California Edison / 2018 GRC activity that would continue to compromise the cable insulation in normal operation and indicate likelihood of imminent failure. These segments are scheduled for replacement. The second activity in the CLE Program is a cable rejuvenation activity ( cable injection ) that provides life extension benefits by improving the insulation characteristics of aged cable. Cable injection involves the physical injection of a silicone-based fluid along the strands of aging underground primary distribution cable. This fluid migrates into the conductor insulation, modifying its chemistry and improving its dielectric strength. The activity of cable injection does not distinguish between cable presently in good condition versus bad condition, but provides aggregate life extension benefit for all injected cable regardless of present insulation condition. a) Program Necessity Failures on CIC continue to be a significant operational and reliability problem, based on the aging nature of SCE s CIC population. Approximately 25% of SCE s primary distribution cable inventory, based on conductor miles, is comprised of CIC. In recent years, SCE has seen a steady increase in the number of distribution cable-related outage events attributed to CIC failures. In 2015, there were 629 such cable-related outage events in the SCE system; this is a 47% increase since See Table III-6 below. 28

30 Workpaper Southern California Edison / 2018 GRC 29 Table III-6 Cable Related Outages Attributed to CIC Failures 22 Cable-related outages Year attributed to CIC failures % increase ( ) 47% Based on current CIC cable inventory and age distribution, the number of conductor miles of existing CIC expected to fail each year can be estimated based on the Weibull characteristics of XLPE cable, which constitutes the majority of CIC cable type. The results are shown below. 22 SCE s outage database does not directly classify cable outage events by cable construction, i.e. CIC versus rigid duct. The database does, however, provide the protective device that operated for all outage events. For the purpose of this analysis, based on the radial nature of typical CIC installations, SCE attributes cablerelated outages to CIC failures when such events include operation of a radial protective device such as a fused underground switch. 29

31 30 Workpaper Southern California Edison / 2018 GRC Figure III-11 Forecast Failures of Existing CIC Inventory ( ) Based on these results, SCE expects CIC failures to increase from approximately 60 miles of CIC failure per year today to approximately 330 miles of CIC failure per year in the next 20 years without mitigation. This indicates that a robust strategy for dealing with aging CIC is required now to address this growing problem. SCE initially implemented the CLE Program in the 2015 GRC as a cable testing activity. This program was designed to more accurately select the segments of cable needed to be replaced and minimize the replacement of cable with useful remaining life. Since the last rate case cycle, SCE has learned more about cable injection techniques in the utility industry. SCE has a solid understanding of cable testing, but relatively limited experience in cable injection. SCE recognizes there are inherent advantages and disadvantages to each of these two types of cable life extension activities. Based on the magnitude of SCE s CIC problem, a one size fits all approach to CIC life extension is not appropriate. SCE is planning a combination of cable testing activities and cable injection activities in its CLE program in this rate case cycle. The cable testing activity will directly identify scope 30

32 Workpaper Southern California Edison / 2018 GRC for the CIC Replacement Program by identifying cable segments that test bad and are at elevated risk of imminent failure. The cable injection activity will directly rejuvenate segments of aged CIC cable without replacement. If cable condition is poor enough that injection is not viable, these cables will be replaced under the CIC Replacement Program. SCE has estimated that pre-emptive replacement of approximately 150 miles of CIC per year based on testing bad, plus injection of 100 miles of CIC per year based on age criteria alone, will be necessary for SCE to adequately address the rapidly growing CIC problem illustrated in Figure III-11 above. b) Cost Forecast Table III-7 and Figure III-12 below indicate SCE s historical and forecast spending for its Testing-Based Cable Life Extension program. 23 To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. Table III-7 24 Historical and Forecast Spending for CIC Testing and Injection Year Recorded/Forecast of Cable Testing (conductor-miles) Forecast Unit Cost of Cable Testing (Nominal dollars x 1,000) Recorded/Forecast Cost Cable Testing (Nominal dollars x 1,000) Recorded/Forecast of Cable Injection (conductor-miles) Forecast Unit Cost of Cable Injection (Nominal dollars x 1,000) Recorded/Forecast Cost Cable Injection (Nominal dollars x 1,000) Total Recorded/Forecast Cost Cable Testing and Injection (Nominal dollars x 1,000) $0 $ $1,361 $1, $7,863 $80 $7, $12,768 $474 $13, $6,852 $4,813 $11, $47 $9, $136 $12,196 $21, $48 $9, $138 $13,833 $23, $49 $9, $142 $14,181 $23, $51 $10, $146 $14,626 $24, $52 $10, $151 $15,091 $25, On September 4, 2008, FERC issued a letter, (Docket No. AC ), which stated that costs associated with a cable injection program meeting certain requirements could be capitalized. In addition, on March 9, 2009, FERC issued a letter, (Docket No. AC ), which stated that costs associated with a cable testing program meeting certain requirements could be capitalized. 24 Refer to WP SCE-02, Vol. 08, pp (Cost of Cable Testing) and WP SCE-02, Vol. 08, pp (Cost of Cable Injection). 31

33 32 Workpaper Southern California Edison / 2018 GRC Figure III-12 Cable Life Extension Program 25 WBS Element CET-PD-IR-LE Recorded /Forecast (CPUC-Jurisdictional Constant 2015 and Nominal $000) c) Justification of Forecast Work As shown in Figure III-11, SCE estimates that without mitigation, approximately 330 miles of CIC cable failures per year will be realized in the next 20 years due to CIC aging. SCE has estimated that pre-emptive replacement of approximately 150 miles of CIC per year based on testing bad, plus injection of 100 miles of CIC per year based on age criteria alone, will be necessary for SCE to adequately address the rapidly growing CIC problem illustrated in Figure III-11. SCE will replace 150 miles of CIC per year through testing and injection. SCE estimates that testing of approximately 200 miles of cable per year (a sustainable level of testing based 25 Refer to WP SCE-02, Vol. 08, pp

34 Workpaper Southern California Edison / 2018 GRC on SCE experience) will yield around 80 miles of CIC for replacement based on an approximate 40% fail rate assumption, which is based on our historical experience in this area. Preliminary data also suggests that 40% of cable that we attempt to inject will not be injectable because of poor physical condition such as concentric neutral conductor degradation or excessive number of cable splices. Such cables will most likely fail cable testing and therefore preemptive replacement is recommended under such conditions. SCE estimates that attempting to inject 170 miles of cable per year will cause approximately 100 miles of cable successfully injected, and approximately 70 miles of cable found needing replacement, each year. Since we identify approximately 80 miles of cable replacement based on testing results and 70 miles of cable replacement from non-injectable cable, and can accomplish 100 miles of cable injection, we estimate mitigating approximately 250 miles of CIC per year 150 miles from cable replacement and 100 miles from cable rejuvenation. This will not only allow SCE to directly address the growing CIC problem in Figure III-11, but will also give SCE valuable experience related to implementation details of cable injection as another cable life extension technique. d) Lessons Learned from CIC Cable Testing In the 2012 GRC decision, the Commission directed SCE to carefully document the data collection from the Cable-in-Conduit (CIC) pilot program, and other efforts it undertakes to develop a best practice and most cost-effective method for replacements. This information shall be submitted to support future GRC requests in this category to assist the Commission and to illustrate that ratepayers achieved value from SCE s lessons learned. (1) CIC Testing Effectiveness SCE has learned a great deal of information about the effectiveness of distinguishing CIC cable in good condition from CIC cable in poor condition. As of May 2016, SCE has performed CIC testing on 98 distribution circuits, spread among 22 districts and 8 regions in SCE s service territory. Among these circuits, the program has tested approximately 862 conductor miles of CIC in 8,961 segments. A summary of the test results, broken down by region, is provided Table III-8. 33

35 34 Workpaper Southern California Edison / 2018 GRC Table III-8 Summary of SCE s CIC Test Results (as of May 2016) NUMBER OF CIRCUITS TESTED CONDUCTOR MILES TESTED REPLACEMENT/REPAIR NEEDED NO REPLACEMENT/REPAIR NEEDED Region Desert % 65% Metro East % 46% Metro West % 46% North Coast % 48% Orange % 42% Rurals % 61% San Jacinto % 62% San Joaquin % 68% OVERALL % 53% These results show that CIC testing is effective at distinguishing cable in good condition from cable in poor condition. As of May 2016, on average, just over half of the CIC cable tested by SCE (53%) has been found to be in good condition and requiring no near-term follow-up action. The results also show high uniformity of test results among various regions in the SCE service territory, with pass rates in the SCE service territory varying between 42% and 68%. This shows that CIC problems are not localized to any portion of the service territory; high volumes of poor condition CIC can be found throughout all regions of the SCE system. (2) CIC Testing Efficiency SCE has also learned a great deal about the efficiency of testing to identify CIC cable in poor condition. As of May 2016, SCE has tested 8,961 segments of cable using 1,371 testing crew days to complete the testing. Based on these results, SCE has learned that testing crews can generally test CIC cable at a pace of approximately 6.5 CIC segments per testing crew per day. This efficiency is largely the result of the physical configuration of CIC on distribution circuits. CIC is typically installed on radial portions of distribution circuits, which are spurs off of circuit mainlines. Each radial serves a few customers relative to the total circuit customer base. CIC is typically connected to mainline through a switch, junction bar, or fuse. This provides a convenient switching point for de-energizing the radial for CIC testing without subjecting larger numbers of customers to the planned outage. CIC cable can be isolated from the rest of the circuit with a 34

36 Workpaper Southern California Edison / 2018 GRC single switching operation, tested efficiently along the full length of the de-energized radial, and then restored to service again with a single switching operation at the end of the testing day. This greatly improves the pace of CIC testing and has a positive impact on cost of testing on a cost-per-conductormile basis. (3) CIC Testing Challenges Finally, SCE has learned some useful lessons about various challenges associated with CIC testing, unforeseen at the onset of the CIC testing program. SCE has learned that impacts of cable faults during testing cannot be ignored when planning a cable testing program. These faults require immediate replacement or repair before service is restored. SCE has learned that a small, but non-negligible, percentage of aged CIC cable can fault during partial discharge testing itself. As of April 2016, among the 8,961 segments of CIC tested, SCE has experienced 74 segments that have faulted during test. See the table below. Table III-9 Summary of CIC Faults During Test NUMBER OF SEGMENTS TESTED NUMBER OF SEGMENTS FAULTING DURING TEST PERCENT Voltage Class 12 kv 7, % 16 kv 1, % OVERALL 8, % Table III-9 shows that approximately 1% of the time, cable being tested for partial discharge activity faults during the test. The percentage is slightly higher for cable operating at 16 kv and slightly lower for cable operating at 12 kv. These faults take place during scheduled outages, so customers do not immediately see the impact of these faults. However, these faulted cables must be replaced on an emergency breakdown basis before customers can be restored at the end of the scheduled outage. Sometimes, replacement of faulted cable may also need to take place in parallel with completing 35

37 36 Workpaper Southern California Edison / 2018 GRC remaining testing activity in the same area. This can be an operational burden on resources and can cause extended outages beyond the planned restoration time for affected customers. Fortunately, since CIC testing takes place on circuit radials, these reliability risks and customer impacts are localized only to the radial under test. Even in these small number of cases, because CIC testing takes place on circuit radials, the resulting operational and reliability problems rarely impact safe and reliable operation of the entire distribution circuit. (4) CIC Replacement Techniques and Associated Costs Finally, SCE s recently implemented changes in CIC replacement techniques that have had a positive impact on CIC replacement unit costs. Historically, SCE replaced CIC with a relatively costly method of open cut trenching. In late 2013, SCE implemented new CIC replacement techniques. SCE replaced CIC with a combination of new lubricating equipment, new cable pulling equipment, and new smaller-diameter replacement cables. Since that time, SCE s construction standards specify that removal of CIC from its tubing using these techniques is the preferred option for replacing CIC. If removal from the tubing fails, then by default the replacement method will be the traditional approach of open cut trenching. These new replacement techniques have reduced CIC replacement unit cost. The present unit cost for CIC replacement in this rate case is approximately 50% lower than the unit cost identified in the 2015 GRC and is approximately 70% lower than the cost request in the 2012 GRC. This illustrates the effectiveness of the new CIC replacement techniques implemented beginning in e) Mainline Cable Testing Discussion As directed by the Commission in D , SCE has re-evaluated the potential use of testing as part of its Worst Circuit Rehabilitation (WCR) program to reduce the likelihood of replacing cable unnecessarily. Specifically, the Commission required SCE to provide a detailed cost benefit analysis to justify cable that SCE replaces without testing, particularly in the WCR Program. The Commission recognized that implementing such a change takes time, but stated the 36

38 Workpaper Southern California Edison / 2018 GRC benefits to customers of reducing good cable replacement can outweigh the benefit to customers of accelerated replacement of more total cable. 26 (1) Differences between Mainline Cable and CIC Because SCE has historically used cable testing on CIC but not for mainline cable, any discussion of mainline cable testing must begin with a relative comparison of the differences between mainline and CIC. Figure III-13 Mainline and Radial Cable Illustrated on a Typical Underground SCE Circuit CIC is typically installed on radial portions of distribution circuits, and is typically connected to mainline cable through fused underground switches, dedicated positions on mainline switches, junction bars, or similar techniques. These provide convenient switching points for de-energizing a single radial for CIC testing without affecting other downstream radials on the circuit. However, mainline switches, installed at less frequent intervals along the circuit, divide circuits into load blocks. These load blocks can be de-energized by use of mainline switches for planned work on mainline cable, but this is done at a price the opening of a mainline switch to isolate mainline cable will de-energize all downstream load blocks, not just the load block targeted for planned work. 26 See D pages

39 38 Workpaper Southern California Edison / 2018 GRC CIC has historically been installed within a non-rigid polypropylene tube in the place of rigid duct. This allowed for inexpensive initial installation costs, but has led to higher end-of-life replacement costs. SCE has successfully reduced CIC replacement costs through new replacement techniques, but the costs for CIC replacement are still higher than the comparable costs for mainline cable replacement. Mainline cable is almost always found to be installed in rigid duct which allows for relatively quick cable replacement. (2) SCE Experience Testing Mainline Cable When SCE first evaluated partial discharge testing as part of its distribution cable asset management strategy, SCE s original intent was to perform cable testing on mainline cable. SCE performed tests of distribution mainline cable in May A total of nine segments of mainline cable, on three distribution circuits, were identified for a mainline cable testing pilot. The circuits tested were selected based on the ability to isolate the cable through switching to minimize impacts of planned outages to large areas of customers. The pilot project demonstrated that testing can be effective at distinguishing between cable in good condition and cable needing repair. Among the nine segments tested, five segments (63% on a conductor mile basis) were partial-discharge free and four segments (37% on a conductor mile basis) were found to have evidence of partial discharge activity. See Table III- 10 below. Table III Mainline Cable Testing Pilot Results TESTED REPLACEMENT/ REPAIR NEEDED NO REPLACEMENT/ REPAIR NEEDED Cable Segments Percent of total - 44% 56% Cable Conductor Miles Percent of total - 37% 63% 38

40 Workpaper Southern California Edison / 2018 GRC The pilot project also demonstrated the relative inefficiency of mainline cable testing. Due to the switching complexity to avoid large planned outages and circuit abnormal conditions, the pace of mainline testing was relatively slow. Specifically, the nine segments were tested over the course of six testing crew days. This means that SCE could only test mainline cable at a rate of 1.5 mainline segments per test crew per day. Mainline testing was halted while the testing strategy was re-evaluated. Ultimately, this experience in 2008 motivated SCE to shift its focus for cable testing from mainline to radial efforts, in particular to the growing problem of CIC. In SCE s initial evaluations of radial cable testing in 2012, SCE could test over 6 segments of radial cable per day with a much greater efficiency. (3) Industry Experience Testing Mainline Cable The 2008 pilot identified a 63% pass rate for mainline cable on a conductor-mile basis, but since that sample size was so small, these percentages are not statistically significant. Because of the small sample size, industry data can be used to improve the estimates of mainline testing effectiveness. SCE s cable testing vendor has experience testing both radial cable and mainline cable throughout the industry. According to their testing data, on average, approximately 40% of aged mainline cable systems meet standards, an additional 20% require repairs or replacement, and the remaining 40% require replacement. In contrast, on average, approximately 70% of aged radial cable systems meet standards, an additional 10% require repairs or replacement, and the remaining 20% require replacement. Therefore, industry data suggests that radial cable testing on average has a 70% to 80% yield rate in avoided cable replacements, and mainline cable testing on average has a lower 40% to 60% yield rate. Based on these industry averages, we used a range of 40% to 60% yield rate for the mainline testing economic analysis below. (4) Economic Analysis of Mainline Testing As discussed in Section I.D above, SCE s 2015 GRC decision directed SCE to explain why a testing-based replacement program is not more cost-effective for mainline cable 39

41 40 Workpaper Southern California Edison / 2018 GRC replacement under the WCR program. In response, SCE performed a financial analysis to determine the cost effectiveness of a mainline testing program. The financial analysis focused on the life-cycle cost of one year s amount of WCR cable replacements of 350 miles under two options for mainline cable testing. The first option (Option 1) determined the costs to replace all 350 miles in the first year with no mainline cable testing. The second option (Option 2) assumed that all 350 miles would be tested in the first year, cable that failed the testing would be replaced in year one, and the cable that passed would be replaced in a future year. In Option 2, two yield rate scenarios were evaluated, with one scenario (Scenario A) assuming a 60% yield rate of avoided replacements and the other scenario (Scenario B) assuming a 40% yield rate of avoided replacements. These yield rates are based on industry data. Cables that test good in the first year will still fail in the future and need to be replaced, but SCE does not yet have statistical information about future failure rates of tested cable. Therefore, for this analysis, we assumed a mean-time-to-future-failure concept for tested cable. Analysis was performed to determine the breakeven point of the mean-time-to-future-failure of tested cable compared to replacing all cable in year one. To account for timing differences, SCE compared the two scenarios using a Present Value of Revenue Requirement (PVRR) analysis. The economic analysis was based on three key assumptions: cable testing yield rates (i.e., effectiveness); cable testing costs (i.e., efficiency); and cable replacement costs. The assumptions used in this analysis was: Cable testing yield rates were based on industry average data for mainline cable testing. The optimistic Scenario A assumed a 60% yield rate of avoided replacements (i.e., 40% failure rate), and the conservative Scenario B assumed a 40% yield rate of avoided replacements (i.e., 60% failure rate). Cable testing costs (i.e., efficiency) for mainline testing were based on radial testing costs, adjusted to account for the slower pace of 40

42 Workpaper Southern California Edison / 2018 GRC mainline testing based on switching requirements and operational complexities. Cable replacement costs were based on recorded costs for cable replacement only work orders that closed in 2015 as part of the WCR Program. The two scenarios used the same replacement costs, adjusted for escalation for scenario two. 7 Figure III-14 summarizes the key assumptions and calculations. Figure III-14 Summary Mainline Testing Assumptions and Calculations (Constant 2015$) 27 Assumptions Annual Miles (A) 350 Scenario 1 Failure Rate (B) 40% Scenario 2 Failure Rate (C) 60% Cost per Mile to Replace (D) $ 245,526 Cost per Mile to Test (E) $ 98,000 Option 1 Replace All without Testing Annual Miles 350 A Total Cost 85,934,150 A x D Option 2 Test & Replace Fail Now and Pass in Future Scenario A: 40% Failure Scenario B: 60% Failure Miles Tested Bad (F) 140 A x B =F 210 A x C =F Cost to Test $ 13,720,000 E x F $ 20,580,000 E x F Cost to Replace 34,373,660 D x F 51,560,490 D x F Total Cost of Bad $ 48,093,660 $ 72,140,490 Miles Tested Good (G) 210 A - F = G 140 A - F = G Cost To Test $ 20,580,000 E x G $ 13,720,000 E x G Total First Year Cost $ 68,673,660 $ 85,860,490 Future Replace Cost $ 51,560,490 D x G $ 34,373,660 D x G Total Cost $ 120,234,150 $ 120,234,150 First Year Cost Comparison Option 1 Total cost $ 85,934,150 $ 85,934,150 Option 2 First Year Cost 68,673,660 85,860,490 Variance $ (17,260,490) $ (73,660) 27 Refer to WP SCE-02, Vol. 08, pp (Mainline Cable Testing Analysis). 41

43 42 Workpaper Southern California Edison / 2018 GRC Figure III-15 summarize the revenue requirement for Option 1 and Option 2 Scenarios A and B describe above. The graph shows the revenue requirement analysis results for when the Option 2 scenarios breakeven with Option 1. On a revenue requirement basis Option 1 Scenario A requires cable that passes the test to last an average of approximately another 14 years. The analysis also shows that Option 2 Scenario B is not cost effective at any time in the analysis timeframe. Figure III-15 Summary of Mainline Testing Cost/Benefit Analysis From this analysis, SCE has concluded that mainline cable testing as part of the WCR Program would not be cost-justified. Figure III-14 shows that under the more optimistic Option 2 Scenario A, mainline cable that tests good must survive another 15 years beyond the test data to break even with Option 1. Under the more conservative Option 2 Scenario B, even if cable that tested good survived another 29 years beyond the test date, it would still be less expensive on a present value revenue requirement basis to not test the cable. Based on a sample of approximately 80 conductor miles of WCR-targeted mainline cable from the 2016 scoping year, the average age of cable replaced under WCR is 42

44 Workpaper Southern California Edison / 2018 GRC approximately 39 years. 28 From the break-even results above, the average cable that passes testing under Option 2 must last to an average age of somewhere between 54 years and 68 years to break even with Option 1. Based on these findings, SCE has concluded that it is not reasonable to expect that mainline cable will last long enough after testing to make Option 2 preferable over Option 1. This analysis supports SCE s current practice of not testing mainline cable before replacement. (5) Qualitative Analysis of Mainline Testing vs. CIC testing The economic analysis above does not capture the many qualitative problems associated with mainline cable testing, such as more significant requirements for customer impacts due to scheduled outages. Cable testing typically involves scheduling customer outages for testing and subsequently scheduling customer outages for replacement. This is relatively manageable when the customer counts are small, such as on circuit radials. However, this becomes much more problematic from a customer service point of view when customer counts are larger, such as for larger mainline load blocks. There are additional operational concerns about mainline testing related to the relatively small, but still existing, risk of cables faulting during test. SCE learned from CIC testing that in the 1% of cases in which cables fault under test, circuit abnormal conditions can extended well beyond the scheduled outage end time. In the case of CIC testing, the associated challenges of extended circuit abnormal conditions only impacts the radial spur off of the circuit mainline. However, if the same were to take place on a circuit mainline, the challenges associated with extended circuit abnormal conditions would affect the entire circuit downstream of the point of fault. In other words, there would be additional switching requirements, greater customer impacts, and greater reliability risks for other unplanned outage events. 28 Refer to WP SCE-02, Vol. 08, pp (Average age of WCR). 43

45 44 Workpaper Southern California Edison / 2018 GRC (6) Mainline Testing Conclusion In summary, SCE has concluded that testing of mainline cable is not costjustified as part of the cable replacement decision process in the WCR program. SCE will continue to implement cable testing as part of its cable replacement decision processes where it makes the most economic and operational sense to do so on CIC radial cables. 3. CIC Replacement Program The CIC Replacement Program preemptively replaces segments of SCE s cable-inconduit (CIC) population approaching the end of their service life. The objective of the program is to reduce the number of in-service failures of CIC cable and thus drive down the number of unplanned outages to SCE customers. a) Program Necessity In the late 1960s, SCE installed an underground cable known as cable-inconduit, or CIC. CIC is distinguished not by its insulation material (its insulation is HMW or XLPE), but by its construction. While cable in the mainline sections of circuits is typically installed in rigid PVC duct, CIC is installed in relatively thin-walled polypropylene tubing. CIC comes from the manufacturer with the conductor already inside the polypropylene tubing and coiled up on a large reel. It was installed primarily in radial branches of circuits serving residential customers. CIC was very attractive at the time due to its: (1) ease of installation which shortened the construction time of residential developments; (2) lower cost relative to cable installed in rigid duct; and (3) greater durability over that of direct buried cable. However, decades later we have found that CIC is very difficult to replace. While cable installed in rigid PVC duct can be removed relatively easily, CIC cable resists being pulled out from its polypropylene tubing. This is particularly true when the polypropylene tubing has been damaged, as is usually the case when a CIC cable faults to ground. The tight clearances between the conductor and the tubing wall, the tendency of the tubing to crush and impinge on the conductor, and the tendency of the concentric neutral wires to break and ball up all make removal of the CIC conductor difficult. As a result, outages to replace failed CIC can be long in duration. 44

46 Workpaper Southern California Edison / 2018 GRC Figure III-12 depicts the forecasted number of failures of CIC in existing inventory without a program of preemptive cable replacement. Approximately 13,000 conductor-miles, one-fourth of SCE s cable population, is CIC type cable. The challenge of an aging cable population cannot be adequately met without addressing CIC. b) Cost Forecast Table III-11 and Figure III-16 below indicate SCE s historical and forecast spending under the CIC Replacement program. To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. Table III-11 Historical and Forecast CIC Cable Replacement 29 Year Recorded/Forecast Replacements of CIC (conductor-miles) Forecast Unit Cost of CIC Replacement (Nominal dollars x 1,000) Recorded/Forecast Cost of CIC Replacements (Nominal dollars x 1,000) $5, $4, $13, $23, $54, $265 $26, $271 $31, $278 $41, $286 $42, $295 $44, Refer to WP SCE-02, Vol. 08, pp (CIC unit cost). 45

47 46 Workpaper Southern California Edison / 2018 GRC Figure III-16 Cable-in-Conduit (CIC) Replacement Program 30 WBS Element CET-PD-IR-CC Recorded /Forecast (CPUC-Jurisdictional Constant 2015 and Nominal $000) c) Justification of Forecast Work As discussed in preceding sections, SCE s cable population is continuing to age. Without a program of preemptive cable replacement, that aging will cause a significant decline in reliability. As shown in Figure III-11, SCE forecasts over 330 miles of CIC cable failures per year in the next 20 years unless we take significant mitigation measures. We estimate that pre-emptive replacement of approximately 150 miles of CIC per year based on testing bad is necessary to adequately address the rapidly growing CIC problem. SCE forecasts that approximately 80 miles of cable will be identified per year based on cable testing, and 70 miles of cable will be identified per year from non-injectable cable. 30 Refer to WP SCE-02, Vol. 08, pp

48 Workpaper Southern California Edison / 2018 GRC d) Conclusion SCE is committed to replacing CIC entirely based on outcomes from the Cable Life Extension Program. Preemptive replacement of 150 miles of CIC per year is necessary to prevent the decline in reliability associated with CIC failures. 4. Overhead Conductor Program The Overhead Conductor Program (OCP) is a new program to address public safety risks. The scope and mitigations used for OCP have been influenced by SCE s risk informed decision making process. The goals of the Overhead Conductor Program are to reduce the frequency and impact of wire down events by executing proactive overhead conductor replacement projects, reactive emergency wire down work during events, and reactive planned conductor work after wire down events. Similar to the WCR program that focuses on the worst performing circuits to addresses reliability risks, OCP ranks overhead circuits based on criteria such as specific increased likelihood of wire down events to address safety and reliability risks. In 2013, SCE started data collection efforts and small wire research. In 2014, we started safety and reliability risk analysis, which demonstrated the safety risk of electrocution caused by energized wire down events is considerable relative to other system risks. 31 OCP started scoping and executing work in 2015 to address these safety risks. SCE has refined its scoping criteria for proactive OCP projects and has identified work for 2016 and As overhead conductor risk analysis continues to be refined, the scoping criteria and mix of mitigations targeted for the OCP will continue to evolve. Please see SCE-02 Vol. 1, for more information on operational OCP decisions influenced by risk analysis. a) Program Necessity Wire down events have numerous triggering causes such as Mylar balloons contacting conductors. Wire down events are potentially hazardous, and this drove SCE s decision to 31 See SCE-02, Vol.1 Appendix. 32 Refer to WP SCE-02, Vol. 08, pp (OCP Scope). 47

49 48 Workpaper Southern California Edison / 2018 GRC implement the OCP program. 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. Central to OCP strategy is an understanding of short circuit duty (SCD). Generally, SCD indicates the relative strength of a system, typically measured by the fault current (in amps) that the system can supply at any location within the system. For older overhead wire installations, existing levels of SCD can result in increased risk of conductor damage during fault conditions, though it is not currently possible to determine the extent of conductor damage on in-service overhead conductor from previous faults. 33 Possible solutions to such problems fall into one of three categories: (1) changes to protection device settings to accomplish faster fault clearing and reduce conductor damage; (2) installation of new protective devices, such as additional fuses or automatic reclosers; and (3) to reconductor smaller-gauge wire to a larger size that reduces the risk of conductor damage during fault conditions. Additional mitigations are being explored as part of the T&D risk informed decision making process. 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. In many cases, existing overhead conductor on SCE s system was sized in compliance with design standards at the time of construction. For example, while distribution design standards of today specify a minimum wire size for new overhead construction in the SCE system, many older overhead installations 33 In other words, for smaller wire sizes in elevated SCD conditions, the amount of fault current may be in excess of the withstand capabilities of the conductor, and under fault conditions, the conductor may be damaged. 48

50 Workpaper Southern California Edison / 2018 GRC have wire sizes smaller than the current standard for new construction still in-service. Generally speaking, smaller wire sizes are less able to withstand high levels of fault current without damage than relatively larger wire sizes. Over the years, these smaller wire sizes can also be subject to many fault current events of high fault current magnitude. Until OCP, SCE s practice for failed overhead conductor was to repair by installing a splice to accomplish the quickest load restoration possible. This can lead to situations where older overhead conductor has numerous splices, reflecting the poor aggregate reliability history of that span in relation to its fault history. b) Cost Forecast Table III-12 and Figure III-17 below indicate SCE s historical and forecast spending under the OCP Replacement program. To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. Year Table III-12 Historical and Forecast Spend for OCP 34 Recorded/Forecast Replacements Overhead Conductor (Circuit-miles) Forecast Unit Cost of Overhead Conductor Replacement (Nominal dollars x 1,000) Recorded/Forecast Cost of Overhead Conductor Replacements (Nominal dollars x 1,000) $58, $444 $142, $454 $136, $465 $139, $480 $143, $495 $148, Refer to WP SCE-02, Vol. 08, pp (OCP unit costs), 49

51 50 Workpaper Southern California Edison / 2018 GRC Figure III-17 Overhead Conductor Program 35 WBS Element CET-PD-IR-OC Recorded /Forecast (CPUC-Jurisdictional Constant 2015 and Nominal $000) c) Justification of Forecast Work OCP aligns with risk-informed decision making supported by the Commission to comprehensively examine risks and mitigations to improve risk reduction activities. As discussed in SCE-02, Volume 1, OCP has been greatly influenced by what SCE learned during the Safety Model Assessment Proceeding (SMAP). Costs associated with OCP work come from proactive projects and reactive projects. 36 Proactive projects include work that is scoped before any failures in service. Reactive projects are scoped and executed in reaction to a wire down event and are considered to be emergency wire down work during events or planned conductor work shortly after wire down events (see section 2 below). 35 Refer to WP SCE-02, Vol. 08, pp. 1-17, 36 SCE does not track costs by proactive vs. reactive OCP Project. 50

52 Workpaper Southern California Edison / 2018 GRC (1) Proactive Projects These proactive projects may be thought of as an overhead equivalent to WCR projects; while WCR projects address the risks associated with significant failures of underground cable, OCP projects address risks associated with overhead conductor failure. The proactive portion of the program performs an annual review of all circuits within the SCE system for their relative prioritization from an OCP perspective. This prioritization is based on circuit breaker operations, customer density, fault duty (SCD) data, and recent history of wire-down events. Every year, the circuits identified as representing the greatest risk for a wire-down event are reviewed to determine the mitigation. The proactive projects typically include replacement of overhead conductor and the installation of additional protective devices such as branch-line fuses. (2) Reactive Projects During emergency load restoration due to a wire down event, work may be accomplished for greatest conductor benefit with minimal customer outage impact. At the time of a wire-down incident, responders may identify and request the immediate installation of branch-line fusing to protect the conductor or the upsizing of existing non-standard conductor to standard conductor, with the objective of not aversely delaying restoration of interrupted customers. If performing the reactive work would materially impact the customer outage, responders will make only the repairs to restore service without the additional work. In these cases, an after-the-fact evaluation is performed of the wire size and protective devices near the wire down event. Based on this evaluation, existing protective device settings may be modified, or a project may be created. The scope of this project will be to: reconductor existing small-gauge wire to larger conductor; or, install or modify protective devices such as fuses and automatic reclosers. However, the mitigation would be made in two phases; the first phase would be immediate restoration efforts to get customers back online, and the second phase would be a follow-up to rebuild or otherwise upgrade those circuit sections that failed. 51

53 52 Workpaper Southern California Edison / 2018 GRC Underground Oil Switch Replacement Program The Underground Oil Switch Replacement Program replaces oil-filled switches in underground structures which we believe are approaching the end of their service lives and pose a threat to both system reliability and public and employee safety. a) Program Necessity Switches are used in the distribution system for opening and closing electrical circuit connections. Switches are found in both overhead and underground circuits, with underground circuits containing both subsurface and padmounted switches. Subsurface switches are inspected every three years in compliance with GO 165. These inspections include visual examination of the enclosure for corrosion, leaks, and hot connections. Every six years, every oil-filled switch is subjected to an oil test to check for water ingress. Unfortunately, inspections cannot detect all imminent failures of switches. Deterioration of the electrical contacts and other components internal to the switch cannot be detected. In-service failures of switches can occur. In years , an average of 44 underground oil-filled switches per year have failed. See Table III-13. Table III-13 Annual Underground Oil-Filled Switch Failures ( ) SWITCH TYPE OIL BURD SWITCH OIL RAC SWITCH OIL RAM SWITCH Total The primary reason for SCE s program to remove old oil-filled switches is that failures of oil-filled switches can be violent. Arcing across electrical components immersed in oil (as can occur inside of oil-filled switches approaching failure) creates dissolved acetylene gas in the oil, which is highly explosive. Violent failures of oil-filled equipment can damage adjacent electrical equipment (e.g., cable, transformers, switches), expanding the scope and duration of the outage. Property damage and injuries can also result from violent oil switch failures. Most of our oil-filled subsurface switches are 52

54 Workpaper Southern California Edison / 2018 GRC in concrete structures, typically 10 feet by 15 feet, underneath streets. Violent failures of oil-filled equipment can release enough energy to send the concrete lid of the structure several feet into the air. This has the potential of causing great bodily harm and damaging property. b) Cost Forecast Table III-14 and Figure III-18 below indicate the number of subsurface oil-filled switches replaced and the recorded spend from 2011 through 2015 and SCE s forecast replacements and spend from 2016 through To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. Table III Historical and Forecast Spending for Underground Oil Switch Replacement Year Number of Underground Switch Replacements 1 Forecast Unit Cost of Underground Switch Replacement (Nominal dollars x 1,000) Recorded/Forecast Cost of Underground Switch Replacements 2 (Nominal dollars x 1,000) $15, $9, $18, $19, $25, $61 $10, $62 $11, $64 $12, $65 $13, $68 $13, Refer to WP SCE-02, Vol. 08, pp (Underground Switch unit costs). 53

55 54 Workpaper Southern California Edison / 2018 GRC Figure III-18 Underground Oil Switch Replacement Program 38 WBS Element CET-PD-IR-SR Recorded /Forecast (CPUC-Jurisdictional Constant 2015 and Nominal $000) c) Justification of Forecast Work There are about 45,626 subsurface and padmounted switches installed in our underground system. The age distribution of these subsurface and padmounted switches is shown in Figure III-19 below. 38 Refer to WP SCE-02, Vol. 08, pp

56 Workpaper Southern California Edison / 2018 GRC 55 Figure III-19 Summary of UG and Padmount Switch Inventory by Installation Year Number of UG Switches Year of Installation SCE performed an engineering analysis of mainline oil-filled switch reliability, which concludes that the relationship between unreliability and age is represented by the curve in Figure III-20 below Refer to WP SCE-02, Vol. 08, pp (Underground Oil Switch Reliability). 55

57 56 Workpaper Southern California Edison / 2018 GRC Figure III-20 Underground Oil Switch Failure Rates This reliability analysis concludes that the mean-time to wear-out of mainline oil switches is about 35 years. As of the end of 2015, SCE had 1,883 mainline oil-filled switches, with 181 older than 35 years. There is a relatively low number of mainline subsurface oil switches older than 35 years results due to recent infrastructure replacement efforts that focused on addressing these before radial switches. The mean-time to wear-out of radial switches is assumed to be comparable to that of mainline switches. As of the end of 2015, SCE had 9,666 radial oil-filled subsurface switches still in service, with 1,429 of them older than 35 years. The relatively large number of radial switches older than 35 years indicates that previous IR program efforts did not target radial subsurface switches for preemptive replacement as the program was focused on mainline switches due to their larger potential 56

58 Workpaper Southern California Edison / 2018 GRC reliability impact upon failure. Now that population mainline switches older than mean-time to failure is relatively low, SCE is now shifting focus to radial switches. Based on a combined population of 11,549 subsurface oil-filled switches with a mean-time to wear-out of 35 years, the long-term-steady-state replacement rate would be about 330 replacements per year. SCE plans to continue its program of preemptively replacing oil-filled subsurface switches with SF6-filled switches (or vacuum switches sometimes) until all have been replaced. Because the present population of oil-filled radial switches is older than the present population of oil-filled mainline switches, SCE intends to focus pre-emptive switch replacements more on radial switches than on mainline switches at this time. SCE intends to remove these switches at a rate of 200 per year in Capacitor Bank Replacement Program The Capacitor Bank Replacement Program replaces failed and obsolete capacitor banks and their appurtenant capacitor switches. a) Program Necessity 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 to below the 95% of nominal service voltage level that SCE must provide per its tariff and its Electric Service Requirements. Inadequate voltage could damage customers electrical equipment and appliances. Serious voltage drops resulting from inadequate capacitance could conceivably lead to grid collapse. b) Cost Forecast Table III-15 and Figure III-21 below indicate SCE s historical and forecast spend for its capacitor bank replacement program. To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. 57

59 58 Workpaper Southern California Edison / 2018 GRC Table III Historical and Forecast Spending for Capacitor Bank Replacement Year Number of Capacitor Bank Replacements Forecast Unit Cost of Capacitor Bank Replacement (Nominal dollars x 1,000) Recorded/Forecast Cost of Capacitor Bank Replacements (Nominal dollars x 1,000) $7, $6, $8, $7, $8, $48 $12, $49 $17, $50 $17, $52 $18, $53 $18, Refer to WP SCE-02, Vol. 08, pp (Cap Bank unit costs). 58

60 Workpaper Southern California Edison / 2018 GRC 59 Figure III-21 Capacitor Bank Replacement Program 41 Multiple WBSs 42 Recorded /Forecast (CPUC-Jurisdictional Constant 2015 and Nominal $000) c) Justification of Forecast Work There are two types of capacitor banks on our system, switched and fixed. Switched capacitor banks turn on and off to accommodate changes in customer load. Fixed capacitor banks are permanently connected to the circuit. Each capacitor bank is composed of three capacitor units, fuses, a rack, and mounting hardware. For switched banks, capacitor switches (either two or three) and a capacitor control are also included. There are approximately 13,640 capacitor banks in SCE s distribution system. Of these, about 11,388 are installed in the overhead system. The other 2,253 are installed in the underground system. 41 Refer to WP SCE-02, Vol. 08, pp These capital expenditures includes two WBS elements: CET-PD-IR-CB and CET-ET-IR-CR. 59

61 60 Workpaper Southern California Edison / 2018 GRC 1 A graph showing the age distribution of existing capacitor banks is provided in 2 Figure III-22. Figure III-22 Age Distribution of Capacitor Banks Number of Capacitor Banks Year of Installation The expected average time to wear-out of an overhead capacitor bank is assumed to be about 30 years. 43 As of year-end 2015, approximately 1,724 capacitor banks (about 13% of the population) are older than 30 years. Based on a combined population of 13,640 capacitor banks with a mean-time to wear-out of 30 years, the long-term-steady-state replacement rate would be about 450 replacements per year. Inspection of the capacitor banks is a part of our preventive maintenance program. Once every five years, each capacitor bank in our system is inspected for proper operation, corrosion, 43 Refer to WP SCE-02, Vol. 08, pp (Capacitor Banks). 60

62 Workpaper Southern California Edison / 2018 GRC leaking oil, and loose connections. Capacitor banks requiring replacement or repair are recorded and prioritized for follow-up work. Problems with newer capacitor banks usually result in repairs. Problems with older banks, where parts are no longer available and/or where, determined by SCE s Apparatus Technicians, repairs cannot be made cost-effectively, often result in replacement. As Table III-15 demonstrates, over the past five years we have been replacing capacitor banks at an average rate of 267 per year. SCE forecasts a need to replace capacitor banks in years at a rate of 350 per year. This replacement rate was determined based on historical capacitor bank replacement rates and the anticipated long-term-steady-state replacement rate SCE will ultimately see. 7. Automatic Reclosers Replacement Program The Automatic Recloser Program replaces automatic reclosers (ARs) which have been identified as being obsolete and/or unreliable. a) Program Necessity Automatic reclosers (ARs) are used in distribution circuits to interrupt the supply of electricity to that portion of the circuit downstream of its location. They act much like a circuit breaker. However, instead of being at the upstream-most end of the circuit, ARs are typically located out toward the end of the circuit. ARs are installed for two reasons, safety and reliability. On long circuits there may be so much impedance in the conductor that if a fault occurred near the end of the circuit, the circuit breaker in the substation could not detect it. This would mean that a broken overhead conductor lying on the ground would remain energized creating a serious public hazard. Energized conductors are also a serious fire hazard. ARs are often on circuits just before the circuit enters an area of dense vegetation. When a fault occurs downstream of an AR, the AR opens before the circuit breaker in the substation responds to the fault. Only the downstream portion of the circuit is interrupted and all customers upstream of the AR remain energized. Therefore, ARs reduce the number of customers affected by a downstream fault to a fraction of what it would otherwise have been. 61

63 62 Workpaper Southern California Edison / 2018 GRC b) Cost Forecast Table III-16 and Figure III-23 below indicate SCE s historical and forecast spend for the AR replacement program. To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. Table III Historical and Forecast Spending for Automatic Recloser Replacements Year Number of Automatic Recloser Replacements Forecast Unit Cost of Automatic Recloser Replacement (Nominal dollars x 1,000) Recorded/Forecast Cost of Automatic Recloser Replacements (Nominal dollars x 1,000) $1, $ $1, $1, $2, $75 $2, $77 $2, $79 $2, $81 $2, $84 $2, Refer to WP SCE-02, Vol. 08, pp (Automatic Reclosures Unit Cost). 62

64 Workpaper Southern California Edison / 2018 GRC 63 Figure III-23 Automatic Reclosers Replacement Program 45 WBS Element CET-PD-IR-AR Recorded /Forecast (CPUC-Jurisdictional Constant 2015 and Nominal $000) c) Justification of Forecast Work We have approximately 1,379 automatic reclosers installed in our distribution system. Of these, approximately 1,332 are in the overhead system and about 47 are in the underground system. A graph showing the age distribution of ARs in the system is provided in Figure III-24 below: 45 Refer to WP SCE-02, Vol. 08, pp

65 64 Workpaper Southern California Edison / 2018 GRC Figure III-24 Inventory of Distribution Automatic Reclosers Number of Automatic Reclosers Year of Installation The estimated time to wear-out of an AR is assumed to be about 25 years. 46 There are approximately 147 ARs (about 11% of the population) that are older than 25 years. Based on a combined population of 1,379 ARs with a mean-time to wear-out of 25 years, the long-term-steady-state replacement rate would be about 55 replacements per year. The design of ARs has undergone extensive changes since we installed them. The early ARs were oil-filled. The latest reclosers have a vacuum switch and electronic control arrangement. Many of the oldest ARs are no longer manufactured and cannot be repaired or are of an obsolete design, which cannot be repaired cost-effectively. 46 Refer to WP SCE-02, Vol. 08, pp (Automatic Reclosers). 64

66 Workpaper Southern California Edison / 2018 GRC As Table III-16 demonstrates, over the past five years we have been replacing ARs at an average rate of 21 per year. SCE forecasts a need to replace 34 ARs in 2016 and 30 ARs per year in PCB Transformers Replacement Program The PCB Transformer Replacement program replaces distribution line transformers suspected of being contaminated with PCB oil. a) Program Necessity For a period of about 20 years, transformer manufacturers such as Westinghouse and General Electric distributed transformers and other oil-filled electrical equipment containing insulating oil with polychlorinated biphenyls (PCBs) to utilities in the United States. While SCE never specifically ordered transformers containing PCB oil, many transformers were received and installed with oil containing PCB (>50ppm PCB) from the manufacturer. PCB oil is believed to pose a threat to public health and the environment because transformers can leak and can occasionally fail catastrophically. The Environmental Protection Agency (EPA) has encouraged all utilities to remove transformers containing significant levels of PCB. It is expected that the federal government will ultimately make elimination of PCB-containing equipment a legal requirement. In the late 1970s, the United States Congress passed the Toxic Substances Control Act which required the EPA to write regulations banning the manufacture, processing, distribution in commerce and use of PCBs. The EPA s PCB regulations subsequently written provided specific authorizations [40 CFR Part ] allowing electric utilities to continue using transformers and other electrical equipment for the remainder of their useful lives subject to use conditions [761.30(a)(1)] and disposal requirements [761.60]. To date, there are no federal regulations requiring utilities to remove equipment containing PCBs. However, the United States was a participant at the Stockholm Convention on Persistent Organic Pollutants, which set non-binding goals for eliminating PCB use in electrical equipment by

67 66 Workpaper Southern California Edison / 2018 GRC In April 2010, the EPA published an Advance Notice of Proposed Rulemaking describing its reassessment of PCB use authorizations and its consideration of mandatory removal of all equipment containing PCB at levels greater than 50 ppm. In February 2014 the EPA presented at the Small Business Advocacy Review ( SBAR ) Panel meeting for rulemaking indicating their continued focus on developing a regulatory timeline for phase-out of PCB Transformers (>500ppm) and PCB-contaminated transformers (50-500ppm) and interim use conditions on known PCB Transformers and PCB-contaminated transformers. The most substantial challenge that electric utilities may face is the adoption of new regulations by EPA that would require analytical testing of in-service equipment to determine if transformers or other oil filled equipment contain regulated levels of PCBs. While EPA does not require testing as a condition of use, the EPA has a very stringent set of assumption standards [761.2 PCB concentration assumptions for use ] when levels are unknown. With few exceptions, all distribution transformers owned by SCE manufactured prior to July 2, 1979 and whose PCB concentration are unknown must be assumed to contain regulated levels of PCBs (i.e., ppm). These existing EPA assumption standards pose significant management and liability concerns for SCE and our customers. This is due primarily to the fact that the EPA regulates spills and even minor leaks from transformers containing PCBs 50 parts per million or greater as improper disposal of PCBs [761.60]. Under the EPA s Penalty Policy, such improper disposal of PCBs can cause penalties/fines of up to $32,500 per day per incident. SCE estimates there are approximately 2,900 older distribution transformers in our system that may contain over 50 ppm PCBs. If and when older transformers in our system leak and a release of PCBs 50 ppm or greater occurs, even if unknown, SCE is out of compliance with the EPA s regulations (i.e., improper disposal of PCBs ) and subject to enforcement action. Although SCE has in place protocols for the proper response and cleanup of PCB spills under the EPA s PCB Spill Cleanup Policy [ ] and performs well in that regard, it is virtually impossible to monitor the tens of thousands of older distribution transformers remaining in our system for the presence of minor oil leaks. 66

68 Workpaper Southern California Edison / 2018 GRC b) Cost Forecast Table III-17 and Figure III-25 below shows SCE s historical and forecast spending for its PCB-Contaminated Transformer Replacement program. To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. Table III Historical and Forecast Spending for PCB-Contaminated Transformer Replacement Year Number of PCB- Contaminated Transformer Replacements Forecast Unit Cost of PCB- Contaminated Transformer Replacement (Nominal dollars x 1,000) Recorded/Forecast Cost of PCB- Contaminated Transformer Replacements (Nominal dollars x 1,000) $ $ $ $1, $1, $6 $1, $6 $1, $6 $1, $6 $1, $6 $1, Refer to WP SCE-02, Vol. 08, pp (PCB Transformers Unit Cost). 67

69 68 Workpaper Southern California Edison / 2018 GRC Figure III-25 PCB Transformer Replacement Program 48 WBS Element CET-PD-IR-PC Recorded /Forecast (CPUC-Jurisdictional Constant 2015 and Nominal $000) c) Justification of Forecast Work The federal government may adopt more restrictive legislation requiring all US utilities to sample transformers in-service and/or remove PCB containing transformers entirely from their electrical systems by a set date. EPA has indicated they have contemplated using the existing Stockholm Convention on Persistent Organic Pollutants, which set non-binding goals for eliminating PCB use in electrical equipment by In light of this, SCE has instituted a proactive PCB transformer replacement program for suspect PCB Transformers (>500 ppm) and suspect PCB-contaminated transformers ( ppm) which has been highly effective at identifying transformers containing PCB. 48 Refer to WP SCE-02, Vol. 08, pp

70 Workpaper Southern California Edison / 2018 GRC SCE has developed a predictive PCB model using manufacturer data captured during normal attrition. Under normal conditions the PCB concentration is recorded with the serial number for all transformers. Because serial numbers are assigned sequentially during the manufacturing process, transformers whose serial numbers are numerically close are likely to have been filled with the same transformer oil. Operating transformers whose serial numbers are close to those of removed transformers found to be contaminated are then targeted for removal by the PCB-contaminated Transformer Removal program. In 2015, approximately 90% of the targeted transformers were found to contain PCB levels greater than or equal to 50 ppm PCB. SCE has approximately 165,000 distribution transformers in-service that were manufactured prior to Based on EPA assumption standards [761.2 PCB concentration assumptions for use ], equipment manufactured prior to July 2, 1979 and whose PCB concentration are unknown must be assumed to contain regulated levels of PCBs (i.e., parts per million). Through typical attrition and other replacement programs, approximately 1% of distribution transformers removed annually contain PCB (>50ppm). Our proactive PCB Transformer Replacement program, with its forecast accelerated replacement rate of 250 transformers per year, will significantly reduce the balance of all PCB contaminated transformers by B. Substation Infrastructure Replacement Program The Substation Infrastructure Replacement program preemptively replaces major pieces of aging or obsolete substation equipment to minimize the negative effect of aging on system reliability, safety, and operability/maintainability. Three functions are overseen by the SIR program: Transformer Replacement; Circuit Breaker Replacement; and Substation Switchrack Rebuild. An overview of the SIR program process flow is provided in workpapers. 49 Figure II-35 below summarizes our transmission and distribution system and the configuration of AA, A, and B substations. 49 Refer to WP SCE-02, Vol. 08, pp (SIR Process Flow). 69

71 70 Workpaper Southern California Edison / 2018 GRC Figure III-26 Overview of SCE s Transmission and Distribution System Transformer Bank Replacement Substation transformers are major pieces of equipment used to either (a) increase electricity voltage to reduce energy losses during its transmission over long distances, or (b) reduce electricity voltage to make it more practical for the customer. 70

72 Workpaper Southern California Edison / 2018 GRC 71 1 SCE has several classes of substation transformers shown in Table III-18. Table III-18 Type and Number of Substation Transformer Banks Primary Voltage Number of Currently in Service, Year-End kv (AA transformers) kv (A-transformers) or 66 kv (B-transformers) a) AA Bank Replacement AA-Bank transformers are in major substations where they take electricity at the 500kV transmission level and transform it down to 220kV. The SIR program identifies and replaces AA-Bank transformers approaching the end of their service lives, which contain parts known to be problematic or are no longer available, or that can no longer be cost-effectively maintained. The costs of AA-Bank transformer replacement scheduled for years 2016 through 2020 are all under FERC jurisdiction and are, therefore, not discussed further in this testimony. See Figure III-27 below. To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. 71

73 72 Workpaper Southern California Edison / 2018 GRC Figure III-27 Substation Transformer Bank Replacement (AA-Bank, A-Bank, & B-Bank) 50 Multiple WBS Elements 51 Recorded /Forecast (Total Company Constant 2015 and Nominal $000) 50 Refer to WP SCE-02, Vol. 08, pp WBS elements include: CET-ET-IR-TB and a portion CIT-00-OP-NS. 72

74 Workpaper Southern California Edison / 2018 GRC 73 Table III A and B Banks Historical/Forecast Unit Costs 53 Year Voltage Class Number of A & B- Bank Transformer Replacements Forecast Unit Cost of A & B-Bank Transformer Replacement (Nominal dollars x 1,000) Recorded/Forecast Cost of A & B-Bank Transformer Replacements (Nominal dollars x 1,000) kV 3 115kV 8 66kV 16 33kV 6 16kV 2 12kV 1 Total 33 $68, kV 4 115kV 3 66kV 13 33kV 5 16kV 4 12kV 2 Total 27 $67, kV 3 115kV 0 66kV 20 33kV 2 16kV 4 12kV 4 Total 30 $63, kV 5 115kV 4 66kV 14 33kV 6 16kV 4 12kV 1 Total 29 $70, kV 2 115kV 6 66kV 18 33kV 2 16kV 3 12kV 4 Total 33 $72, kV 11 $5,085 $55, kV 1 $1,628 $1,628 66kV 14 $1,256 $17,588 33kV 3 $1,138 $3,414 16kV 6 $1,221 $7,324 12kV 4 $1,174 $4,695 Total 28 $90, kV 3 $5,198 $15, kV 1 $1,662 $1,662 66kV 15 $1,282 $19,236 33kV 3 $1,162 $3,485 16kV 5 $1,246 $6,230 12kV 4 $1,198 $4,793 Total 28 $50, kV 3 $5,326 $15, kV 1 $1,704 $1,704 66kV 15 $1,315 $19,721 33kV 3 $1,191 $3,573 16kV 5 $1,277 $6,387 12kV 4 $1,228 $4,913 Total 28 $52, kV 3 $5,466 $16, kV 1 $1,757 $1,757 66kV 15 $1,356 $20,339 33kV 3 $1,228 $3,685 16kV 5 $1,317 $6,587 12kV 4 $1,267 $5,067 Total 28 $53, kV 3 $5,609 $16, kV 1 $1,813 $1,813 66kV 15 $1,399 $20,986 33kV 3 $1,267 $3,802 16kV 5 $1,359 $6,797 12kV 4 $1,307 $5,229 Total 28 $55, A bank operational dates assume 4 in 2016, 8 in 2017, 4 in 2018, 4 in 2019, and 3 in Refer to WP SCE-02, Vol. 08, pp (A & B Bank Unit Costs). 73

75 74 Workpaper Southern California Edison / 2018 GRC b) A Bank Replacement A-Bank transformers are in major substations where electricity at the 220kV transmission level is transformed down to a subtransmission voltage, either 115kV or 66kv. The SIR program identifies and replaces A-Bank transformers approaching the end of their service lives, which contain parts known to be problematic or are no longer available, or that can no longer be costeffectively maintained. (1) Program Necessity An in-service failure of an A-Bank transformer can be catastrophic. A- Bank transformers typically supply power to large portions of SCE s distribution system servicing hundreds of thousands of customers. While a certain amount of redundancy is built into the A-Bank system, an in-service failure of one piece of equipment would place the system into an N-1 condition, wherein a second failure or system disturbance could cause a massive blackout affecting significantly large areas. So severe are the consequences of such a blackout that SCE believes every reasonable precaution must be taken to prevent it. Although infrequent, in-service failures of A-Bank transformers can be violent. These transformers are oil-filled. Catastrophic failures and ensuing fires can endanger the safety of SCE employees and the operability of nearby equipment. Inspections are helpful in identifying many incipient failures. However, because of the speed at which failure mechanisms can arise and progress, inspections cannot prevent all failures. Therefore, planned preemptive replacements under controlled conditions of transformers approaching the end of their service lives is a prudent and responsible action to minimize the risk of in-service failures. (2) Cost Forecast Table III-19 above indicates the number of A-Bank transformers replaced and the recorded spend from 2011 through 2015 and SCE s forecast work and spend from 2016 through In , all forecast A-Bank replacements are under CPUC jurisdiction. 74

76 Workpaper Southern California Edison / 2018 GRC (3) Justification of Forecast Work Prior to 2012, SCE identified the volume and transformers to be replaced each year largely on the basis of expert judgment, i.e., the judgment of those individuals responsible for inspecting and maintaining the equipment. Beginning in 2012, SCE incorporated formal engineering analysis in combination with expert judgment into the evaluation process. Since then, SCE has been improving the data quality and the data refresh frequency supporting the evaluation process for transformer replacements. The initial stage of that engineering analysis was an evaluation of historical removals and failures of substation transformers to develop a relationship between age and the probability of in-service failure. That relationship for A-Bank transformers is shown as a Weibull curve in Figure III-28 below Refer to WP SCE-02, Vol. 08, pp (Substation Transformer Reliability Model). 75

77 76 Workpaper Southern California Edison / 2018 GRC Figure III-28 A-Bank Weibull Curve 1 The meantime to wear-out of A-Bank transformers is determined to be 43 2 years SCE s system contains 158 A-Bank transformers. The age distribution of these transformers is shown in Figure III-29 below. The average age of SCE s A-Bank transformers is 22 years. 76

78 Workpaper Southern California Edison / 2018 GRC 77 Figure III-29 A-Bank Transformer Inventory 12 Inventory of A Bank Transformer by Year of Installation (as of year end 2015) 10 Number of A Bank Transformer By multiplying the probability of failure by the number of transformers in each age group, the forecasted number of transformers expected to reach the end of their service lives can be determined. That forecast is shown in Table III-20 below. Table III-20 A-Bank Transformers End of Service Life Forecast Based on Weibull Analysis Number of A-Bank Transformer forecast reach end of service life Having forecast the number of A-Bank transformers expected to reach the end of their service lives each year, what remains is the selection of those transformers whose replacement is most urgent. To accomplish this, SCE developed a process to assess each transformer s physical condition or Health Index. This Health Index is a function of factors such as a transformer s age, loading, fault counts, maintenance orders, oil quality, oil dissolved gas analysis results, 77

79 78 Workpaper Southern California Edison / 2018 GRC manufacturer, and severity of consequences that would result from an in-service failure. 55 From this algorithm-derived replacement prioritization, a five-year replacement schedule is drafted. Two adjustments are then made to this draft replacement schedule. First, the draft schedule is adjusted by a team of technical experts by incorporating factors difficult to quantify into the prioritization process such that high-risk transformers are not overlooked. In some cases, this results in additional transformers identified as requiring replacement compared to the Weibull analysis. Since the A-bank population is relatively small, this can cause large percent variances in units requiring replacement in any given year. Second, if equipment conditions allow, the draft schedule is adjusted to optimize the construction aspects of the replacements. In other words, efforts are made to combine multiple projects (involving transformer replacement, circuit breaker replacement, or some other major work activity) at a substation into a larger single project to avoid return visits to that substation within a reasonable planning horizon. In some cases, this results in variances in construction year of transformer replacements compared to the levelized annual results from the Weibull analysis. Scope for 2016 is impacted by three delayed A-banks from previous years and five additional A banks requiring replacement based on the health index analysis and scoring. These factors have resulted in additional A bank replacements in 2016 compared to forecasts for The names, locations, ages, and other specific justification for each A-Bank Transformer to be replaced each year are provided in workpapers. 56 In summary, the replacement of A-Bank transformers is managed by the SIR program, which combines engineering analysis and expert judgment to determine the number of A- Bank transformer replacements and which specific transformers need replacement. c) B Bank Replacement B-Bank transformers are typically in neighborhood substations where they take electricity at the subtransmission level, usually 66kV but sometimes 115kV, transform it down to 33kV, 55 Refer to WP SCE-02, Vol. 08, pp (A-Bank Health Index). 56 Refer to WP SCE-02, Vol. 08, pp (A-bank Replacements ). 78

80 Workpaper Southern California Edison / 2018 GRC kv, 12 kv, or 4 kv, and then send it out onto distribution circuits to feed pole-mounted, padmounted, or subsurface line transformers. The SIR program identifies and replaces B-Bank transformers approaching the end of their service lives, that contain parts known to be problematic or are no longer available, or that can no longer be cost-effectively maintained. (1) Program Necessity The consequences of an in-service failure of a B-Bank transformer can be significant. B-Bank transformers typically supply power to multiple distribution circuits and an inservice failure could cause an outage to thousands of customers. Although infrequent, in-service failures of B-Bank transformers can be violent. These transformers are oil-filled and catastrophic failures and ensuing fires can endanger the safety of SCE employees and the operability of nearby equipment. Inspections are helpful in identifying many incipient failures. However, because of the speed at which failure mechanisms can arise and progress, inspections cannot prevent all failures. Therefore, planned preemptive replacement of transformers approaching the end of their service lives is a prudent and responsible action to minimize the risk of in-service failures. (2) Justification of Forecast Work Prior to 2012, SCE identified the volume and transformers to be replaced each year largely on the basis of expert judgment, i.e., the judgment of those individuals responsible for inspecting and maintaining the equipment. Beginning in 2012, SCE incorporated formal engineering analysis in combination with expert judgment into the evaluation process. Since then, SCE has been improving the data quality and the data refresh frequency supporting the evaluation process for transformer replacements. The initial part of that engineering analysis was an evaluation of historical removals and failures of substation transformers to develop a relationship between age and the 79

81 80 Workpaper Southern California Edison / 2018 GRC 1 2 probability of in-service failure. That relationship for B-Bank transformers is shown in Figure III-30 below as a Weibull curve. 57 Figure III-30 B-Bank Transformer Weibull Curve The meantime to wear-out of B-Bank transformers is 60 years. There are 2,226 B-Bank transformers in SCE s system. The age distribution of these transformers is shown in Figure III-31 below. The average age of SCE s B-Bank transformers is 37 years. 57 Refer to WP SCE-02, Vol. 08, pp (Substation Transformer Reliability Model). 80

82 Workpaper Southern California Edison / 2018 GRC 81 Figure III-31 B-Bank Transformer Inventory 80 Inventory of B Bank Transformer by Year of Installation (as of year end 2015) 70 Number of B Bank Transformer By multiplying the probability of failure by the number of transformers in each age group, the number of transformers reaching the end of their service lives in each future year, , can be determined. For B-Bank transformers, the average number of transformers reaching the end of their service lives each year is shown below in Table III-21. Table III-21 B-Bank Transformers End of Service Life Forecast Voltage Class KV kV kV kV kV Total Having forecasted the number of B-Bank transformers expected to reach the end of their service lives each year, what remains is the selection of those transformers whose replacement is most urgent. To accomplish this, SCE developed a process to assess each transformer s 81

83 82 Workpaper Southern California Edison / 2018 GRC physical condition or Health Index. This Health Index is a function of factors such as a transformer s age, loading, fault counts, maintenance orders, oil quality, oil dissolved gas analysis results, manufacturer, and severity of consequences that would result from an in-service failure. From this algorithm-derived replacement prioritization, a five-year replacement schedule is drafted. Two adjustments are then made to this draft replacement schedule. The draft schedule is adjusted as necessary by a team of technical experts to ensure that factors difficult to quantify are incorporated into the prioritization process such that high-risk transformers are not overlooked. In some cases, this results in additional transformers, or a different mix of transformers, identified as requiring replacement compared to the Weibull analysis. Second, if equipment conditions allow, the draft schedule is adjusted to optimize the construction aspects of the replacements. In other words, efforts are made to combine multiple projects (involving transformer replacement, circuit breaker replacement, or some other major work activity) at a substation into a larger single project to avoid return visits to that substation within a reasonable planning horizon. In summary, the replacement of B-Bank transformers is managed by the Substation Infrastructure Replacement program which combines engineering analysis and expert judgment to ensure that the appropriate number of B-Bank transformers are replaced each year and those which are replaced are the most risk-significant. The names, locations, ages, and reasons for replacement of all B-Bank Transformers to be replaced each year are provided in workpapers. 58 In its Decision on SCE s 2012 GRC, the CPUC stated on p. 155: We expect that SCE will continue to make 30 replacements annually through To assist the Commission, SCE shall document the replacements performed and submit the names, locations, and ages of the replaced transformers in support of future GRC requests in this category. SCE provides the names, locations, and ages of B-Bank transformer historical replacements in workpapers Refer to WP SCE-02, Vol. 08, pp (B Bank Replacements ). 59 Refer to WP SCE-02, Vol. 08, pp (B Bank Replacements ). 82

84 Workpaper Southern California Edison / 2018 GRC Circuit Breaker Replacement Circuit breakers are major pieces of equipment used to interrupt the flow of electricity through a transmission or distribution circuit. Circuit breakers are essential in preventing equipment damage and public injury when faults occur in their downstream circuits. SCE has 12,509 circuit breakers at several major voltage classes: Table III-22 Summary of Circuit Breakers Voltage Bulk Power or Distribution Number Currently in Service 500 kv Bulk Power kv Bulk Power 1, kv Bulk Power kv Distribution kv Distribution 3, kv Distribution kv Distribution kv Distribution 3 16 kv Distribution 1, kv Bulk Power (see note) kv Distribution 4,001 7 kv Distribution kv Distribution 1,342 Total 12,509 Note: 13.8 kv Circuit Breakers typically are used for switching of shunt reactors on tertiary windings of AA-bank transformers in the Bulk Electric System. Therefore, despite the lower voltage, 13.8 kv Circuit Breakers are typically FERC jurisdiction. 83

85 84 Workpaper Southern California Edison / 2018 GRC Figure III-32 Circuit Breaker Replacement Program (220kV 2.4kV) 60 WBS Element CET-ET-IR-CB Recorded /Forecast (Total Company Constant 2015 and Nominal $000) a) Bulk Power Circuit Breaker Replacement Bulk power circuit breakers interrupt the flow of electricity in transmission lines, typically at the 500kV and 220kV voltage levels. The SIR program identifies and replaces bulk power circuit breakers approaching the end of their service lives that contain parts known to be problematic or no longer available, or that can no longer be cost-effectively maintained. The replacement of bulk power circuit breakers is under FERC jurisdiction and is, therefore, not discussed further in this testimony. b) Distribution Circuit Breaker Replacement Distribution circuit breakers are typically located in residential and commercial area substations ( B substations) where electricity is transformed from a subtransmission level voltage, 60 Refer to WP SCE-02, Vol. 08, pp

86 Workpaper Southern California Edison / 2018 GRC usually 66kV but sometimes 115kV, down to a distribution level voltage of either 33kV, 16 kv, 12kV, 4 kv, or 2.4kV. The Distribution Circuit Breaker Replacement program identifies and replaces breakers approaching the end of their service lives and therefore becoming increasingly unreliable that contain parts known to be problematic or unavailable, or that can no longer be cost-effectively maintained. (1) Program Necessity Circuit breakers perform the critical function of turning off the flow of electricity to a circuit which has encountered a problem. These problems are typically events which result in conductors coming into contact with the ground, i.e., faults. These faults, left unmitigated, would allow massive amounts of energy to be drawn through the upstream portion of the circuit destroying the conductor/cable, distribution switches, and substation buses and transformers. Most importantly, faulted conductors which remain energized are serious hazards to the public. Circuit breakers must operate quickly. A properly functioning circuit breaker is expected to detect the overcurrent condition and isolate the circuit in less than one-tenth of a second. (2) Cost Forecast Table III-23 indicates the number of distribution circuit breakers replaced and the recorded spend from 2011 through 2015 and SCE s forecast work and spend from 2016 through To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. 85

87 86 Workpaper Southern California Edison / 2018 GRC Year Table III Circuit Breakers Historical/Forecast Unit Costs Voltage Class Number of Distribution Circuit Breaker Replacements Forecast Unit Cost of Distribution Circuit Breaker Replacement (Nominal dollars x 1,000) Recorded/Forecast Cost of Distribution Circuit Breaker Replacements (Nominal dollars x 1,000) kV kV 31 Total 113 $28, kV kV 52 Total 90 $24, kV kV 94 Total 108 $35, kV kV 217 Total 253 $60, kV kV 180 Total 217 $47, kV 36 $278 $9, kV 184 $162 $29,868 Total 220 $39, kV 37 $283 $10, kV 184 $166 $30,474 Total 221 $40, kV 37 $290 $10, kV 184 $170 $31,227 Total 221 $41, kV 38 $300 $11, kV 182 $175 $31,857 Total 220 $43, kV 39 $309 $12, kV 180 $181 $32,493 Total 219 $44, Refer to WP SCE-02, Vol. 08, pp (CB Unit Costs). 86

88 Workpaper Southern California Edison / 2018 GRC (3) Justification of Work Forecast Prior to 2012, SCE identified the volume and specific circuit breakers to be replaced each year largely on the basis of expert judgment, i.e., the judgment of those individuals responsible for inspecting and maintaining the equipment. Beginning in 2012, SCE incorporated formal engineering analysis in combination with expert judgment into the evaluation process. Since that time, SCE has been improving the data quality and the data refresh frequency supporting the evaluation process for transformer replacements. Part of that engineering analysis was an evaluation of historical removals and failures of circuit breakers to develop a relationship between age and the probability of in-service failure. That relationship for distribution circuit breakers is shown below in Figure III-33 as a Weibull curve Refer to WP SCE-02, Vol. 08, pp (Circuit Breaker Reliability). 87

89 88 Workpaper Southern California Edison / 2018 GRC Figure III-33 Circuit Breaker Weibull Curve From this curve, the mean time to wear-out of distribution circuit breakers was determined to be 49 years. There are 11,229 distribution circuit breakers in SCE s system, of which 4,051 breakers are 115kV or 66kV and 7,178 breakers are 55 kv through 2.4kV. The average age of the population of 115kV 66kV circuit breakers is 18 years. The average age of the population of 55kV and below circuit breakers is 28 years. Distributing the 115kV 66kV circuit breakers by year of installation is shown below in Figure III

90 Workpaper Southern California Edison / 2018 GRC 89 Figure III kV and 66kV Circuit Breaker Inventory 300 Inventory of 115kV and 66kV Circuit Breakers by Year of Installation (as of December 31, 2015) Number of Distribution Circuit Breakers shown below in Figure III-35. Distributing the 33kV 2.4kV circuit breakers by year of installation is Figure III-35 33kV and 2.4kV Circuit Breaker Inventory 450 Inventory of 33kV 2.4kV Circuit Breakers by Year of Installation (as of December 31, 2015) 400 Number of Distribution Circuit Breakers By multiplying the probability of failure of distribution circuit breakers (shown in Figure III-34 and Figure III-35) by the number of circuit breakers in each age group, the 89

91 90 Workpaper Southern California Edison / 2018 GRC number of circuit breakers reaching the end of their service lives in future years can be determined. For 115kV 66kV circuit breakers, the forecast average number of circuit breakers reaching the end of their service lives each year is provided below in Table III-24. Table III-24 Forecast Circuit Breaker Wear-Out Rate, 115kV 66kV Forecast Circuit Breaker Wear Out Rate, 66kV 115kV Circuit Breaker Voltage kv For 55kV and below circuit breakers, the forecast average number of circuit breakers reaching the end of their service lives each year is provided below in Table III-25. Table III-25 Forecast Circuit Breaker Wear-Out Rate, 55kV and Below Forecast Circuit Breaker Wear Out Rate, 2.4kV 33kV Circuit Breaker Voltage kv Having forecasted the number of circuit breakers expected to reach the end of their service lives each year, what remains is the selection of those circuit breakers whose replacement is most urgent. To accomplish this, SCE developed a process to assess each circuit breaker s physical condition or Health Index. This Health Index is a function of factors such as a circuit breaker s age, number of operations, number of faults experienced, the mechanism, the number of corrective maintenance orders, results of oil analysis if applicable, and severity of consequences that would result from an in-service failure. From this algorithm-derived replacement prioritization, a fiveyear replacement schedules are drafted. Two adjustments are made to these draft schedules. First, the draft schedule is adjusted by a team of technical experts to ensure that factors difficult to quantify are incorporated into the prioritization process such that high-risk circuit breakers are not overlooked. In 90

92 Workpaper Southern California Edison / 2018 GRC some cases, this may mean that circuit breakers are added to the five-year plan based on real-time condition. Second, if equipment conditions allow, the draft schedule is adjusted to optimize the construction aspect of the replacements. In other words, efforts are made to combine multiple projects (involving transformer replacement, circuit breaker replacement, or some other major work activity) at a substation into a larger single project to avoid return visits to a substation within a reasonable planning horizon. In summary, the replacement of distribution circuit breakers is managed by the Substation Infrastructure Replacement program which combines engineering analysis and expert judgment to ensure that the number of circuit breakers are replaced each year and those which are replaced are most risk-significant. The names, locations, ages, and reasons for replacement of all Circuit Breakers to be replaced each year are provided in workpapers. 63 (4) Analysis of Circuit Breaker Replacements in Other Programs In the 2015 GRC decision, SCE was directed to provide analysis of the preemptive replacements of Circuit Breakers in the Substation Infrastructure Replacement program in combination with other types of replacements. In addition to breakdown replacements, SCE removes existing circuit breakers and replaces them with new circuit breakers in programs such as the existing Substation Infrastructure Replacement program, existing load growth programs such as the Substation Equipment Replacement Program (SERP), and the new Substation Switchrack Rebuild Program. The current forecast of Circuit Breaker wear-out rate shown in Table III- 22 and Table III-23 sums to approximately 220 circuit breakers per year. This forecast is based on existing inventory, and does not account for forecast changes in the existing inventory for other programs. Therefore, SCE has performed a sensitivity analysis on the wear-out rate forecast for Circuit 63 Refer to WP SCE-02, Vol. 08, pp (CB Replacements ). 91

93 92 Workpaper Southern California Edison / 2018 GRC Breakers. Specifically, SCE has performed analysis to determine what impact replacements forecast under SERP and the Substation Switchrack Rebuild Program would have on the forecast wear-outs described above in Table III-22 and Table III-23. The results of this sensitivity analysis have shown that replacements in SERP and the Substation Switchrack Rebuild Program in the coming years will have a negligible impact on the system-wide circuit breaker wear-out forecast. 64 Specifically, the wear-out rate would be reduced by approximately 26 Circuit Breakers in total during years , or on average approximately 5 Circuit Breakers per year. Since the forecast based on existing inventory anticipates wear-out of approximately 220 circuit breakers per year, the impact of circuit breaker replacements in other programs on this forecast is very small (less than 2.5% per year, on average, over the next five years). 2. Substation Switchrack Rebuild This capital activity captures the expenditures related to rebuilding existing substation racks based on conditions found in the field. The Substation Switchrack Rebuild program scope is initially driven by the CB or Transformer Bank replacement programs. For each CB or transformer bank replacement project, Engineering performs a pre-engineering job walk to gather site information that is not available prior to a site visit. The need to rebuild a particular rack at a substation would be identified during the job walk. We develop specific switchrack rebuild requirements to support the completion of the CB and Transformer Bank replacements. Without the substation rack rebuilds, the existing switchracks could not accommodate the infrastructure replacement equipment for CB and Transformer Banks. Substation Switchrack Rebuild projects typically include the following elements: the demolition of existing inadequate substation switchrack structures; the construction of new switchrack structures consistent with current design standards; and the removal and replacement of substation assets within the switchrack, such as circuit breakers, disconnects, and transformers. Sometimes, these projects 64 Refer to WP SCE-02, Vol. 08, pp (Circuit Breaker Forecast Sensitivity Analysis). 92

94 Workpaper Southern California Edison / 2018 GRC also include: the relocation of existing substation fences within the SCE property line; the addition of a Mechanical Electrical Equipment Room (MEER); and other scope as needed. a) Program Necessity We started the Substation Switchrack Rebuild program in 2015 to address problems related to planning and prioritizing Infrastructure Replacement projects at substation locations with inadequate switchrack structures. Typical Substation IR projects, such as circuit breaker replacement projects or B- bank or A-bank transformer replacement projects, involve a pre-engineering jobwalk after project initiation to identify field-specific conditions that must be considered during design. These jobwalks typically take place the year prior to the operating date of the project. Usually, scope changes identified during these jobwalks are relatively minor in magnitude, have little or no impact in operating date of the project, and are processed, approved and tracked usually. In certain cases, however, serious problems with existing substation switchracks are identified during the pre-engineering jobwalks. These problems are typically much larger in magnitude and have significant impacts on the project scope, cost, and schedule. These problems are typically related to substation switchrack structures being antiquated, not having enough room for larger modern equipment, or not being up to existing design standards such as not having sufficient clearances or not being seismically sound due to deterioration of foundation or structure. Such cases fall into one of three categories: (1) Lattice-steel or pipe-steel structures; (2) wood-pole structures; and (3) cubicle switchgear. In cases of lattice-steel or pipe-steel structures, these structures do not meet current seismic requirements. These structures typically cannot support new disconnect installations due to equipment weight and clearance considerations. Usually, these structures also cannot accommodate modern circuit breaker model types. Finally, often, the ground grid that exists is inadequate and foundations that support equipment within these structures are crumbling or in otherwise poor condition. Wood pole structures have many of the same problems that lattice-steel or pipesteel structures have, including lack of meeting seismic requirements, inability to support new 93

95 94 Workpaper Southern California Edison / 2018 GRC disconnect installations, inability to accommodate modern circuit breaker types, and inadequate ground grids. Wood pole structures have additional problems, however, related to the wood poles themselves. Specifically, it is typically impossible to replace a single wood pole attached to an existing wood pole switchrack. Intrusive inspections are typically not performed on wood poles used within a substation wood pole switchrack. In cases of cubicle switchgear, the biggest problem is that maintenance work areas are tight and pose significant operational and safety concerns for SCE operations and maintenance personnel. In addition, existing equipment in this switchgear is antiquated and spare parts are extremely limited or unavailable. Existing cubicle switchgear cannot accommodate modern circuit breaker model types. Sometimes, cubicle switchgear substations do not have ties to neighboring stations, and if switchgear equipment fails in service, there are limited abilities to restore load in a quick fashion. Such substations may require additional land to build either new modern cubicle switchgear or a new conventional open-air switchrack. Finally, duct replacements, underground vault replacements, cable termination replacements, and cable replacements are often necessary based on field conditions at these cubicle switchgear locations. b) Cost Forecast Table III-26 and Figure III-36 below shows SCE s historical and forecast spending for its Substation Switchrack Rebuild program. To develop the cost estimates, SCE analyzed historical cost data to develop unit costs. This unit cost was applied to our forecast units to develop our annual forecasts. 94

96 Workpaper Southern California Edison / 2018 GRC 95 Table III Substation Switchrack Rebuild Historical/Forecast Unit Costs Year Number of Switchrack Rebuilds Forecast Unit Cost of Switchrack Rebuilds (Nominal dollars x 1,000) Recorded/Forecast Cost of Switchrack Rebuilds (Nominal dollars x 1,000) $ $ $ $ $1, $5,244 $5, $5,353 $16, $5,488 $16, $5,660 $16, $5,840 $17, Refer to WP SCE-02, Vol. 08, pp (Switchrack Rebuild Unit Costs). 95

97 96 Workpaper Southern California Edison / 2018 GRC Figure III-36 Substation Switchrack Rebuild Program 66 WBS Element CET-ET-IR-RB Recorded /Forecast (CPUC-Jurisdictional Constant 2015 and Nominal $000) c) Justification of Work Forecast Our process for substation switchrack rebuild projects is: existing projects ( driver projects such as SubIR or load growth) in the current 5-year horizon are reviewed by substation engineering for their potential to drive the need for a substation switchrack rebuild. When identified projects are found anticipated to pose switchrack issues, jobwalks focused on rebuild evaluation are initiated. These jobwalks are typically organized based on geographic location, include stakeholders such as SC&M, Infrastructure Replacement, Substation Engineering, Civil/Structural Engineering, and Field Engineering. At the jobwalks, the condition of foundations, equipment, structures, and working areas/clearances are evaluated as a whole. 66 Refer to WP SCE-02, Vol. 08, pp

98 Workpaper Southern California Edison / 2018 GRC If the driver project can be constructed without a rebuild of the substation switchrack, then the driver project continues through its lifecycle. However, if it is determined that a switchrack rebuild is required, the driver project is typically cancelled and a new switchrack rebuild project is initiated. Importantly, the operating date of the driver project is typically extended out one year from the previous operating date, to accommodate the more complex design and construction issues that will result from the added scope. 97

99 2018 General Rate Case Index of Workpapers SCE-02, Vol. 08 DOCUMENT PAGE(S) WP SCE-02 T&D-Vol. 08-Capital Detail by WBS Element 1-17 WP SCE-02 T&D-Vol. 08-Capital Authorized vs. Recorded Detail WP SCE-02 T&D-Vol. 08-Cost of Worst Circuit Rehabilitation WP SCE-02 T&D-Vol. 08-Underground Cable Reliability WP SCE-02 T&D-Vol. 08-Impact of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability WP SCE-02 T&D-Vol. 08-Cost of Cable Testing WP SCE-02 T&D-Vol. 08-Cost of Cable Injection WP SCE-02 T&D-Vol. 08- Mainline Cable Testing Analysis WP SCE-02 T&D-Vol. 08-Average Age of WCR WP SCE-02 T&D-Vol. 08-Cost of Cable in Conduit Replacements WP SCE-02 T&D-Vol. 08-OCP Scope for 2016 and WP SCE-02 T&D-Vol. 08-Cost of Overhead Conductor Rebuilds WP SCE-02 T&D-Vol. 08-Cost of Underground Oil Switch Replacements WP SCE-02 T&D-Vol. 08-Underground Oil Switch Reliability WP SCE-02 T&D-Vol. 08-Cost of Capacitor Bank Replacements WP SCE-02 T&D-Vol. 08-Capacitor Banks WP SCE-02 T&D-Vol. 08-Cost of Automatic Reclosure Replacements WP SCE-02 T&D-Vol. 08-Automatic Reclosures WP SCE-02 T&D-Vol. 08-Cost of PCB Transformer Replacements WP SCE-02 T&D-Vol. 08-Substation IR Process WP SCE-02 T&D-Vol. 08-Cost of Transformer Bank Replacements WP SCE-02 T&D-Vol. 08-Substation Transformer Reliability Model WP SCE-02 T&D-Vol. 08-A Bank Health Index WP SCE-02 T&D-Vol. 08-A Bank Replacements WP SCE-02 T&D-Vol. 08-B Bank Replacements WP SCE-02 T&D-Vol. 08-B Bank Replacements WP SCE-02 T&D-Vol. 08-Cost of Circuit Breaker Replacements WP SCE-02 T&D-Vol. 08-Substation Circuit Breaker Reliability Model WP SCE-02 T&D-Vol. 08-CB Replacements WP SCE-02 T&D-Vol. 08-CB Forecast Sensitivity Analysis WP SCE-02 T&D-Vol. 08-Cost of Switchrack Rebuilds

100 Workpaper Southern California Edison / 2018 GRC 1 Workpaper Title: Capital Detail by WBS Element

101 2 Workpaper Southern California Edison / 2018 GRC Southern California Edison Company CAPITAL WORKPAPERS - T&D OU SUMMARY, BY WITNESS, BY EXHIBIT (Nominal $000) Item Testimony Description Forecast Capital Expenditures Total 1 Jose Ramon Goizueta 527, , , , ,679 2,562, SCE-02, Vol , , , , ,679 2,562, CET-ET-IR-RB CET-ET-IR-TB CET-ET-IR-TB CET-ET-IR-TB CET-ET-IR-CB CET-ET-IR-CB CET-PD-IR-WC-MTW CET-PD-IR-SR-MTW CET-PD-IR-PC-MTW CET-PD-IR-OC-MTW CET-PD-IR-CB-MTW CET-PD-IR-CC-MTW CET-PD-IR-LE-MTW CET-PD-IR-LE-MTE CET-PD-IR-AR-MTW 5,244 16,060 16,464 16,981 17,521 72,270 31,172 15,932 16,325 16,755 34, ,572 55,932 15,593 15,978 16,399 16, ,731 34,649 35,406 36,298 37,437 38, ,417 39,861 40,957 41,974 43,239 44, ,575 5,279 5,876 6,021 6,179 6,859 30, , , , , , ,611 10,923 11,150 12,701 13,099 13,516 61,389 1,385 1,413 1,449 1,494 1,542 7, , , , , , ,160 12,005 17,156 17,588 18,140 18,717 83,605 26,528 31,142 41,643 42,949 44, ,578 12,196 13,833 14,181 14,626 15,091 69,928 9,374 9,569 9,810 10,117 10,439 49,309 2,565 2,310 2,368 2,443 2,520 12,206

102 Workpaper Southern California Edison / 2018 GRC 3 Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4057 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: AUTO RECLOSURE REPL SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Distribution Infrastructure Replacement AUTO RECLOSURE REPL 611 INFRASTRUCTURE REPL AUTO RECLOSURE REPL CET-PD-IR-AR-MTW 3. PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Lines 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: No 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,520 Total 12,206 2,600 2,550 2,500 2,450 2,400 2,350 2,300 2,250 2,200 2, GRC - Capital Expenditures Forecast (a). SYSTEM SHORT TEXT: Metro West 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Auto Reclosure Repl (486) See workpaper for details 9. JUSTIFICATION: See workpaper for details

103 4 Workpaper Southern California Edison / 2018 GRC Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4057 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS SCE-02, Vol. 08 PROJECT DETAIL WORKSHEET: CABLE LIFE EXT INFRASTRUCTURE REPLACEMENT PROGRAMS Distribution Infrastructure Replacement CABLE LIFE EXT 613 INFRASTRUCTURE REPL CABLE LIFE EXT CET-PD-IR-LE-MTW 3. PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Lines 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: Yes 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,091 Total 69,928 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2, GRC - Capital Expenditures Forecast (a). SYSTEM SHORT TEXT: Metro West 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Cable Injection See workpaper for details 9. JUSTIFICATION: See workpaper for details

104 Workpaper Southern California Edison / 2018 GRC 5 Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4057 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS SCE-02, Vol. 08 PROJECT DETAIL WORKSHEET: CABLE LIFE EXT INFRASTRUCTURE REPLACEMENT PROGRAMS Distribution Infrastructure Replacement CABLE LIFE EXT 612 INFRASTRUCTURE REPL CABLE LIFE EXT CET-PD-IR-LE-MTE 3. PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Lines 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: Yes 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,439 Total 49,309 10,600 10,400 10,200 10,000 9,800 9,600 9,400 9,200 9,000 8, GRC - Capital Expenditures Forecast (a). SYSTEM SHORT TEXT: Metro East 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Cable Testing See workpaper for details 9. JUSTIFICATION: See workpaper for details

105 6 Workpaper Southern California Edison / 2018 GRC Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4057 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: CBLE IN CONDUIT REPL SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Distribution Infrastructure Replacement CBLE IN CONDUIT REPL 614 INFRASTRUCTURE REPL CBLE IN CONDUIT REPL CET-PD-IR-CC-MTW 3. PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Lines 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: Yes 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , ,949 50,000 40,000 30,000 20, GRC - Capital Expenditures Forecast ,315 10,000 Total 186, (a). SYSTEM SHORT TEXT: Metro West 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Cable In Conduit Repl See workpaper for details 9. JUSTIFICATION: See workpaper for details

106 Workpaper Southern California Edison / 2018 GRC 7 Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4057 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: DISTR CAP BANK REPL SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Distribution Infrastructure Replacement DISTR CAP BANK REPL 617 INFRASTRUCTURE REPL DISTR CAP BANK REPL CET-PD-IR-CB-MTW 3. PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Lines 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: No 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,717 20,000 15,000 10,000 5, GRC - Capital Expenditures Forecast Total 83, (a). SYSTEM SHORT TEXT: Metro West 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Distr Cap Bank Repl (485) See workpaper for details 9. JUSTIFICATION: See workpaper for details

107 8 Workpaper Southern California Edison / 2018 GRC Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4057 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS SCE-02, Vol. 08 PROJECT DETAIL WORKSHEET: OH CONDUCTOR INFRASTRUCTURE REPLACEMENT PROGRAMS Distribution Infrastructure Replacement OH CONDUCTOR 618 INFRASTRUCTURE REPL OH CONDUCTOR CET-PD-IR-OC-MTW 3. PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Lines 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: No 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , , , , , GRC - Capital Expenditures Forecast , ,000 Total 710, , (a). SYSTEM SHORT TEXT: Metro West 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Overhead Conductor Program See workpaper for details 9. JUSTIFICATION: See workpaper for details

108 Workpaper Southern California Edison / 2018 GRC 9 Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4057 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: PCB TRANSFORMERS SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Distribution Infrastructure Replacement PCB TRANSFORMERS 619 INFRASTRUCTURE REPL PCB TRANSFORMERS CET-PD-IR-PC-MTW 3. PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Lines 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: No 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,542 1,600 1,550 1,500 1,450 1,400 1, GRC - Capital Expenditures Forecast Total 7,283 1, (a). SYSTEM SHORT TEXT: Metro West 7 (b). DETAILED DESCRIPTION: 8. SCOPE: PCB Transformers See workpaper for details 9. JUSTIFICATION: See workpaper for details

109 10 Workpaper Southern California Edison / 2018 GRC Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4057 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: SWITCH REPLACEMENT SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Distribution Infrastructure Replacement SWITCH REPLACEMENT 621 INFRASTRUCTURE REPL SWITCH REPLACEMENT CET-PD-IR-SR-MTW 3. PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Lines 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: Yes 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,516 Total 61,389 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2, GRC - Capital Expenditures Forecast (a). SYSTEM SHORT TEXT: Metro West 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Switch Replacement (483) See workpaper for details 9. JUSTIFICATION: See workpaper for details

110 Workpaper Southern California Edison / 2018 GRC 11 Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4057 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: WORST CIRCUIT REHAB SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Distribution Infrastructure Replacement WORST CIRCUIT REHAB 625 INFRASTRUCTURE REPL WORST CIRCUIT REHAB CET-PD-IR-WC-MTW 3. PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Lines 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: Yes 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , , , , , GRC - Capital Expenditures Forecast , ,000 Total 651, , (a). SYSTEM SHORT TEXT: Metro West 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Worst Circuit Rehab (481) See workpaper for details 9. JUSTIFICATION: See workpaper for details

111 12 Workpaper Southern California Edison / 2018 GRC Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4329 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: CIRCUIT BREAKERS SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Substation Infrastructure Replacement CIRCUIT BREAKERS 616 INFRASTRUCTURE REPL CIRCUIT BREAKERS CET-ET-IR-CB PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Substations 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: Yes 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,544 Total 210,575 45,000 44,000 43,000 42,000 41,000 40,000 39,000 38,000 37, GRC - Capital Expenditures Forecast (a). SYSTEM SHORT TEXT: Replace Non-Bulk CBs-115kV & Below(CPUC) 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Substation IR-Circuit Breakers See workpaper for details 9. JUSTIFICATION: See workpaper for details

112 Workpaper Southern California Edison / 2018 GRC 13 Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4211 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: CIRCUIT BREAKERS SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Substation Infrastructure Replacement CIRCUIT BREAKERS 615 INFRASTRUCTURE REPL CIRCUIT BREAKERS CET-ET-IR-CB PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Transmission Substations 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: Yes 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,859 Total 30,213 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1, GRC - Capital Expenditures Forecast (a). SYSTEM SHORT TEXT: Replace Bulk CBs - 220kV & 500kV (FERC) 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Substation IR-Circuit Breakers See workpaper for details 9. JUSTIFICATION: See workpaper for details

113 14 Workpaper Southern California Edison / 2018 GRC Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 7713 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: SUB SW RACK REBUILDS SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Substation Infrastructure Replacement SUB SW RACK REBUILDS 620 INFRASTRUCTURE REPL SUB SW RACK REBUILDS CET-ET-IR-RB PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Substations 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: No 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,521 20,000 15,000 10,000 5, GRC - Capital Expenditures Forecast Total 72, (a). SYSTEM SHORT TEXT: Substation Switchrack Rebuilds (CPUC) 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Substation Switchrack Rebuilds See workpaper for details 9. JUSTIFICATION: See workpaper for details

114 Workpaper Southern California Edison / 2018 GRC 15 Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 5210 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: TRANSFORMER BANKS SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Substation Infrastructure Replacement TRANSFORMER BANKS 624 INFRASTRUCTURE REPL TRANSFORMER BANKS CET-ET-IR-TB PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Transmission Substations 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: Yes 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,388 Total 114,572 40,000 35,000 30,000 25,000 20,000 15,000 10,000 5, GRC - Capital Expenditures Forecast (a). SYSTEM SHORT TEXT: Replace Bulk Xfmrs - AA & A Banks (FERC) 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Transformer Banks See workpaper for details 9. JUSTIFICATION: See workpaper for details

115 16 Workpaper Southern California Edison / 2018 GRC Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 5210 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: TRANSFORMER BANKS SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Substation Infrastructure Replacement TRANSFORMER BANKS 623 INFRASTRUCTURE REPL TRANSFORMER BANKS CET-ET-IR-TB PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Transmission Substations 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: Yes 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,828 60,000 50,000 40,000 30,000 20,000 10, GRC - Capital Expenditures Forecast Total 120, (a). SYSTEM SHORT TEXT: Replace Bulk Xfmrs - AA & A Banks (CPUC) 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Transformer Banks See workpaper for details 9. JUSTIFICATION: See workpaper for details

116 Workpaper Southern California Edison / 2018 GRC 17 Organizational Unit Summary: T&D Testimony Summary: Major Program: Category: Detail: RO Model ID: 1. WITNESS: Jose Ramon Goizueta 2. PROJECT DETAIL: Program Group: Program: Pin #: 4331 WBS Element: SOUTHERN CALIFORNIA EDISON COMPANY CAPITAL WORKPAPERS - INFRASTRUCTURE REPLACEMENT PROGRAMS PROJECT DETAIL WORKSHEET: TRANSFORMER BANKS SCE-02, Vol. 08 INFRASTRUCTURE REPLACEMENT PROGRAMS Substation Infrastructure Replacement TRANSFORMER BANKS 622 INFRASTRUCTURE REPL TRANSFORMER BANKS CET-ET-IR-TB PROJECT TYPE: Select one Compliance: Safety, Environmental, Licenses Customer Growth X Replacements in Kind Load Growth Upgrades Reliability Blanket: F&E, Tools, Spare Parts, Lab, Computer Equip Capitalized Software Various (See Below) 4. ASSET TYPE: Distribution Substations 5. CLOSE DATE: Specific Blanket 6. RIIM ELIGIBLE: Yes 7. COST ESTIMATES (NOMINAL $000): Year SCE $ , , , , ,627 Total 182,417 39,000 38,000 37,000 36,000 35,000 34,000 33,000 32, GRC - Capital Expenditures Forecast (a). SYSTEM SHORT TEXT: Replace Non-Bulk Xfmrs - B Banks (CPUC) 7 (b). DETAILED DESCRIPTION: 8. SCOPE: Transformer Banks See workpaper for details 9. JUSTIFICATION: See workpaper for details

117 18 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Capital Authorized vs. Recorded Detail

118 Workpaper Southern California Edison / 2018 GRC 19 IR Authorized vs. Recorded Workpaper - $Millions Program Waterfall Chart Category 2015 Requested 2015 Authorized 2015 Recorded Rec.vs Auth. Var. Worst Circuit Rehabilitation Worst Circuit Rehabilitation Cable Life Extension Cable Life Extension (15) CIC Replacement CIC Replacement (22) Overhead Conductor Program Overhead Conductor Program Underground Oil Switch Replacement Other Capacitor Bank Replacement Other (5) Automatic Reclosure Replacement Other PCB Transformer Replacement Other (1) Substation Transformer Bank Replacement Substation Transformer Bank Replacement Substation Circuit Breaker Replacement Substation Circuit Breaker Replacement Substation Switchrack Rebuilds Other Total $350 $309 $392 $83

119 20 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Cost of Worst Circuit Rehabilitation

120 Workpaper Southern California Edison / 2018 GRC 21 Work paper Title: COST OF WORST CIRCUIT REHABILITATION WBS Element: CET PD IR CR and CET PD IR WC Witness: Jose Ramon Goizueta This workpaper establishes the average unit cost of rehabilitating a worst performing circuit Worst Circuit Rehabilitation Counts 1,2,& Total Count Worst Circuit ,474 Worst Circuit Rehabilitation Expenditures 2015 $ 1& Total Worst Circuit $ 60,134,200 $ 60,264,171 $ 102,188,313 $ 131,494,875 $ 95,766,017 $ 449,847,575 Worst Circuit Rehabilitation Unit Costs ' Average Worst Circuit $ 203,212 $ 313,963 $ 340,621 $ 325,977 $ 338,180 $ 305,098 Unit Cost Used for Forecasting 4& Worst Circuit $ 338,180 $ 344,568 $ 351,732 $ 360,591 $ 371,903 $ 383,728 1 Annual expenditures and unit counts based on projects that occurred from 2011 to Annual expenditures and units counts based on projects that have been closed (construction completed, finalized accounting and financials) 3 Unit cost per year is derived by taking total expenditures and dividing by total unit counts 4 Unit Cost used for forecasting includes escalation 5 Last year recorded was chosen because of increasing cost over the last five years 6 Unit counts are in conductor miles

121 22 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Underground Cable Reliability

122 Workpaper Southern California Edison / 2018 GRC 23 CableFailureModel_Dec2015 SUBSTATION APPARATUS & STANDARDS GROUP REPORT Underground Cable Reliability Model Update May 2016 Prepared by: S. H. Chien 6/26/2016

123 24 Workpaper Southern California Edison / 2018 GRC Underground Cable Reliability Model Update I. Objective Purpose of this analysis is to update the underground (UG) cable Weibull failure model based on the recent fourteen years of SCE cable inventory records (2000 thru 2013). This is a follow-up study of the June 2010 report generated for the 2015 GRC. Results of this analysis can be used to develop infrastructure replacement (IR) strategy, assess effectiveness of PRISM risk ranking and other asset management programs. Results of this assessment include: - Weibull distribution parameters for four types of UG cable - Age dependent cable failure rates, and - Cable mean-time-to-failure (MTTF) II. Approach This effort begins with the collection of cable removal record through service years and establishment of cable histogram for all cable inventory. For the four types of cable installed in the SCE territory over the past sixty years, separate Weibull parameters were derived based on individual retirement/replacement record. Adjustment of cable data is required to account for effects of cable infrastructure replacement program implemented in recent years. Validation of the derived Weibull cable failure rates was performed by comparing the calculated cable failures against the distribution system interruption records queried from the ODRM outage data. The cable MTTF can be derived from the respective reliability model then modified with engineering judgment, if needed. III. Cable Weibull failure rate development Both main line and radial cable are included in the scope of analysis. Major effort is the preparation of input files for Reliasoft Weibull ++9 software package. Use XLPE cable as example, the data span covers full spectrum of XLPE installed years (1971 thru 1998). After test out Weibull results on reasonable combinations of cable removal records, the most representative cable reliability is derived from XLPE installed between 1982 and 1990 (justification of this decision is provided in Sec III.a). Also addressed by this assessment are: non-age dependent random failure and new cable break-in failures. Examples of non-age dependent cable failures are: dig-in, animal, vandalism, etc. Based on the ODRM data, non-age related failure is a small contributor (but not-negligible) to overall cable induced outages. This failure mechanism amounts to 7.8% of total cable introduced outages. 2

124 Workpaper Southern California Edison / 2018 GRC 25 Cable break-in failures mostly associated with quality of extrusion technology, insulation material cleanness control, and workmanship during cable fabrication, etc. Operating experiences cumulated from early removal of TR-XLPE forms foundation to derive the infant-fatality of cable failure rate. For completeness of the analysis, this failure rate model also credits effects of the cable infrastructure replacement (IR) program implemented since Following is the summary of specific tasks associated with Weibull parameter development for each cable type. a. Develop XLPE failure rate Series of Weibull++ runs were carried out to establish the best-estimate combination of parameters to represent the reliability of XLPE. After test out Weibull results on all reasonable combinations of cable removal records, the most representative cable reliability is derived from XLPE installed between 1982 and 1990, justification of this selection includes: 1982 thru 1990 represent the mid-section of XLPE service years 1982 thru 1990 represents the largest in-service inventory of XLPE. For older/younger cable population, Weibull++ generated results can be biased. b. Develop TR-XLPE failure rate based on XLPE results As this type of cable design was adopted by SCE after year 1999, the oldest TR-XLPE is at age 15 and there is in-sufficient data to forecast end of life behavior for this new type of cable. XLPE is the cable with design parameters close to that of TR-XLPE, we can use some of its performance records to project future behavior of the tree-resistant XLPE. Compare removal rate of XLPE and TR-XLPE at the same vintage, it is evident that TR-XLPE does show improved performance over XLPE during early stage of operation. By collecting inputs of available cable experts and Apparatus Engineering cable specialist, we choose to establish 5 year difference in median/mean life for the two types of cables. The choice of TR-XLPE mean-time-to fail (MTTF) to be five years longer than XLPE may still be conservative. Another improvement in the TR-XLPE performance is the deferred initiation of electric tree defect within the cable insulator. It is estimated (Ref 1 & 2) that the introduction of tree defect inside TR-XLPE can be delayed by about 20 years due to adoption of the improved cable fabrication process. Based on these inputs, the derived TR-XLPE Weibull parameters are shown in Table 1. c. Develop PILC failure rate All combinations of PILC Weibull runs gives consistent mean life for this type of cable. Reliability parameters calculated for PILC installed between 1964 and 1967 was chosen, as this leads to a somewhat smaller beta value. i.e. less dramatic rise in failure rate at final stage of service life (which is closer to the trend shown by the PILC removal records). 3

125 26 Workpaper Southern California Edison / 2018 GRC d. Develop HMW failure rate and expected mean life: In earlier cable model development, we have a calculated HMW MTTF of 22 year and initial failure rate 2E-2/yr/mile (more than 3 times higher than that of XLPE). This accounts for poor HMW performance during early years of its service. There is no change on this assessment and its conclusion. On the other hand, the Weibull generated MTTF for the remaining HMW cable is 50.9 year. This reflects the low HMW cable removal volume during final stage of service shown by this type of cable. Note for the limited amount of remaining HMW, its MTTF is similar to that of PILC. We use this set of Weibull parameters to ensure reliability projection can be consistent with the ODRM outage record. Updated HMW Weibull parameters are in Table 1. IV. Results: Weibull parameters of UG cable Results of the updated cable model with full set of Weibull parameters has been generated via Weibull ++9 and an Excel based data file. Table 1 lists Weibull parameter for each type of the UG cable. Attachment One provides full set of Weibull probability values for the four types of cable as function of age. Table 1: Weibull parameters of underground cable Cable type Beta of Weibull Eta of Weibull Mean-time-tofail (year) Initial failure rate (/yr/conductor mile) XLPE E-3 TR-XLPE E-3 PILC E-3 HMW (Note 1) 2.0E-2 Note 1: Based on early operating record of HMW, its generic mean life is set to 22 year. For the remaining operating HMW cable, the mean life derived from Weibull++ is 50.9 year. V. Validate cable failure rate against distribution outage record (ODRM) Independent Verification of Weibull failure rate against outage data allows us to build-up confidence on this analytically generated cable failure model. The validation process can be carried out by using Weibull projection to compare against actual cable failures recorded by the ODRM. As shown in the following table, difference between these two independent approaches is less than 1% (0.75% given below). In addition, this verification has been separately validated by comparing results of the PRISM projection model. 4

126 Workpaper Southern California Edison / 2018 GRC 27 Table 2: Compare Weibull projected cable failure with ODRM cable outage data Projected failures for each type of cable per year TR-XLPE XLPE HMW PILC Total projected cable failure per year (Note 1) 1211 ODRM recorded 2014 cable failures 1202 Compare projected failures with ODRM data 0.75% Note 1: Independent validation by PRISM model is VI. Reasonableness of the new failure model - Comparison between new and old cable Weibull model: Compare the new XLPE cable failure rate against the 2007 developed failure rate at the same cable age, most of the new cable failure rates are lower than that of the old cable model. This new cable model is derived from the recent cable removal record and a better representation of cable deterioration with age. In addition, the higher beta parameter associated with the new cable model allows the two failure rate to crossover at advanced operating age. Use PILC as example, these two failure rates intersect at age 51, then the new cable model shows faster retirement of old cable than that of the 2007 model. This accelerated cable removal at advanced service age matches observation from the latest cable retirement data. In summary, the new cable Weibull function is a better representation of cable aging and improves our projection of distribution system reliability. Table 4: Compare new failure rate ( -2015) against the earlier version of cable failure rate ( ) at selected cable age Year Install XLPE PILC TR-XLPE E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E-01 VII. Conclusions Cable failure derived from SCE cable removal records forms the basis for the development of age dependent failure rate model. This updated cable reliability model gives a better representation of cable removal record and can be used to project cable induced distribution system power outages. Results of this analysis can be summarized as: 1. Age dependent cable failure rate curves are shown in Figure For most categories of UG cable (PILC, XLPE, and TR-XLPE), the MTTF can be derived from the respective Weibull parameters and the results are given in Table 1. 5

127 28 Workpaper Southern California Edison / 2018 GRC 3. Full set of Weibull parameter values as function of cable age is provided in Attachment One. Figure 1: Cable failure rate as function of age Reference: 1. Discussions with industry cable leaders (July 2015): -Joe McAuliffe, Senior product engineer of Southwire Rick Hartlein, NEETRAC Director Nigel Hampton, NEETRAC Principal Investigator Long-life XLPE insulated power cable Neetrac Jicable 07_life expectancy paper,

128 Workpaper Southern California Edison / 2018 GRC 29 Attachment One Cable failure model Weibull parameters for: XLPE, PILC, TR-XLPE, & HMW 7

129 30 Workpaper Southern California Edison / 2018 GRC SCE UG CABLE DATA (July 2015) Selcted Weibull parameters for SCE UG cables PILC HMW XLPE TR-XLPE Eta Beta Mini failure rate 6.00E E E E-03 Max failure rate 9.00E E E E-01 MTTF (Year) Yrs of Installation XLPE Eta 44.6 Beta 5 F(t) 5.67E E E E E E E E E E E E E 03 f(t) 6.00E E E E E E E E E E E E E 03 Lambda 6.00E E E E E E E E E E E E E-03 MTTF 41.0 PILC Eta 52.7 Beta 11.3 F(t) 3.50E E E E E E E E E E E E E 07 f(t) 6.00E E E E E E E E E E E E E 03 Lambda 6.00E E E E E E E E E E E E E-03 MTTF 50.4 TR-XLPE Eta 50 Beta 5.5 F(t) 4.53E E E E E E E E E E E E E 04 f(t) 3.00E E E E E E E E E E E E E 03 Lambda 3.00E E E E E E E E E E E E E-03 MTTF 46 HMW Eta 53.6 Beta 9.7 F(t) 1.69E E E E E E E E E E E E E 06 f(t) 2.00E E E E E E E E E E E E E 02 Lambda 2.00E E E E E E E E E E E E E-02 MTTF

130 Workpaper Southern California Edison / 2018 GRC 31 Selcted Weibull parameters for SCE UG cables Yrs of Installatio XLPE F(t) 3.04E E E E E E E E E E E E E E 02 f(t) 6.00E E E E E E E E E E E E E E 02 Lambda 6.00E E E E E E E E E E E E E E 02 PILC F(t) 3.12E E E E E E E E E E E E E E 04 f(t) 6.00E E E E E E E E E E E E E E 03 Lambda 6.00E E E E E E E E E E E E E E 03 TR XLPE F(t) 9.10E E E E E E E E E E E E E E 02 f(t) 3.00E E E E E E E E E E E E E E 03 Lambda 3.00E E E E E E E E E E E E E E 03 HMW F(t) 2.21E E E E E E E E E E E E E E 03 f(t) 2.00E E E E E E E E E E E E E E 02 Lambda 2.00E E E E E E E E E E E E E E 02 9

131 32 Workpaper Southern California Edison / 2018 GRC Selcted Weibull parameters for SCE UG cables Yrs of Installatio XLPE F(t) 9.29E E E E E E E E E E E E E E E 01 f(t) 1.58E E E E E E E E E E E E E E E 02 Lambda 1.74E E E E E E E E E E E E E E E 02 PILC F(t) 7.88E E E E E E E E E E E E E E E 02 f(t) 6.00E E E E E E E E E E E E E E E 02 Lambda 6.00E E E E E E E E E E E E E E E 02 TR XLPE F(t) 4.04E E E E E E E E E E E E E E E 01 f(t) 7.77E E E E E E E E E E E E E E E 02 Lambda 8.10E E E E E E E E E E E E E E E 02 HMW F(t) 1.84E E E E E E E E E E E E E E E 02 f(t) 2.00E E E E E E E E E E E E E E E 02 Lambda 2.00E E E E E E E E E E E E E E E 02 10

132 Workpaper Southern California Edison / 2018 GRC 33 Selcted Weibull parameters for SCE UG cables Yrs of Installatio XLPE F(t) 5.65E E E E E E E E E E E E E E E 01 f(t) 4.21E E E E E E E E E E E E E E E 03 Lambda 9.69E E E E E E E E E E E E E E E 01 PILC F(t) 9.55E E E E E E E E E E E E E E E 01 f(t) 2.39E E E E E E E E E E E E E E E 02 Lambda 2.64E E E E E E E E E E E E E E E 01 TR XLPE F(t) 3.54E E E E E E E E E E E E E E E 01 f(t) 3.61E E E E E E E E E E E E E E E 02 Lambda 5.58E E E E E E E E E E E E E E E 01 HMW F(t) 1.11E E E E E E E E E E E E E E E 01 f(t) 2.36E E E E E E E E E E E E E E E 02 Lambda 2.66E E E E E E E E E E E E E E E 01 11

133 34 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Impact of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability

134 Workpaper Southern California Edison / 2018 GRC 35 SUBSTATION APPARATUS & STANDARDS GROUP REPORT Impact of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability June 2016 Prepared by: Henry Hou Reviewed by: Sam Chien Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 1

135 36 Workpaper Southern California Edison / 2018 GRC Table of Contents Executive Summary 1 Introduction 1.1 VBA Program 1.2 CYMDIST Circuit Reliability Modeling Tool 2 Cluster Analysis for Distribution Circuits 3 System Reliability Simulation and Forecast Model Development 3.1 Circuit Reliability Simulation Model for Distribution System 3.2 Equipment Age Data 3.3 Base Year Model Model Assumptions 3.4 Base Line Model 3.5 Reliability Improvement Model Infrastructure Replacement Simulations WCR Reliability Improvement Other Than Cable IR 4 System Reliability Simulation and Forecast Results 5 References Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 2

136 Workpaper Southern California Edison / 2018 GRC 37 Executive Summary Southern California Edison has a large amount of distribution equipment that is approaching the end of its useful life. In addition, the amount of old equipment is increasing every year, resulting in more frequent interruptions and longer outage times. Several cost effective mitigation measures (or reliability improvement programs) such as preemptive replacement of underground cables and worst circuit rehabilitation (WCR) program are in place to maintain the system reliability at an acceptable level. To justify these plans with focus on the aging infrastructure and WCR, a predictive reliability simulation model for future system reliability is developed. This study also provides a basis for the request for funding in the upcoming general rate case (2018 GRC). In general, distribution equipment replacement can be done in two ways: reactive or proactive replacement. If SCE relies solely on reactive replacement, it can expect a very large number of in service failures over the next twenty years. Use underground cables as an example, from the age profile of SCE existing cables and the associated failure rates, it can be computed that about 1,000 conductor miles of cable will fail per year. These cables will have to be replaced either when it fails during service (reactively) or in a planned manner (proactively). Reactive replacement results in unplanned service interruptions, worsening reliability, and higher replacement costs (due to overtime wages and fault location costs). In addition, replacing cables only when they fail would eventually lead SCE system into an undesirable situation where the underground cables have more in service failures than can be effectively managed. For these reasons and others (such as safety, work planning and resources alignment), the optimum option for SCE is to have an effective proactive replacement program. The system reliability simulation model predicts the future reliability performance in response to the infrastructure aging and different preemptive replacement strategies along with WCR program. When only equipment aging and reactive replacement are evaluated, at 2034, the model predicts a downgrade of the system reliability into a state where the interruption frequency increases per year and the outage time raises about 35 minutes. To cope with such reliability degradation, various strategies of proactive replacement are simulated in the reliability predictive model. Based on the results, it is recommended that SCE implement a proactive replacement of distribution mainline cables at a rate of 350 conductor miles per year in the future years. Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 3

137 38 Workpaper Southern California Edison / 2018 GRC 1. Introduction Southern California Edison (SCE) has a large amount of distribution equipment that is approaching the end of its useful life. The amount of old equipment is increasing every year, resulting in higher customer interruptions and longer outage times. If no action is taken, age related failures will continue to aggregate and eventually cause system reliability to dramatically worsen. Several mitigation measures including preemptive replacement of distribution equipment, such as underground cables, and WCR program, are evaluated to identify a cost effective way to maintain system reliability at an acceptable level. To address the aging impacts, SCE has adopted a dynamic circuit modeling and time series reliability simulation methodology to study the impact of aging equipment, mainly underground cable, and other WCR practice to the system reliability and to investigate the amount of proactive cable replacement (a.k.a. infrastructure replacement, IR) that can cost effectively balance the long term cable replacement expenditure and the system reliability trend. The study has been continuously conducted to support the past general rate cases. Most existing reliability studies focus on snapshot reliability and do not evaluate the dynamic effects of the equipment aging. This is a proper approach when assessing short term or non age related system reliability, since the impact due to increases of the equipment failure rate over times is not considered. Such studies typically assume that component failure rates are constant, and do not change from year to year. However, the current project specifically aims to evaluate the impact of aging infrastructure over the years on system reliability. The approach is thus unlike the traditional reliability studies, the model keeps changing over the time. As the equipment ages, the failure rate increases, and the system reliability gets worse. SCE has distribution circuits. It is neither possible nor necessary to include each circuit in the reliability model. A statistical clustering method employs the representative circuits, instead of using every circuit, in the model. In the clustering approach, circuits are grouped by their reliability similarities and the representative circuits (or cluster circuits) are identified. Each cluster circuit with a proper weight represents the reliability performance of the entire cluster. In this study, 20 representative circuits are identified from the cluster analysis. In 2009, SCE started a pilot reliability project to predict the reliability trends of the distribution system in response to equipment aging over a 20 year period. Quanta Technology was hired as a consultant to develop a reliability simulation model. The model contains two major parts. The first part involves a visual basic application program (VBA Program) to perform the equipment database updates (due to aging and replacement) and then mapping them to the age dependent equipment failure rate data. The second part utilizes the commercial software CYMDIST (Ref.1) to build the distribution circuit reliability model and calculating the circuit reliability indices. Running the entire model is an iterative process between the VBA program and CYMDIST software. Because of schedule and budget constraints, the project had only been able to achieve to a point where the quantification is a manual process and has to be done in two separate steps (one for VBA and CYMDIST each). In 2012, as part of the Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 4

138 Workpaper Southern California Edison / 2018 GRC 39 preparation of 2015 GRC filing, SCE began its efforts to validate the reliability simulation model and renovate the model quantification into an automated process. The in house project involves Updated cluster analysis based on 2011 operational data, New CYMDIST version allowing input and output data communicating with Excel VBA program A modified interface control between the VBA program and CYMDIST In 2015, SCE continues improving the methodology to enhance various technical details so that the study can closely reflect SCE s field practice. The project is a joint effort task between Nexant consultants and SCE engineers. Nexant is responsible to enhance mainly in VBA code with a focus on two major technical areas: (1) how the annual budget of cable IR is allocated to the candidate circuits, and (2) the development of the cable IR package for each representative circuit to accommodate all possible budgets. While SCE performs: (i) cluster analysis to identify a new set of cluster circuits from most updated circuit characteristic data, (ii) circuit reliability model for each cluster representative circuit, and (iii) WCR work scoping (other than cable IR) for reliability improvement. All the SCE works serve as inputs to Nexant s enhanced system simulation and forecast model. As a summary, the model with the enhancement is able to simulate SCE distribution system in the following 5 major categories: 1. Equipment aging 2. Equipment replacement because of in service failures 3. Equipment replacement due to infrastructure replacement 4. WCR implementation 5. Interactions of WCR and equipment replacement Further discussions of the VBA program and the CYMSDIST and their roles in the simulation model are provided in the following sections. 1.1 VBA Program The reliability simulation model uses Visual Basic for Applications (VBA) programming language in Excel in conjunction with the CYMDIST Reliability Assessment Module (RAM) model to perform various simulations for future years. The VBA program works extensively with RAM through Component Object Module (COM) to update the equipment failure rates and the circuit configurations to reflect equipment aging, replacement, and WCR work scope in the circuit reliability models. For equipment aging and replacement simulations, VBA directs RAM to update the age dependent equipment failure rates in the Access Database of RAM model based on the new age profiles for every modeled year. The Access Database contains the equipment failure rate data which links to an age dependent equipment failure rate file. The VBA keep tracks of the equipment type and age in the database. When the model ages one year, the age of the equipment is updated by the VBA program. While the equipment is replaced, either reactively or preemptively, the VBA resets the age to zero. In this way, for any given year, an updated equipment age Access database will be used in CYMDIST RAM.. Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 5

139 40 Workpaper Southern California Edison / 2018 GRC The preemptive equipment replacement is performed to reflect current SCE practice of the infrastructure replacement program. Previously, the annual budget of proactive replacement is allocated to all the circuits in a selected cluster, with each circuit getting a portion of the budget. Depending on the number of circuits in the cluster, each circuit may only get a limited replacement budget for a year and get another amount of replacement budget for the following year. However, in practice, if a WCR circuit is selected for reliability improvement, engineers tend to perform most, if not all, of the identified projects on the circuit at once and move on to other circuits instead of coming back to visit the same circuit over again. In the modified algorithm, the annual cable IR budget is allocated to a certain number of circuits in which each circuit would have all the identified equipment replaced. The enhanced simulation is closer to the current Field Engineer practice. Moreover, in selecting the worst performing circuits, the VBA program commands RAM in CYMDIST, through COM, to calculate circuit reliability indices, which later become the basis for determining the equipment preemptive replacement and calculating the reliability indices at the system level. Also through COM of CYME, VBA program modifies circuit reliability model according to identified WCR work scope other than cable IR. The updated circuit models serve as the inputs of WCR simulations. 1.2 CYMDIST Circuit Reliability Modeling Tool CYMDIST is typically thought of as a distribution power flow calculation tool, but it also has a state of the art RAM that has been tested and validated for accuracy. In the RAM, circuit connectivity can be developed by importing the SCE Geographic Information System (GIS) data and provided the foundation of the study. The RAM is able to simulate the reliability experiences, in terms of power interruption frequency and duration, of each customer based on the connectivity models, operational data, and equipment failure data. These customer experiences can then be aggregated into reliability indices such as System Average Interruption Duration Index (SAIDI) and System Average Interruption Frequency Index (SAIFI), the two most commonly used measures of reliability performance in the power industry. In a typical circuit reliability model, calculation of reliability indices requires the following elements: Likelihood of fault occurrence such as equipment failure rate Circuit responses to faults including protective devices and automatic features Affected customers given that faults occurred Duration parameters for interruptions such as traveling, detection, isolation, switching, and repairing during power restoration It should be pointed out that, based on SCE outage experiences, underground cable is one of the major contributors to affect SCE system reliability. In addition, the cable failure rate is age dependent, older cables has higher failure rates. Moreover, the maintenance and inspection of UG cables are much more time consuming and costly than others. Therefore, more attentions should be placed onto cables and hence the current simulation model stays focus on the impact from underground cable and the way to mitigate it. Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 6

140 Workpaper Southern California Edison / 2018 GRC 41 In the study, simulations and predictions of reliability of the distribution system for the next 20 years are conducted. To develop the model, first, CYMDIST is used to build the connectivity model, or reliability simulation model, for each of the 20 cluster representative circuits. Then, the circuit reliability models are run for 20 times separately, one year each time, with random failure simulation to calculate the reliability indices for the 20 year period. This is to establish a baseline case. During the baseline run, each year RAM interfaces with the VBA program to export the calculated SAIDI and SAIFI results and then import the updated equipment age database, as a result of aging and reactive replacement simulations, for the run of next year. After the baseline circuit reliability model quantification, the model can be run for different IR strategies. The details will be provided in Section 3. The remainder of the report depicts cluster analysis, model development, base year model, base line model, reliability improvement model, and system reliability prediction. 2. Cluster Analysis for Distribution Circuits SCE s distribution system has more than four thousand circuits. If any study requires to analyze every single circuit, it would be extremely costly and time consuming. Instead of working on each individual circuit, the reliability model adopts an alternate approach, cluster analysis, which uses a set of representative circuits to approximate the reliability performance of SCE system. The cluster analysis statistically assesses the circuit characteristic data (such as customer count, circuit age, sectionalizing devices, conductor length, etc.), groups similar circuits together and then selects one representative circuit from each group or cluster. In the same cluster, circuits are considered similar to each other; it is thus expected that the studies performed on the selected circuit can represent the results of the same study on the entire circuits in the same cluster. Therefore, the results on cluster circuits can be used to represent the results on their associated group of circuits and further enable the extrapolations of study to the whole system. This method has been effectively used in T&D impact and Integrated Resource Planning (IRP) studies over the past twenty years (Ref. 2). Selection of the set of representative circuits can be achieved by using a statistical clustering method. This clustering method groups circuits that are similar into the same cluster but dissimilar into other clusters. The circuit that is the closest to the common or averaged circuit characteristics of each cluster is selected as the representative circuit of the cluster (or cluster center circuit). Due to the diversity of equipment types, circuit configurations, voltage levels, and the geographic and demographic variations of customer base, a distribution circuit has various features. Every feature is not necessarily relevant in all studies, making the selection of circuit features is very specific to each study. Step wise regression is one standard procedure for selecting the best subset of features (Ref. 3), and is applied in this study for circuit feature selection. The details of the cluster analysis pertains to this study can be found in Section 2 of 2009 Quanta study report (Ref. 4). In current reliability model validation, customer Minutes of interruption (CMI), same as Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 7

141 42 Workpaper Southern California Edison / 2018 GRC 2010 approach, is used as the dependent variable and other circuit related data as independent variables (as shown in Table 2.1) in the cluster analysis to identify the representative circuits for reliability simulation model. The correlations factor (or Coefficient of determination), R 2, between each independent variable and CMI also listed in Table 2.1. In the current study, the cluster analysis includes 4429 distribution circuits and identifies 20 clusters based on the similarity of the variables listed in Table 2.1. Table 2.2 is the list of the representative circuit as the result of the cluster analysis. In the table, it is noted that cluster weight indicates the number of circuits in each cluster group. Variable Designator Average CMI ( ) R 2 V9 Average Customer Count ( ) V10 AR Count V17 Total Trans Count V25 CIC Cable (miles) V28 Fault Indicators V14 Total RCS V20 Tie Count V29 Fused Equipment V19 UG Trans KVA V16 Total Trans KVA V21 UG 3 Phase Miles V11 Circuit Age V15 All Weather Zone (ODRM) V12 Total OH Manual Switch V13 UG Trans Count V24 OH 3 Phase Miles V18 Total Miles Total R Table 2.1 Cluster Analysis Variables and R 2 Values Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 8

142 Workpaper Southern California Edison / 2018 GRC 43 Cluster # Cluster Center Circuit Name Cluster Weight 1 ALTURA BAUXITE 55 3 BIG RIGG BONANZA 67 5 CASTLEROCK CONCOURSE CORSAIR 29 8 DELFORD DITCH LLOYD MAGOO NELSON OPPORTUNITY OTIS REDSKIN SIZZLER SPRAGUE TALC TARGET THORNBURG 203 Table 2.2 Representative Circuits and weights for 20 Clusters The identified cluster circuits are used extensively in the simulation model and the extrapolated results represent the features of the entire distribution system, it would be prudent to validate the representativeness of the selected circuits before applying them in the simulation model. Table 2.3 provides a comparison of actual values at system level and extrapolated values from the cluster circuits. The comparisons between the actual data and extrapolated value of different circuit attributes show that the representative model approach can represent the system very well: majority of the circuit attributes extrapolated values are within 8 9 % difference of the actual data. For example, underground miles has 6% of difference for all circuits. Even though some of the circuit attributes are not included in the cluster analysis to determine the representative circuit, they are still well represented by the cluster (see UG 3 Phase Miles and OH3 Phase Miles) Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 9

143 44 Workpaper Southern California Edison / 2018 GRC Circuit Features Actual Total Extrapolated Total Difference (%) Average CI ( ) 21,574 21,306-1% Average CMI ( ) 527,006, ,579,377-7% Customer Count 4,858,675 4,865,432 0% AR Count 1, % Circuit Age 91,810 94,897 3% Total Manual Switch 28,360 26,005-8% UG Trans Count 263, ,019-1% Total RCS 7,233 6,624-8% All Weather Zone 11,713 12,098 3% Total Trans KVA 48,398,140 52,774,384 9% Total Trans Count 712, ,267 4% Total Miles 66,820 61,907-7% UG Trans KVA 35,046,998 36,856,853 5% Tie Count 21,626 25,615 18% UG 3 Phase Miles 17,286 17,430 1% OH Miles 36,155 32,990-9% UG Miles 30,665 28,916-6% OH 3 Phase Miles 23,786 22,803-4% CIC Miles 10,947 10,508-4% Fault Indicators 13,184 13,098-1% Fuses 36,213 35,224-3% Table 2.3 Comparison of System Actual Data and Extrapolation Results from Cluster Circuits Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 10

144 Workpaper Southern California Edison / 2018 GRC 45 3 System Reliability Simulation and Forecast Model Development This section describes the details of development of the reliability simulation model for the representative circuits of SCE distribution system. Also provided are the procedure and assumptions used for various model development phases including connectivity model, base year model, base line model and reliability improvement model. 3.1 Circuit Reliability Simulation Model for Distribution System From the Cluster analysis, 20 cluster circuits have been identified and are used to represent the reliability performance of SCE distribution system. Since each circuit represents an averaged measure of entire cluster, the reliability associated activities (e.g., interruptions) modeled in each representative circuit can be extrapolated to a cluster or group level. By summing up all those circuit level power interruptions of 20 cluster groups, the system level reliability can be obtained. In evaluating circuit reliability, CYMDIST is used as a tool. CYMDIST is widely used throughout the power industry for various analysis and calculations such as load flow analysis, fault analysis, load allocation, load balancing, and capacitor placement calculation. Among many features, RAM is specifically designed to support distribution engineers in assessing the reliability of electric distribution networks. The module is equipped with the capability to simulate the circuit features in response to faults and SCE practices in power restoration to the affected customers. The simulations with equipment failure rate and restoration timing data enable RAM to compute reliability indices such as SAIFI and SAIDI for overall system or local networks.. For each of the 20 circuits, CYMDIST built a basic connectivity model, which includes the distribution system configuration, protective and switching devices, customers, and loads. A manual importing process, based on image maps, was used since these feeders were not modeled in any software that could be directly imported into CYMDIST. Circuit Connectivity is imported using the CYMDIST Mapping module which has the ability to retrieve GIS maps from SCE emaps system. Figure shows a section of GIS map as an illustrative example. Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 11

145 46 Workpaper Southern California Edison / 2018 GRC Figure GIS Map for a Sample Circuit The retrieved maps can then be used as a background for tracing specific features such as wires, cables, and distribution equipment. The background map is generally scaled so wires and cables that are traced on top of the map are of the correct length as well as geographically correct. It should be noted that there is one exception in the scale mapping system, i.e., the structures layout shown in emap is not in the same scale as is designed for of the conductor routes and their connections. This is because if the original scale would follow, the detailed information of the equipment or components in the structure would have been too small to be visible. To address this scaling issue, after the maps imported into CYMDIST software, a manual trace effort of all the cable lengths, using FIM information, was performed to ensure that the correct length data is used in the connectivity model. As a result, the total cable length of the 20 cluster circuits is reduced by about 20%. After circuit model with three and single phase primary was built, the correct wire type was assigned, Figure shows a sample of the circuit model corresponding to same section in Figure In the Figure, the circuit lines traced over the GIS background are shown in red. Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 12

146 Workpaper Southern California Edison / 2018 GRC 47 Figure Sample of CYMDIST Circuit Trace The distribution transformers and protective and switching devices were also included in the model. Each distribution transformer was modeled as a spot load with the appropriate capacity and specific load information assigned to the load allocation, as well as the customer count. From the reliability point of view, location and number of customers of each spot load are crucial for assessing the impacts during a power interruption. When a fault occurs in a distribution circuit, it would cause the immediate upstream protective device to activate. In general, the protection device can be a branch line fuse or substation circuit breaker depending on specific fault location and circuit feature. With the fuse or circuit breaker triggered, power to the spot loads or customers downstream of the device is interrupted. Then, event mitigations including automations or manual actions are taken to further reduce number of impacted customers. At last, the faulty equipment is either replaced or repaired to restore power back to those last customers. By applying proper equipment failure rates date and time parameters, RAM is able to simulate the power interruption activities in a reliability circuit model from fault occurrence, circuit response, mitigation, and corrective actions (e.g., detections, trouble shooting, power switching). Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 13

147 48 Workpaper Southern California Edison / 2018 GRC The load transformers are shown as small red arrows in the following diagram Figure Spot Load Spot Load Figure Distribution Transformers or Spot Loads (red arrows) in a Sample Circuit 3.2 Equipment Age Data The aging impact of underground cables is the major focus of this study. Therefore, the age related information for the cables was essential in the circuit model. SCE s FIMs contain the installation year of the cable segments and they were used directly in the circuit model. Figure shows a section of FIM map as an illustrative purpose. The age assignment is a manual process too. Each of the developed circuit models was compared with FIM maps to find the age of the underground cables. During the course, it is discovered that the definition of a cable segment is different between CYMDIST and FIM. In FIM, a cable segment is defined from a structure to another structure, while CYMDIST is on a node to node basis. In CYMDIST, a node is a connecting point to a single reliability significant equipment such as cable, switch, fuse, recloser, breaker, etc... Therefore, a cable segment in CYMDIST may contain several FIM structure to structure cable segments, which may create an inconsistent age problem in the CYMDIST cable segment. Besides Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 14

148 Workpaper Southern California Edison / 2018 GRC 49 the inconsistent age issue, there are few cases where the age information is not available in the FIM map, thus, several assumptions are made: 1. Multiple installation years were found for the same cable segment in CYMDIST connectivity model. For age selection, the age of the majority of associated FIM segments is assigned. O Otherwise, the oldest age neighboring segments is applied. 2. When no neighbor cable age is available, the FIM structure built year is assumed. 3. Averaged age of cables in the circuit model is used if no other assumptions can be applied. Figure Example of FIM Map After having the age data, the connectivity model is ready to run for base year case that reflects the system reliability performance at current time and will be discussed more in the next section. Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 15

149 50 Workpaper Southern California Edison / 2018 GRC 3.3 Base Year Model With the equipment data linking to circuit connectivity model built in CYMDIST RAM, the reliability performance of SCE distribution system and the impact of implementing different reliability improvement strategies can then be evaluated. As discussed previously, quantification of the reliability simulation model contains two phases: circuit level and system level. It starts with calculation of the circuit SAIFIs and SAIDIs in CYMDIST RAM and later extrapolating the circuit reliability indices with their weights to obtain the system reliability. The foundation of entire process is the base year model which (1) represents the circuit configurations, design, practice, and performance of SCE distribution system at current year, and (2) serves as the basic model for subsequent simulations of equipment aging, replacement, and circuit reliability improvement for the future years. CYME connectivity model is used directly to establish the circuit reliability base year model for 20 cluster circuits. The CYMDIST circuit reliability model includes all major distribution equipment (i.e., cables, OH conductors, transformers etc.), protective and switching devices along with customer distribution. Based the equipment ages (or age dependent failure rates) at the current year, RAM calculates the reliability indices SAIFI and SAIDI for the base year reliability model. In a circuit model, RAM evaluate all possible faults in the circuit and take the associated failure rates of equipment with affected customers to determine circuit SAIFI first. Then it further brings in interruption times of affected customers into calculation of circuit SAIDI. The computation of circuit SAIFI and SAIDI can be expressed by It should be pointed out that the location of protective devices in a circuit plays an important role to determine the size of affected customers after a fault. To effectively calculate the interruption frequency and outage duration for impacted customers, RAM divides a circuit into zones by protective device (e.g., breaker, recloser, or fuse). For any fault occurring in the same zone, the impacted customers are all the same. Therefore, to calculated circuit SAIFI, instead of by each fault, RAM takes the total fault occurrence frequency and affected customers by zones. Further, within a zone, if there are sectionalizing devices (such as switch) for fault isolations and power restorations, RAM further separates the zone into sub zones to accurately account for various downtimes that customers may experience. The algorithm to calculate circuit SAIDI in RAM simulates the event sequences after faults occur, which includes all the possible post event activities such as traveling, inspection, testing, detection, isolation, diagnosis, restoration, and repair. In other words, connectivity and features of a circuit define the fault impact zone, isolation means, and restoration paths for the modeled faults and hence affect the way to calculate the reliability indices. Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 16

150 Workpaper Southern California Edison / 2018 GRC 51 To determine circuit performance, the calculation requires 3 major elements. They are (1) equipment failure rate, (2) number of customer affected, and (3) restoration time. Among the three, equipment failure rate is most complicated to develop since it is based on an extensive historical outage data derived from equipment failure events. To evaluate the cable aging impact, SCE has developed failure rate models of underground cables (Ref.5). Figure shows the failure rates for different types of underground cables. The presented failure rate, also called hazard function, is the likelihood of a component failing if it has not already failed, representing how the failure rate of component changed as it ages. Even though newer cable types (i.e., XLPE and TR XLPE) have longer expected lifespan and less steep failure rate increase at high age comparing to older cable types, the failure rates of older cable are still significantly higher than younger cable. The failure rates for other equipment included in the simulation modeled will be discussed in the model assumptions section. 8.0E 01 Cable Failure Rate Cable failure/mile/yr 7.0E E E E E E 01 XLPE PILC TR XLPE 1.0E E Cable Age Figure Underground Cable Failure Rates Once the circuit level indices have been determined, it is assumed that, because of the similarity of circuit features, the experience of specific cluster center circuit applies to every circuit in the same cluster. Later, by totaling all the reliability contributions from all the clusters, system reliability indices can be calculated by the following expressions: Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 17

151 52 Workpaper Southern California Edison / 2018 GRC And The number of customers included in a circuit cluster, or represented by a cluster circuit, is listed in Table Cluster Circuit ALTURA BAUXITE BIG_RIGG BONANZA CASTLEROCK CONCOURSE CORSAIR DELFORD DITCH LLOYD MAGOO NELSON OPPORTUNITY OTIS REDSKIN Cluster Group Customer 243, , ,128 37,204 65, ,080 56, , ,632 57, , , , , ,748 Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 18

152 Workpaper Southern California Edison / 2018 GRC SIZZLER 17 SPRAGUE 18 TALC 19 TARGET 20 THORNBURG Table List of Circuit Cluster Customer Counts 512, ,348 42, , , Model Assumptions In the reliability simulation model, RAM parameters in CYMDIST are selected to closely reflect the SCE actual practice and operations during fault conditions. The practice or operational responses along with circuit designs and features (e.g., protection, detection, response, isolation, switching, restoration, and repair) are the RAM parameters used in the simulation model to accommodate versatile modeling approaches. With the parameters applied properly, the reliability model would be able to simulate the circuit responses to power interruptions in the distribution system. It would be extremely time consuming if all the possible and unique features in the complex distribution networks are analyzed. As an optimum approach, the acceptable industrial and reliability practice and SCE experiences suggest the use of the following assumptions or parameters: Fuse clearing scheme was used to model over current protection practices. For reclosing devices, it was assumed that single phase trip and all phases lock out during momentary and permanent faults. When transferring load to adjacent feeders during contingency, the load transfer tries to maximize the number of restored customer. Momentary events have duration less than 5 minutes. The switching time for automated devices is 30 seconds and the switching time for remotely controlled devices is 3 minutes. The failure rate of underground cable is determined by the cable failure rate model (Ref. 5). The calibration aims to match the system level SAIFI and SAIDI instead of circuit level reliability performance. Thus, a uniform overhead failure rate and an averaged repair time (MTTR) are applied for all representative circuits. Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 19

153 54 Workpaper Southern California Edison / 2018 GRC The number of customers served by SCE is 4,858,675. It was assumed that single phase, two phase and three phase faults represent 70%, 10%, and 20% of all permanent faults, respectively. The same is also assumed for momentary faults. All feeders are allowed to reclose. For overhead lines, failure rates of momentary faults are 2 times that of permanent faults. This is a standard assumption in reliability modeling. Failure rate of overhead lines ise greater than 0.03 failure / year / mile. For underground cables, failure rates of momentary faults are assumed to be zero. All protective devices have a manual operation time ranging from 30 minutes to 1 hour. MTTR of overhead lines are greater than 2 hours. Underground cables have a MTTR of 3 hours. The contribution of protective and switching devices (e.g., Fuses, Reclosers, and Switches), unless explicitly modeled, to the circuit reliability is reflected by the line failures. i.e., the failure rates of protective devices are lumped into the failure rate of the overhead line or underground cable it belongs to. Once the circuit reliability models (or base year model) are established, a quick examine is to check the model results (system SAIFI and SAIDI) that are in the reasonable range of the SCE system reliability historical data, i.e., the average value of historical SAIFI and SAIDI indices from the most recent 3 years (data from June 2012 to May 2015). Using averaged value of historical reliability indices as the target for determining equipment contribution is a common industry practice that smooths down random reliability data variations that occur in the system. In the analysis, the three year average SAIFI and SAIDI values, as the historical target data to be matched with the reliability model base year results, are provided in Table Historical SAIFI (outage/customer/year) Historical SAIDI (hour/customer/year) Distribution Circuit (System) Table Historical Reliability Indices Based on 3 Year Data (from June, 2012 to May, 2015) To stay focus on our objective of the study, which is to predict the underground cable contribution due to aging in the future 20 years, segregation of the reliability contributions other than underground cables is necessary. The non cable portions of the distribution system contain either the equipment and devices that are easier to inspect and maintain and most of times are replaced reactively; or other Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 20

154 Workpaper Southern California Edison / 2018 GRC 55 causes (e.g., vehicle hit pole, animal, Mylar balloon.., etc.). All these contributors in the reliability simulation model are assumed to remain as constants allowing only the underground portion changes in the future years. Another reason in segregation is, in reality, to match the system SAIFI and SAIDI directly from the model simulation run is very difficult since many parameters (such as failure rates and durations) used in the simulations and each parameter impacts the model result through its own unique scheme of algorithm. Therefore, an alternate approach using a breakdown of system reliability contribution as the matching targets is applied. The system reliability can be classified into 3 categories by major contributors: underground cables, overhead conductors, and transformers and others. Based on SAIFI/SAIDI cause breakdowns of 2014 data, the 3 major components contribute approximately 50% (overhead), 35% (underground), and 15% (transformer and others) of system reliability. For the dominant contributor, overhead conductor, the base year model is run exclusively by using typical failure rates and MTTR values obtained from industry and manufacturers data, as well as peer reviewed publications are used as initial values. Then, the failure rates and MTTR are iteratively adjusted until the extrapolated system level SAIDI and SAIFI values match the targets, i.e. SAIFI and SAIDI As a result, the overhead failure rates (sustained and momentary faults) and the MTTR are listed in Table Sustained Failure Rate (failure/year/mile) Momentary Failure Rate (failure/year/mile) MTTR (hour) All Circuit (System) Table Overhead Line Failure Rates and MTTR For distribution transformers, the reliability modeling of transformers contains two components: localized impact and propagated impact. The localized scenario simulates failures of a transformer causing power interruption to the customers it serves and thus the impact is limited to its downstream customers. However, a review of SCE distribution outage data shows that for some events, transformers fail in such way that activate the immediate upstream protection device and therefore the impact is not limited to its local customers. From the review, it also provides the ratio of localized versus propagated, which is 85% localized and 15% propagated. Combine with SCE averaged transformer failure rate data (Ref. 6), the localized failure rate and propagated failure rate can be derived, as shown in Table For the others category, the model includes miscellaneous equipment, e.g., old underground oil switches, which is to be replaced for safety concerns, and is deemed to have a small contribution on system reliability. Transformer Failure Mode Type Failure Rate (failure/year) MTTR (hour) Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 21

155 56 Workpaper Southern California Edison / 2018 GRC localized OH & UG 3.27E OH 7.05E 04 2 Propagated UG 3.60E 04 3 Table Distribution Transformer Failure Rates and MTTRs 3.4 Base Line Model The base year model is a snapshot of the model at current time, which is the year that simulation of cable aging and equipment replacement begins. To assess future system reliability impact by aging and different equipment replacement strategies, a base line model is required and served as a basis for comparison for different reliability improvement options, what if no implementation of reliability improvements It should be kept in mind that the equipment replacement in the model has two different reasons: (1) reactive replacement and (2) preemptive replacement. In the base line model, the replacement is limited to equipment fail in service (or random failures). In the reliability base line simulation model, equipment random failure is determined by Monte Carlo random sampling approach. The Monte Carlo approach provides a simulation of random choice mechanism for cable failures based on their failure likelihood. To illustrate the technique, the algorithm picks every cable segment in the circuits modeled and evaluates its failure likelihood according to its failure rate. For each segment, a random number between 0 and 1 is generated. This specific number determines if the cable segment fails or not. For example, an 800ft cable segment has a failure rate of 0.1 /yr/mi, which means this cable segment has 1.5% chance of failure in the year. The random numbers generated in an even distribution between 0 and 1 also have 1.5% chance of being smaller than or equal to Therefore, if the cable associated random number from Monte Carlo generator is smaller than or equal to 0.015, the cable fails; otherwise, the cable survives. As the simulation adopts a random number generator to sample the equipment failure, there is an uncertainty in the resulted reliability indices, which can vary from simulation to simulation. Different from many other stochastics studies in which an average result from multiple scenarios is sufficient, this study needs an actual circuit model with a set of equipment replacement defined dynamically, such that various reliability improvement options can be simulated for the following years. In order to minimize the impact of randomness in the base line study, in each simulation year, multiple runs of equipment failure sampling are performed, and the round of equipment failure sampling that leads to an average reliability performance will be adopted to generate the base line model for that given year. The accordant list of random device failures is then saved for use in the infrastructure replacement simulations so that the random failure patterns are not deviated due to proactive cable replacement scheme. Once a cable failure occurs, this cable segment will be replaced during the year. Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 22

156 Workpaper Southern California Edison / 2018 GRC 57 For each replacement, an average length of conductor miles of cable is used to reflect the current cable replacement experience data. Also, a new cable type and a new installation year are assigned to the segment in the circuit model. The VBA program in the reliability simulation model acts as the role of performing the equipment aging and replacement functions and linking to CYMDIST to form a loop process as given in Figure Starting with the base year, the simulation model first uses CYMDIST to obtain circuit reliabilities and system reliability. Then, the VBA program performs the equipment aging and equipment replacement tasks through updating the Access database for CYMIDT RAM calculations. The equipment aging simulation is achieved by re mapping the equipment new age for a given year and linking the circuit equipment to their failure rates based on the new age. The equipment replacement due to random failures is implemented by replacing the year of installation and equipment type for the failed cable segments in the equipment database. When the Access equipment database is updated with new ages and newly replaced items, the model is ready for the next year. By looping the process 20 times, once every year, the system reliability indices for the years 2015 through 2034 can be obtained. Run CYMDIST for Circuit SAIFI & SAIDI System SAIDI and SAIFI calculation Equipment Aging & Replacement Update Access Database of CYME for next year Run Loop Runs for 20 Years Figure Illustration of Reliability model quantification process The graphic representations of the Base Line model result are provided in Figures and From these two figures, it is clear to see both SAIDI and SAIFI are getting worse if no preemptive replacement program is in place. System SAIFI increases from 0.91 in 2015 to 1.42 in 2034, with an increase of 55.9%; system SAIDI increases from 104 minutes in 2015 to 139 minutes in 2034, with an increase of 33.8%. Both numbers are significant, especially given the fact that the increases are due to device aging only and failed devices are still being replaced every year Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 23

157 58 Workpaper Southern California Edison / 2018 GRC System SAIDI Trend for Base Line Case SAIDI (Min) Figure SAID Trend from 2012 to 2032 for Base Line Case 1.5 System SAIFI Trend for Base Line Case SAIFI Figure SAIFI Trend from 2012 to 2032 for Base Line Case 3.5 Reliability Improvement Model In order to provide a better service, SCE is interested in exploring the options of proactively replacing aged underground cables (i.e. cable IR). Replacing old but still in service equipment or components will prevent future in service failures and hence improve system reliability. But, proactive replacement has Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 24

158 Workpaper Southern California Edison / 2018 GRC 59 its cost, which requires an earlier investment with the interest. Plus, in reality, the IR program always has a funding constraint, therefore, it is very important to follow the optimum strategy for infrastructure replacement. The current cable IR strategy is implemented on the worst circuits that contribute significantly to the increases of either SAIDI or SAIFI at system and circuit levels. Therefore, the underground cable IR is considered as part of WCR program. The cable IR strategy is to replace the most failure prone cables with the worst impact in the worst performing circuits based on a funding allocation guideline as below 33% to worst circuit SAIFI circuits 17% to worst circuit SAIDI circuits 33% to worst system SAIFI circuits 17% to worst system SAIDI circuits In the study, the guideline is followed in allocating the amount of cables to be replaced in the identified worst circuits. The candidates for replacement are those mainline cables with the oldest age in the circuit model since they have the highest failure tendency and the worst impact. Every year, the circuits with the worst reliability performance will be selected as the WCR circuits for next year s reliability improvement projects. Based on the guideline, half of the WCR funding is allocated to worst individuals (worst circuit SAIFI and worst circuit SAIDI) and the other half is allocated to the circuits that have biggest impact to system reliability (worst system SAIFI and worst system SAIDI). On the other hand, two thirds of the WCR funding is allocated to the worst SAIFI circuits and the remaining one third is allocated to the worst SAIDI circuits. If one circuit is selected as WCR circuit by multiple criteria, it will get the aggregated budget from these categories. For example, if a circuit has the worst circuit SAIFI and the worst circuit SAIDI for a given year, then it will get 50% of the annual funding for the WCR program for the following year. During the model quantifications, the worst circuit list might change over times. Some previous worst circuits are not in the worst circuit group anymore because their reliability has been improved (e.g., their old cables were replaced with new ones). In contrast, some previously regular circuits might become worst circuits when their reliability gets worse because of aging. The simulation model captures all these changes and dynamically identifies the candidates for worst circuits from year to year Infrastructure Replacement Simulations This IR simulation is the key part of the study and also the part that undergoes the most improvement from previous versions of the methodology. To properly emulate cable proactive replacement in the simulation model, the algorithm should be able to reflect the field practice. It is typical that SCE engineers tend to replace all the old cable on the identified circuits at once, instead of replacing the oldest cable and come back for the next oldest cable segment another year (which was the simplified scheme used in previous versions of the methodology); similarly, engineers implement all the cost effective reliability improvement options to the WCR circuits being worked on instead of implementing only the top selective options this year and come back for Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 25

159 60 Workpaper Southern California Edison / 2018 GRC another round of improvement later. For every year within the simulated 20 year horizon, WCR candidates are first selected based on previous year s reliability indices, following the selection criteria described earlier. For the circuits that are not selected for WCR program, only reactive replacement scheme is simulated; while for the representative circuits selected for WCR program, reactive replacement is first performed and proactive replacement is then simulated. WCR program budget allocation is mainly determined by the underground cable replacement, although there are other work scopes for circuit improvement. Thus, in the simulation, the annual budget for proactive replacement program is expressed as the miles of underground cable to be replaced (e.g., 500 miles of proactive cable replacement per year), instead of dollars. Depending on the IR strategy for proactive cable replacement, the number of circuits that can undergo proactive replacement program varies. The set of mainline cable segments to be proactively replaced has been pre determined by SCE engineers based the circuit map of each representative circuit. The amount of cable replacement needed for a WCR circuit with a specific IR strategy is determined from the pre selected set and the aging criteria (i.e., 34 years). With a given budget, the number of circuits in a cluster can be identified, which may be only a portion of the cluster or the entire cluster. If the selected cluster has more circuits than the budget can support, the rest of the cluster will be subject to proactive cable replacement when the cluster is selected as WCR again. In some cases, if a small cluster is selected as WCR circuits or a large WCR cluster only has few circuits left for proactive cable replacement, there will be additional budget available after all the necessary work is performed, then the next worst circuit with the same criterion (e.g., worst system SAIDI) will be selected to use the available funding. Following is an example to illustrate this process in more details. The IR strategy or annual budget for proactive cable replacement is 500 miles. In a particular year, circuit ALTURA is selected as a WCR circuit because it has the worst circuit SAIFI in previous year; according to the budget allocation guidelines, it gets 33% of the annual budget, which is 165 miles. According to the pre defined replacement package, circuit ALTURA has 3 miles of cable segments subject to proactive cable replacement; assuming only 2.5 miles of cables are older than the age criterion, 34 years, for the given simulation year, then 66 circuits in the ALTURA cluster can perform proactive cable replacement. However, the cluster that ALTURA represents has 635 circuits; the remaining 569 circuits will not undergo proactive cable replacement for the given simulation year. For the next few years, ALTURA is not selected as the WCR circuit because the reliability indices of ALTURA gets better, when compared with other circuits, due to the reactive and proactive programs. Thus, only reactive program (replaced upon failure) is performed on the cluster during these years. However, after 5 years, the cluster that ALTURA represents is selected as WCR again because of its circuit SAIFI performance, and then 165 miles of cable can be proactively replaced (33% of the budget). By this year, all the pre defined cable segments for the untouched 569 circuits are older than 34 years; therefore, each circuit has 3 miles of cable to be replaced and the annual budget can support another 55 circuits for proactive cable replacement. In these two rounds of proactive cable replacement, 121 circuits have undergone the program and there are still 514 circuits to be worked on in the future once the cluster is selected as WCR again. Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 26

160 Workpaper Southern California Edison / 2018 GRC 61 In the example above, circuits in the same cluster are broken into different groups because different circuits in the same cluster experience proactive replacement at different times. Circuits in different group causes a change of their cable age profile and thus different amounts of cables are replaced, which then affects the interactions between IR and random failures in a different way. In the simulation model, the circuits with cable IR at a particular year form a unique group that has its own 20 years of reliability indices simulated. Therefore, bookkeeping of all these groups and the associated results is a critical task to ensure the accuracy of the simulation model. In the reliability simulation, when cable IR budget increases, the old cable would be replaced more, and the system cable age profile would be younger. As a result, with more new cables in place, the chance of random failures would decrease. This results in a more reliable system. On the other hand, if the IR amount is not sufficient enough to cope with cable aging effects, the cable age profile gets less improved and the old cables remain at high failure rates. In such a case, clearly the system performance would be going down WCR Reliability Improvement Other Than Cable IR In addition to the aging related equipment replacement, i.e., cable IR program, other reliability improvement options such as adding fuse and fault indicator, adding sectionalizing devices and implementing distribution automation are also very cost effective. When a circuit is selected as WCR circuit, SCE engineers consider all the possible options, age related and non age related, to improve its reliability. As previously discussed, engineers tend to implement most of the cost effective reliability improvement projects, within the available budget, on a selected WCR circuit at once to avoid repetitive (planned) power interruptions to customers in the near future. To reflect this practice in the study, a list of possible reliability improvement options for each representative circuit is developed beforehand. These options are selected through extensive reliability simulations. Ideally, the simulation of non age dependent projects shall consider the available WCR budget, similar to the simulation of proactive cable replacement. However, the interactions of non age related projects and proactive replacement program will significantly increase the number of possible outcomes or scenarios. For example, a set of circuit selected for proactively cable replacement may not have sufficient budget to have non age related projects implemented on all of them at the same year, some circuits may not be able to have WCR program at all and some circuits may only have budget of partial WCR program. Simulating the WCR program in details will inevitably introduce tremendous complication in bookkeeping the various scenarios and significantly increase the simulation time due to the number of possible combinations. As the focus of this study is the impact of aging underground cable to the system reliability, the effect of non age dependent WCR program is evaluated with a simplified approach, which assumes, if a set of circuits are selected and have adequate budget for cable IR, then all the identified non age related program options will be implemented. This is because, in general, the associated cost is relatively low when compared to cable IR expenditures. By doing this, it greatly simplify the simulations and no additional bookkeeping is necessary in the model. In other words, the non age related projects will be implemented when the circuit is first selected for proactive Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 27

161 62 Workpaper Southern California Edison / 2018 GRC cable replacement; if the circuit is selected again at later years, since the non age dependent improvement options have been done, no more work is required except cable IR. 4. System Reliability Simulation and Forecast Results This study examines various infrastructure replacement (IR) strategies in order to evaluate the reliability benefits of different levels of proactive replacement program and make decisions for the best program. The simulated IR scenarios include 200 mile IR program (i.e., the annual budget for proactive cable replacement is 200 miles), 350 mile IR program, 500 mile IR program and 600 mile IR program. Historically, SCE has been conducting infrastructure replacement program on 207 miles every year in average, based on cable replacement data from 2009 and The average length of cable segments replaced upon random failures is 338 miles per year based on the same 5 year period. While the simulation results for 200 mile IR program from the reliability model, the average amount of IR is 198 miles per year and the average amount of cable replacement because of in service failures is 349 miles per year. The results from simulation model show a consistency with the actual historical replacement data, indicating the simulation model is capable to represent SCE distribution system and to reflect the actual cable replacement practice in both areas. Among all the simulated IR scenarios, the 600 mile IR program is the one in which all the representative circuits are selected (because of sufficient budget) for both cable IR and non age WCR programs. This scenario thus can be used to evaluate the holistic benefit of the non age related WCR projects on SCE distribution system. The cost needed for implementing the entire WCR package on one representative circuit is calculated based on the identified project list and the typical costs for various project options. The total cost of the WCR program for the entire cluster (or the applicable portion of the cluster) is extrapolated using the number of circuits having the WCR program implemented. As listed Table 4.1, the WCR cost needed for all the applicable circuits in the 600 mile IR scenario is million dollars. Representative Circuit / Cluster ALTURA BAUXITE BIG RIGG BONANZA CASTLEROCK CONCOURSE CORSAIR DELFORD DITCH LLOYD MAGOO NELSON OPPORTUNITY OTIS Total WCR Cost of each Cluster $46,990,000 $1,760,000 $50,568,000 $34,170,000 $17,472,000 $41,600,000 $17,922,000 $22,110,000 $27,318,000 $13,668,000 $49,920,000 $16,168,000 $25,840,000 $67,326,000 Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 28

162 Workpaper Southern California Edison / 2018 GRC 63 REDSKIN $47,360,000 SIZZLER $50,600,000 SPRAGUE $39,990,000 TALC $17,490,000 TARGET $17,368,000 THORNBURG $99,876,000 TOTAL $705,516,000 Table 4.1 WCR Program Cost for Clusters Currently, SCE s WCR program spends an average of 13.3 million dollars on 87 circuits every year. Assuming the budget remains the same and there is no inflation for the next 20 years, the total WCR program budget available is million dollars, significantly less than the budget needed for the WCR program simulated in the study (705.5 million dollars). Apparently, the benefits simulated from the WCR program are much higher than actual during the modeled 20 year period and need to be adjusted to reflect the current practice properly. However, before any adjustments made to the simulated WCR benefits, a sanity check is performed to ensure that the simulation process aligns closely with the actual practice. As SCE s non cable IR WCR program covers 87 circuits every year on average and there are 4429 circuits in the system, it would take about 51 years from now to complete all the circuits. On the other hand, the total simulated WCR budget is million dollars, which can support about 53 years of WCR program given that the annual WCR budget is 13.3 million dollars. The numbers of years needed for a completed round of WCR program are consistent between the simulation and actual estimation. The heavy WCR (non cable IR) simulations in the early years would improve system reliability significantly earlier than it should be and the improvement will slow down in the later years. However, as the WCR projects address the non aging reliability improvement (such as fusing and automation), even though the interaction of these projects and cable aging still exits, it is considered to have a very limited impact to the long term system reliability performance, whether the projects are implemented earlier or later. As a result, the aggregated impact of all the WCR projects simulated in the 20 year period can be adjusted by actual budget and spread out across the entire 20 year period to estimate the annual reliability improvement from the WCR program. The full benefits (based on M) of the simulated WCR program over the 20 years can reduce the system SAIFI by 0.31 and reduce the system SAIDI by 37 minutes. When adjusted based on the actual WCR budgets available over the 20 simulation years (266.9 M), the system SAIFI can be reduced by and the system SAIDI can be reduced by 14 minutes. These aggregated WCR benefits will be evenly distributed to different years and added to the benefits gained from aging cable IR program. In other words, the WCR program can reduce the system SAIFI by every year and reduce the system SAIDI by 0.7 minute every year. Figure s 4.1 and 4.2 show the trends of system SAIFI and SAIDI over the 20 simulation years under various IR strategies, respectively. The 0 IR scenario shows the base line scenario, which represent the system reliability performance with only the cable reactive replacement program (with 0 mile of cable Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 29

163 64 Workpaper Southern California Edison / 2018 GRC IR). In other IR scenarios including 200 IR, 350 IR, 500 IR, and 600 IR, the benefit of full WCR projects (both cable IR and non age related WCR improvement) is included. With only reactive cable replacement, 0 IR case, the average age of system component continues to increase and the system reliability worsens over the year. With the implementation of IR and WCR programs, the population of aged underground cable reaching its expected lifespan, is suppressed to certain extent, depending on the level of IR strategy. As a result, the system reliability shows a slower rate of deterioration. Depending on the circuit configuration and cable age distribution, some proactive replacement package is more effective than others. This uncertainty results in the different reliability impact from year to year, as shown in the figures that the SAIFI/SAIDI trends are not increasing smoothly; instead, some dips can be observed (e.g., the period between 2021 and 2023 for the 200 IR scenario, the period between 2015 and 2018 for the 500 IR scenario) when the cluster with effective cable replacement package is selected for infrastructure replacement. 1.5 System SAIFI Forecast for Interruptions/Custonmer/Year IR 200 IR 350 IR 500 IR 600 IR Year Figure 4.1 System SAIFI Forecast for Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 30

164 Workpaper Southern California Edison / 2018 GRC 65 System SAIDI Forecast for Minutes/Customer/Year IR 200 IR 350 IR 500 IR 600 IR Figure 4.2 System SAIDI Forecast for References Year 1. CYMDIST, Distribution System Analysis, CYME International. 2. A Cluster Based Method of Building Representative Models of Distribution Systems, H. L. Willis H. N. Tram, and R. W. Powell, IEEE Transactions on Power Apparatus and Systems, March 1983, p Probability & Statistics for Engineers & Scientists, Ronald E. Walpole, Raymond H. Myers, Sharon L. Myers, Prentice Hall College Div; Impact of Aging Infrastructure on System Reliability at SCE, Prepared for Southern California Edison, Le Xu, Quanta Technology, June 28, Substation Apparatus & Standards Group Report: Underground Cable Reliability Model Update. S.H.Chien, May, from Sam Chien, Distribution Transformer Failure Rate, October 7, Impacts of Infrastructure Aging and a Program of Worst Circuit Rehabilitation on Future System Reliability 31

165 66 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Cost of Cable Testing

166 Workpaper Southern California Edison / 2018 GRC 67 Work paper Title: COST OF CABLE TESTING WBS Element: CET PD IR PC Witness: Jose Ramon Goizueta This workpaper establishes the average unit cost of testing a mile of cable Total Count Cable Life Extension CLE Testing Expenditures 2015 $ 1& Total Cable Life Extension $ 7,862,851 $ 12,768,100 $ 6,851,823 $ 27,482, ' Average Cable Life Extension $ 26,277 $ 41,297 $ 45,624 $ 36,229 Unit Cost Used for Forecasting 4& Cable Life Extension $ 46,000 $ 46,869 $ 47,843 $ 49,048 $ 50,587 $ 52,196 1 Annual expenditures and unit counts based on projects that occurred from 2011 to Annual expenditures and units counts based on projects that have been closed (construction completed, finalized accounting and financials) 3 Unit cost per year is derived by taking total expenditures and dividing by total unit counts 4 Unit Cost used for forecasting includes escalation 5 Last year recorded was chosen because of increasing cost over the last five years 6 Unit counts are in conductor miles

167 68 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Cost of Cable Injection

168 Workpaper Southern California Edison / 2018 GRC 69 Work paper Title: COST OF CABLE INJECTION WBS Element: CET PD IR PC Witness: Jose Ramon Goizueta This workpaper establishes the average unit cost of injecting a mile of cable Total Count Cable Life Extension CLE Injection Expenditures 2015 $ 1& Total Cable Life Extension $ 80,161 $ 474,162 $ 4,812,990 $ 5,367, ' Average Cable Life Extension $ 84,922 $ 71,165 $ 133,172 $ 122,687 Unit Cost Used for Forecasting 4& Cable Life Extension $ 133,000 $ 135,512 $ 138,330 $ 141,814 $ 146,262 $ 150,913 1 Annual expenditures and unit counts based on projects that occurred from 2011 to Annual expenditures and units counts based on projects that have been closed (construction completed, finalized accounting and financials) 3 Unit cost per year is derived by taking total expenditures and dividing by total unit counts 4 Unit Cost used for forecasting includes escalation 5 Last year recorded was chosen because of increasing cost over the last five years 6 Unit counts are in conductor miles

169 70 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Mainline Cable Testing Analysis

170 Workpaper Southern California Edison / 2018 GRC 71 Assumptions List of Model Assumptions: Item Input Value Source Cost to test 1 mile of Main Line cable based on analysis by Seema Abraham - Included 1 Main Line - Cost to Test 98,000 in model 2 Main Line - Cost to Replace Cost to replace 1 conductor mile of Main Line cable, calculated using 2015 projects 245,526 that closed to plant that had cable replacement only in order to isolate costs. 3 PVRR - Depreciable Cable Life 40 Years GRC approved depreciable life 4 PVRR - Net Salvage Value 15% PD GRC value 5 PVRR - Fed Depreciation 15 Yr DB GRC Approved 6 PVRR - CA Depreciation 30 Yr DDB GRC Approved 7 Main Line - Industry Failure Rate 59% Mainline Cable Study - IMCORP 8 Main Line - Factored Failure Rate The failure rate of cable less the percentages of failures cuased by non-cable 40% conditions 9 CIC - SCE Failure Rate 32% Failure rate realized over three year period Analysis Notes Column Name Calculation Description The model uses 2016 as the base year. All Present Value (PV) calculations are in 2016 A Year $'s. B Year The model looks at 30 years. After 30 years the PV values are de minimus. C Replacement Cost (Nominal) Cost to replace 1 conductor mile of Main Line cable escalated at approved escalation rates for distribution capital. D PVRR of Replacement Calculates the Present Value Revenue Requirement (PVRR) of replacing 1 mile of Main Line conduit using the assumptions in the table above. E CY Testing PVRR This is the PVRR calculation of testing 1 conductor mile of Main Line cable. This assumes cable testing for Main Line cable is capitalizable as is the treatment of CIC testing. The value is constant as the model evaluates the testing and of 1 conductor mile in the current year. F PVRR Discount to 2016 Replacement This column calculates the difference between the current year (2016) replacement PVRR and the future year PVRR. G Option 1 - Replace All (No Testing) This column calculates the 2016 PVRR cost of replacing the target quantity (350 miles) of conductor. It multiplies the target quantity by the PVRR to replace 1 conductor mile of Main Line cable. H Test + Replace This column calculates the cost to test the target quantity of cable (350 miles) and replace that which fails to pass inspection in The 2016 value multiplies the target quantity by the cost to test. Subsequent years (1-29) calculate PVRR of replacing the passed cable at that year in the future. I J K Option 2 (Scenario A) Test & Replace - 40% Failure Rate Option 2 (Scenario B) Test & Replace - 60% Failure Rate B/C Ratio This column calculates the cost to test the target quantity of cable (350 miles) and replace that which fails to pass inspection in 2016 plus the PVRR cost to replace the passed cable in a future year. The calculation is used to determine how long the tested and "passed" cable would have to last past that time in order to justify the cost of testing. Scenario A suggests a 40% failure rate of the population of tested cable. This column calculates the cost to test the target quantity of cable (350 miles) and replace that which fails to pass inspection in 2016 plus the PVRR cost to replace the passed cable in a future year. The calculation is used to determine how long the tested and "passed" cable would have to last past that time in order to justify the cost of testing. Scenario A suggests a 59% failure rate of the population of tested cable. This column calculates the Benefits to Cost Ratio of the Replace All Strategy to the Test and Replace Strategy to determine the number of years "passed" cable must remain in service to provide positive value. WPSCE-02V08-Mainline Cable Testing Analysis Workpaper.xlsx

171 72 Workpaper Southern California Edison / 2018 GRC Mainline Testing Analysis Cost Per Conductor Mile for Replacement versus Current Year Test & Pass Chart Description: The analysis shows the cost of replacing 350 miles of conductor in year 0 vs. the Test and Replace strategy. T&R includes testing 350 miles and replacing the bad cable in the Year 0 with the replacement of the additional passed cable in Year X. Cost to Test 2016: $ 98,000 Cost to Replace 2016: $ 245,526 Annual Miles of Conductor 350 Industry Failure Rate: 40% A B C D E F G H I J K Option 2 (Scenario B) Test & PVRR of Option 1 - PVRR Discount to Replace All (No Option 2 (Scenario A) Test & Replace - 40% Replace - 60% Failure Year Year Replacement Cost (Nominal) Replacement CY Testing PVRR 2016 Replacement Testing) Test + Replace Failure Rate Rate B/C Ratio $ $ $ $ - $ 109,574 $ 87,565 $ 87, , $ $ $ $ 21.6 $ 109,574 $ 61,199 $ 148, , $ $ $ $ 41.2 $ 109,574 $ 57,087 $ 144, , $ $ $ $ 60.7 $ 109,574 $ 52,989 $ 140, , $ $ $ $ 79.0 $ 109,574 $ 49,163 $ 136, , $ $ $ $ 96.1 $ 109,574 $ 45,565 $ 133, , $ $ $ $ $ 109,574 $ 42,180 $ 129, , $ $ $ $ $ 109,574 $ 38,959 $ 126, , $ $ $ $ $ 109,574 $ 36,173 $ 123, , $ $ $ $ $ 109,574 $ 33,586 $ 121, , $ $ $ $ $ 109,574 $ 31,184 $ 118, , $ $ $ $ $ 109,574 $ 28,954 $ 116, , $ $ $ $ $ 109,574 $ 26,883 $ 114, , $ $ $ $ $ 109,574 $ 24,960 $ 112, , $ $ $ $ $ 109,574 $ 23,175 $ 110, , $ $ $ $ $ 109,574 $ 21,518 $ 109, , $ $ 95.1 $ $ $ 109,574 $ 19,979 $ 107, , $ $ 88.3 $ $ $ 109,574 $ 18,550 $ 106, , $ $ 82.0 $ $ $ 109,574 $ 17,223 $ 104, , $ $ 76.1 $ $ $ 109,574 $ 15,991 $ 103, , $ $ 70.7 $ $ $ 109,574 $ 14,848 $ 102, , $ $ 65.6 $ $ $ 109,574 $ 13,786 $ 101, , $ $ 61.0 $ $ $ 109,574 $ 12,800 $ 100, , $ $ 56.6 $ $ $ 109,574 $ 11,884 $ 99, , $ $ 52.5 $ $ $ 109,574 $ 11,034 $ 98, , $ $ 48.8 $ $ $ 109,574 $ 10,245 $ 97, , $ $ 45.3 $ $ $ 109,574 $ 9,512 $ 97, , $ $ 42.1 $ $ $ 109,574 $ 8,832 $ 96, , $ $ 39.0 $ $ $ 109,574 $ 8,200 $ 95, , $ $ 36.3 $ $ $ 109,574 $ 7,614 $ 95, , $160,000 Present Value of Revenue Requirement Analysis $150,000 $140,000 $130,000 ($000'S) $120,000 $110,000 $100,000 $90,000 $80, Years after Testing Option 1 - Replace All (No Testing) Option 2 (Scenario B) Test & Replace - 60% Failure Rate Option 2 (Scenario A) Test & Replace - 40% Failure Rate WPSCE-02V08-Mainline Cable Testing Analysis Workpaper.xlsx

172 Workpaper Southern California Edison / 2018 GRC 73 Mainline Testing Analysis Assumptions Annual Miles (A) 350 Scenario 1 Failure Rate (B) 40% Scenario 2 Failure Rate (C) 60% Cost to Replace (D) $ 245,526 Cost to Test (E) $ 98,000 Option 1 Replace All without Testing Annual Miles 350 A Total Cost 85,934,150 A x D Option 2 Test & Replace Bad Now and Good in Future Scenario A: 40% Failure Scenario B: 60% Failure Miles Tested Bad (F) 140 A x B =F 210 A x C =F Cost to Test $ 13,720,000 E x F $ 20,580,000 E x F Cost to Replace 34,373,660 D x F 51,560,490 D x F Total Cost of Bad $ 48,093,660 $ 72,140,490 Miles Tested Good (G) 210 A - F = G 140 A - F = G Cost To Test $ 20,580,000 E x G $ 13,720,000 E x G Total First Year Cost $ 68,673,660 $ 85,860,490 Future Replace Cost $ 51,560,490 D x G $ 34,373,660 D x G Total Cost $ 120,234,150 $ 120,234,150 First Year Cost Comparison Option 1 Total cost $ 85,934,150 $ 85,934,150 Option 2 First Year Cost 68,673,660 85,860,490 Variance $ (17,260,490) $ (73,660) WPSCE-02V08-Mainline Cable Testing Analysis Workpaper.xlsx

173 74 Workpaper Southern California Edison / 2018 GRC 60 Percental fail rate Scenario Cost Per Conductor Mile for Replacement versus Current Year Test & Pass Chart Description: The analysis shows the cost of replacing 350 miles of conductor in year 0 vs. the Test and Replace strategy. T&R includes testing 350 miles and replacing the bad cable in the Year 0 with the replacement of the additional passed cable in Year X. Cost to Test 2016: $ 98,000 Cost to Replace 2016: $ 245,526 Annual Miles of Conductor 350 Industry Failure Rate: 60% A B C D E F G H I J K PVRR of PVRR Discount to Option 1 - Replace All Test + Option 2 Test & Replace Failed Year Year Replacement Cost (Nominal) Replacement CY Testing PVRR 2016 Replacement (No Testing) Replace CY/Passed FY B/C Ratio $ $ $ $ - $ 109,574 $ 109,480 $ 109, $ $ $ $ 21.6 $ 109,574 $ 40,799 $ 150, $ $ $ $ 41.2 $ 109,574 $ 38,058 $ 147, $ $ $ $ 60.7 $ 109,574 $ 35,326 $ 144, $ $ $ $ 79.0 $ 109,574 $ 32,775 $ 142, $ $ $ $ 96.1 $ 109,574 $ 30,377 $ 139, $ $ $ $ $ 109,574 $ 28,120 $ 137, $ $ $ $ $ 109,574 $ 25,973 $ 135, $ $ $ $ $ 109,574 $ 24,115 $ 133, $ $ $ $ $ 109,574 $ 22,391 $ 131, $ $ $ $ $ 109,574 $ 20,789 $ 130, $ $ $ $ $ 109,574 $ 19,302 $ 128, $ $ $ $ $ 109,574 $ 17,922 $ 127, $ $ $ $ $ 109,574 $ 16,640 $ 126, $ $ $ $ $ 109,574 $ 15,450 $ 124, $ $ $ $ $ 109,574 $ 14,345 $ 123, $ $ 95.1 $ $ $ 109,574 $ 13,319 $ 122, $ $ 88.3 $ $ $ 109,574 $ 12,367 $ 121, $ $ 82.0 $ $ $ 109,574 $ 11,482 $ 120, $ $ 76.1 $ $ $ 109,574 $ 10,661 $ 120, $ $ 70.7 $ $ $ 109,574 $ 9,898 $ 119, $ $ 65.6 $ $ $ 109,574 $ 9,190 $ 118, $ $ 61.0 $ $ $ 109,574 $ 8,533 $ 118, $ $ 56.6 $ $ $ 109,574 $ 7,923 $ 117, $ $ 52.5 $ $ $ 109,574 $ 7,356 $ 116, $ $ 48.8 $ $ $ 109,574 $ 6,830 $ 116, $ $ 45.3 $ $ $ 109,574 $ 6,342 $ 115, $ $ 42.1 $ $ $ 109,574 $ 5,888 $ 115, $ $ 39.0 $ $ $ 109,574 $ 5,467 $ 114, $ $ 36.3 $ $ $ 109,574 $ 5,076 $ 114, WPSCE-02V08-Mainline Cable Testing Analysis Workpaper.xlsx

174 Workpaper Southern California Edison / 2018 GRC 75 Rev Req Table Rev Req for Capital Investment in Future Year **$100K Base** Year PVRR Discount Factor 2016 $ % 2017 $ % 2018 $ % 2019 $ % 2020 $ % 2021 $ % 2022 $ % 2023 $ % 2024 $ % 2025 $ % 2026 $ % 2027 $ % 2028 $ % 2029 $ % 2030 $ % 2031 $ % 2032 $ % 2033 $ % 2034 $ % 2035 $ % 2036 $ % 2037 $ % 2038 $ % 2039 $ % 2040 $ % 2041 $ % 2042 $ % 2043 $ 9.7 8% 2044 $ 8.8 7% 2045 $ 8.0 6% 2046 $ 7.3 6% 2047 $ 6.6 5% 2048 $ 6.0 5% 2049 $ 5.5 4% 2050 $ 5.0 4% 2051 $ 4.5 4% 2052 $ 4.1 3% 2053 $ 3.7 3% 2054 $ 3.4 3% 2055 $ 3.1 2% 2056 $ 2.8 2% WPSCE-02V08-Mainline Cable Testing Analysis Workpaper.xlsx

175 76 Workpaper Southern California Edison / 2018 GRC Escalation Rates Year Year Capital Esc. Rates O&M Esc. Rates Cap Esc Factor O&M Esc Factor % 2.49% % 2.56% % 2.46% % 2.28% % 2.26% % 2.21% % 2.20% % 2.22% % 2.22% % 2.22% % 2.22% % 2.22% % 2.22% % 2.22% % 2.22% % 2.22% % 2.22% % 2.22% % 2.22% % 2.22% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% % 2.23% WPSCE-02V08-Mainline Cable Testing Analysis Workpaper.xlsx

176 Workpaper Southern California Edison / 2018 GRC 77 Mainline Cost Projection Mainline Testing Cost Estimate Detail Activity Actual Spend 2014 Spend per cond mile Escalated to 2016 $ Notes Testing Mainline Traffic control $ 72, $ $ (1) $ 5,000 Vendor $ 2,719, $ 8, $ 9, (2) $ 18,493 Contractor Support $ 7,824, $ 25, $ 26, (3) $ 39,908 District Support $ 207, $ $ (4) $ 14,008 O/H & Chargebacks $ 1,530, $ 4, $ 5, (5) $ 12,187 T&D Support $ 839, $ 2, $ 2, (6) $ 3,502 A/P Accruals $ (425,093.70) $ (1,374.91) $ (1,445.44) Permitting $ 5,000 Unit Cost per cond mile $ 41,000 $ 43,000 $ 98,000 Assumptions Conductor miles tested in Acctual Spend based on 2014 data (1) It is expected that mainline testing would require a specific traffic control plan for both pre-fielding and testing.. Average of $2500 per traffic control plan (2) Vendor support is expected to be higher to minimize customer impact. Expect to use an additional MDU with crew.. (3) Additonal contractor crew requirements to support additional MDU (4) Dedicated switching crew and vehicle support per day (5) Labor and Vehicle O/H at 87% of labor cost (6) T&D Support at 25% of labor (7) Amount escalated at approved capital Distribution rates ( % and %) WPSCE-02V08-Mainline Cable Testing Analysis Workpaper.xlsx

177 78 Workpaper Southern California Edison / 2018 GRC Mainline Cable Unit Cost This workpaper establishes the average unit cost of replacing mainline cable Mainline Cable Replacement - Counts 1, 2, & Total Count Mainline Cable Mainline Cable Replacement Expenditures $ 1& Total Mainline Cable $ 4,763,648 $ 5,118,328 $ 3,743,663 $ 3,852,402 $ 2,726,968 $ 20,205,009 Mainline Cable Replacement - Unit Costs ' Average Mainline Cable $ 155,419 $ 146,050 $ 177,439 $ 224,009 $ 245,526 $ 175,546 1 Annual expenditures and unit counts based on projects that occurred from 2011 to Annual expenditures and units counts based on projects that have been closed (construction completed, finalized accounting and financials) 3 Unit cost analysis uses a one year weighted average by taking total expenditures and dividing by total unit counts for Last year recorded was chosen because of a general upward trend in unit cost over the last five years. 4 Unit counts are in conductor miles WPSCE-02V08-Mainline Cable Testing Analysis Workpaper.xlsx

178 Workpaper Southern California Edison / 2018 GRC 79 Workpaper Title: Average Age of WCR

179 80 Workpaper Southern California Edison / 2018 GRC Sample of Cable Segments Scoped in 2016 WCR Projects "Batch 1" submission (approximately 20% of scoping year) Details based on project specific engineer scoping notes Average Year 1977 Average Age (2016) 39 Conductor Miles 78 Circuit Name City From Structure To Structure Install Year Circuit Length (Feet) Conductor Length (miles) Albatross 16kV Santa Monica E V Albatross 16kV Santa Monica V E Albatross 16kV Santa Monica E V5XXXX Albatross 16kV Santa Monica V5XXXX E Almansor 16KV ALHAMBRA V M Almansor 16KV ALHAMBRA M M Almansor 16KV ALHAMBRA M V Almansor 16KV ALHAMBRA M M Almansor 16KV ALHAMBRA M V Almansor 16KV ALHAMBRA V E Almansor 16KV ALHAMBRA V M Almansor 16KV ALHAMBRA M V Almansor 16KV ALHAMBRA V V Almansor 16KV ALHAMBRA V V Almansor 16KV ALHAMBRA V V Almansor 16KV ALHAMBRA V V Almansor 16KV ALHAMBRA M V Almansor 16KV ALHAMBRA V M Almansor 16KV ALHAMBRA M V Almansor 16KV ALHAMBRA M V Almansor 16KV ALHAMBRA M E Almansor 16KV ALHAMBRA V M Almansor 16KV ALHAMBRA M E Almansor 16KV ALHAMBRA V M Almansor 16KV ALHAMBRA M M Almansor 16KV ALHAMBRA M V Almansor 16KV ALHAMBRA V M Almansor 16KV ALHAMBRA V V Almansor 16KV ALHAMBRA V E Almansor 16KV ALHAMBRA V M Aspen 16kV Torrance S E Aventura DESERT HOT SPRINGS V M Aventura DESERT HOT SPRINGS M E Benedict 4kV Beverly Hills V V Benedict 4kV Beverly Hills V X Benedict 4kV Beverly Hills X V Benedict 4kV Beverly Hills V V Benedict 4kV Beverly Hills V V Benedict 4kV Beverly Hills V V BILLINGS 12KV COSTA MESA V M BILLINGS 12KV COSTA MESA V X BILLINGS 12KV COSTA MESA X B BILLINGS 12KV COSTA MESA B P BILLINGS 12KV COSTA MESA X9920 B BILLINGS 12KV COSTA MESA B P BILLINGS 12KV COSTA MESA B B BILLINGS 12KV COSTA MESA B P BILLINGS 12KV COSTA MESA B B BILLINGS 12KV COSTA MESA B P BILLINGS 12KV COSTA MESA B V

180 Workpaper Southern California Edison / 2018 GRC 81 BILLINGS 12KV COSTA MESA B B BILLINGS 12KV COSTA MESA B P BILLINGS 12KV COSTA MESA B P BILLINGS 12KV COSTA MESA P X BILLINGS 12KV COSTA MESA X9920 B BILLINGS 12KV COSTA MESA B P BILLINGS 12KV COSTA MESA B B BILLINGS 12KV COSTA MESA B P BILLINGS 12KV COSTA MESA V B BILLINGS 12KV COSTA MESA B P BILLINGS 12KV COSTA MESA V B BILLINGS 12KV COSTA MESA B P BILLINGS 12KV COSTA MESA B P BILLINGS 12KV IRVINE S B BILLINGS 12KV IRVINE B S BILLINGS 12KV IRVINE RACK M BILLINGS 12KV IRVINE M M BILLINGS 12KV IRVINE M M BILLINGS 12KV IRVINE M V BILLINGS 12KV IRVINE V V BILLINGS 12KV IRVINE V S BILLINGS 12KV IRVINE S E BILLINGS 12KV IRVINE V S BILLINGS 12KV COSTA MESA P P Blend DESERT HOT SPRINGS E M Blend DESERT HOT SPRINGS M V Blend DESERT HOT SPRINGS 1909CWT V Blend DESERT HOT SPRINGS V V Blend DESERT HOT SPRINGS E V Blend DESERT HOT SPRINGS V E Blend DESERT HOT SPRINGS V M Blend LANCASTER M S Bryant 16kV Torrance M M Bryant 16kV Torrance M X Bryant 16kV Torrance X E Chimay 12KV CORONA V M Clarinet Valencia V V Clarinet Valencia V E Clifford 12kV Downey E V Clifford 12kV Downey M9464Y V Clifford 12kV Downey V V Clifford 12kV Downey V V Clifford 12kV Downey V M Clifford 12kV Downey M V Clifford 12kV Downey V X Clifford 12kV Downey X V Clifford 12kV Downey V E Clifford 12kV Downey V V Commonwealth 4KV ALHAMBRA M M Commonwealth 4KV ALHAMBRA M V Commonwealth 4KV ALHAMBRA V E Commonwealth 4KV ALHAMBRA V M Commonwealth 4KV ALHAMBRA M V Commonwealth 4KV ALHAMBRA V M Commonwealth 4KV ALHAMBRA V V Commonwealth 4KV ALHAMBRA V V Commonwealth 4KV ALHAMBRA V V Commonwealth 4KV ALHAMBRA V V Cresta 16KV ALHAMBRA V M Cresta 16KV ALHAMBRA M M Elroy 16kV Torrance E M

181 82 Workpaper Southern California Edison / 2018 GRC Elroy 16kV Torrance E M Elroy 16kV Torrance M V Elroy 16kV Torrance V M Elroy 16kV Torrance M E Elroy 16kV Torrance S S Elroy 16kV Torrance E S Elroy 16kV Torrance M M Elroy 16kV Torrance M M Elroy 16kV Torrance M M Elroy 16kV Torrance M M Elroy 16kV Torrance M M Elroy 16kV Torrance M E Elroy 16kV Torrance S S Elroy 16kV Torrance X S Elroy 16kV Torrance S S Elroy 16kV Torrance S S Elroy 16kV Torrance S S Elroy 16kV Torrance S X Elroy 16kV Torrance S E Fuerte 12KV WALNUT E M Fuerte 12KV WALNUT M V Fuerte 12KV WALNUT V S Fuerte 12KV WALNUT V S Fuerte 12KV WALNUT S S Fuerte 12KV WALNUT S S Fuerte 12KV WALNUT S S Fuerte 12KV WALNUT S V Fuerte 12KV WALNUT V S Fuerte 12KV WALNUT S V Fuerte 12KV WALNUT S S Fuerte 12KV WALNUT E V Fuerte 12KV WALNUT V S Fuerte 12KV WALNUT S V Fuerte 12KV WALNUT V M Fuerte 12KV WALNUT S M Fuerte 12KV WALNUT M V Fuerte 12KV WALNUT M V General Petroleum 16 Torrance E M General Petroleum 16 Torrance M M General Petroleum 16 Torrance M M General Petroleum 16 Torrance M X General Petroleum 16 Torrance X E Hilton 16kV Beverly Hills E M Hilton 16kV Beverly Hills M V Hilton 16kV Beverly Hills V S Hilton 16kV Beverly Hills S V Hilton 16kV Beverly Hills V V Hilton 16kV Beverly Hills V X Hilton 16kV Beverly Hills X V Hilton 16kV Beverly Hills V V Hilton 16kV Beverly Hills V V Hilton 16kV Beverly Hills V V Hilton 16kV Beverly Hills V V JOAQUIN 12KV IRVINE M M JOAQUIN 12KV IRVINE M M JOAQUIN 12KV IRVINE M V JOAQUIN 12KV IRVINE V V JOAQUIN 12KV IRVINE V S JOAQUIN 12KV IRVINE S E JOAQUIN 12KV IRVINE V S JUSTICE 12KV SAN BERNARDINO M M

182 Workpaper Southern California Edison / 2018 GRC 83 JUSTICE 12KV SAN BERNARDINO V M JUSTICE 12KV SAN BERNARDINO M M JUSTICE 12KV SAN BERNARDINO M V JUSTICE 12KV SAN BERNARDINO V V JUSTICE 12KV SAN BERNARDINO V V JUSTICE 12KV SAN BERNARDINO V M JUSTICE 12KV SAN BERNARDINO M V JUSTICE 12KV SAN BERNARDINO V M JUSTICE 12KV SAN BERNARDINO M V JUSTICE 12KV SAN BERNARDINO M V JUSTICE 12KV SAN BERNARDINO V V JUSTICE 12KV SAN BERNARDINO V M JUSTICE 12KV SAN BERNARDINO M M JUSTICE 12KV SAN BERNARDINO M V JUSTICE 12KV SAN BERNARDINO M V JUSTICE 12KV SAN BERNARDINO V M JUSTICE 12KV SAN BERNARDINO M V JUSTICE 12KV SAN BERNARDINO V M JUSTICE 12KV SAN BERNARDINO M M JUSTICE 12KV SAN BERNARDINO M V JUSTICE 12KV SAN BERNARDINO V M JUSTICE 12KV SAN BERNARDINO M E JUSTICE 12KV SAN BERNARDINO V V JUSTICE 12KV SAN BERNARDINO V V JUSTICE 12KV SAN BERNARDINO V V JUSTICE 12KV SAN BERNARDINO V V JUSTICE 12KV SAN BERNARDINO V V JUSTICE 12KV SAN BERNARDINO V V MEADOWBROOK 12KVSAN BERNARDINO M M MEADOWBROOK 12KVSAN BERNARDINO V M MEADOWBROOK 12KVSAN BERNARDINO V V Millennium 12KV WALNUT V E Millennium 12KV WALNUT E V Millennium 12KV WALNUT V S Millennium 12KV WALNUT S V Millennium 12KV WALNUT V V Millennium 12KV WALNUT V X Millennium 12KV WALNUT X E Millennium 12KV WALNUT V S Millennium 12KV WALNUT S E Millennium 12KV WALNUT Y S Millennium 12KV WALNUT S V Millennium 12KV WALNUT V E Millennium 12KV WALNUT E M Millennium 12KV WALNUT M V Millennium 12KV WALNUT E V Millennium 12KV WALNUT V E Millennium 12KV WALNUT V V Millennium 12KV WALNUT V V Millennium 12KV WALNUT V V Millennium 12KV WALNUT V V Millennium 12KV WALNUT V E Millennium 12KV WALNUT E V Morse 12kv IRVINE V M Morse 12kv IRVINE M V Morse 12kv IRVINE V S Morse 12kv IRVINE S V Morse 12kv IRVINE V S Morse 12kv IRVINE S V Morse 12kv IRVINE V V Morse 12kv IRVINE V S

183 84 Workpaper Southern California Edison / 2018 GRC Morse 12kv IRVINE S V Morse 12kv IRVINE V B Morse 12kv IRVINE B B Morse 12kv IRVINE B B Morse 12kv IRVINE B B Morse 12kv IRVINE V M Morse 12kv IRVINE M M Morse 12kv IRVINE M V Morse 12kv IRVINE S S Morse 12kv IRVINE S S Morse 12kv IRVINE S S Morse 12kv IRVINE S V Morse 12kv IRVINE V S Morse 12kv IRVINE S X Morse 12kv IRVINE V V Morse 12kv IRVINE V V Morse 12kv IRVINE V V Morse 12kv IRVINE V P Morse 12kv IRVINE V V Morse 12kv IRVINE V V Oxy 12kv IRVINE S B Oxy 12kv IRVINE B B Oxy 12kv IRVINE V M Oxy 12kv IRVINE M V Oxy 12kv IRVINE V M Oxy 12kv IRVINE V S Oxy 12kv IRVINE V S Oxy 12kv IRVINE S V Oxy 12kv IRVINE S V Oxy 12kv IRVINE M M Oxy 12kv IRVINE M M Oxy 12kv IRVINE M M Oxy 12kv IRVINE S V Oxy 12kv IRVINE S S Oxy 12kv IRVINE S P Oxy 12kv IRVINE P S Oxy 12kv IRVINE P P Oxy 12kv IRVINE P P Oxy 12kv IRVINE S P Oxy 12kv IRVINE P P Oxy 12kv IRVINE P P Oxy 12kv IRVINE P P Oxy 12kv IRVINE P S Oxy 12kv IRVINE S S Oxy 12kv IRVINE S V Oxy 12kv IRVINE V V Oxy 12kv IRVINE S S Oxy 12kv IRVINE S S Oxy 12kv IRVINE S V Oxy 12kv IRVINE P V POLLARD 12KV IRVINE M M REDHILL 12KV IRVINE M M Sapporo 12KV CORONA V V Sapporo 12KV CORONA V M Seminary Ventura T E Seminary Ventura S S Seminary Ventura S X Seminary Ventura X S Seminary Ventura S S Seminary Ventura S E Seminary Ventura E V

184 Workpaper Southern California Edison / 2018 GRC 85 Seminary Ventura E V Seminary Ventura V S Spinnaker OXNARD V V Spinnaker OXNARD V V Spinnaker OXNARD V E Spinnaker OXNARD S V Spinnaker OXNARD V V Story 4KV ALHAMBRA V M Story 4KV ALHAMBRA M M Story 4KV ALHAMBRA M V Story 4KV ALHAMBRA V M Story 4KV ALHAMBRA M V Story 4KV ALHAMBRA M V Stringer 4kV Torrance H X Stringer 4kV Torrance X M Stringer 4kV Torrance E M Stringer 4kV Torrance M M Stringer 4kV Torrance M M Stringer 4kV Torrance M M Stringer 4kV Torrance M E Stringer 4kV Torrance M M Stringer 4kV Torrance M E Stringer 4kV Torrance M E Surfside PORT HUENEME V V Surfside PORT HUENEME V V Surfside PORT HUENEME V V Surfside PORT HUENEME V V Surfside PORT HUENEME V V Surfside PORT HUENEME V V Surfside PORT HUENEME E M Surfside PORT HUENEME M V Surfside OXNARD S S Surfside OXNARD S E Surfside PORT HUENEME V V Surfside PORT HUENEME V V Surfside PORT HUENEME V S

185 86 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Cost of Cable in Conduit Replacements

186 Workpaper Southern California Edison / 2018 GRC 87 Work paper Title: COST OF CABLE IN CONDUIT REPLACEMENTS WBS Element: CET PD IR CC Witness: Jose Ramon Goizueta This workpaper establishes the average unit cost of replacing cable in conduits Cable in Conduit Counts 1,2,& Total Count Cable in Conduit Cable in Conduit Expenditures 2015 $ 1& Total Cable in Conduit $ 353,935 $ 3,289,740 $ 5,547,557 $ 22,722,821 $ 34,717,211 $ 66,631,264 Cable in Conduit Unit Costs ' Average Cable in Conduit $ 431,588 $ 222,112 $ 260,190 $ 335,888 $ 260,366 $ 280,031 Unit Cost Used for Forecasting 4& Cable in Conduit $ 260,366 $ 265,284 $ 270,799 $ 277,620 $ 286,329 $ 295,433 1 Annual expenditures and unit counts based on projects that occurred from 2011 to Annual expenditures and units counts based on projects that have been closed (construction completed, finalized accounting and financials) 3 Unit cost is derived by taking total expenditures and dividing by total unit counts 4 Unit Cost used for forecasting includes escalation 5 Last year recorded was chosen because of a changes in the program in recent years to drive cost down 6 Unit counts are in conductor miles

187 88 Workpaper Southern California Edison / 2018 GRC Workpaper Title: OCP Scope for 2016 and 2017

188 Workpaper Southern California Edison / 2018 GRC 89 OCP Scope PIF Substation Circuit District Region BI Category Project Year Lindsay 66/12 (D) Booster 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Telegraph 66/12 (D) Beta 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Piute 66/12 (D) Forage 36 North Coast Plant Betterment OH Conductor Program-Non 4 kv Little Rock 66/12 (D) Sun Village 36 North Coast Plant Betterment OH Conductor Program-Non 4 kv Fogarty 115/12 (D) Kleven 88 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Lancaster 66/12 (D) Joshua 36 North Coast Plant Betterment OH Conductor Program-Non 4 kv Pepper 115/12 (D) Lancers 30 Desert Plant Betterment OH Conductor Program-Non 4 kv Torrance 66/16 (D) Sago 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Puente 66/12 (D) Stafford 26 Metro East Plant Betterment OH Conductor Program-Non 4 kv Venida 66/12 (D) Wells 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Fillmore 66/16 (D) Barrington 39 North Coast Plant Betterment OH Conductor Program-Non 4 kv San Miguel 66/16 (D) Gringo 39 North Coast Plant Betterment OH Conductor Program-Non 4 kv Homart 115/12 (D) Silvertone 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Slater 66/12 (D) Chargers 33 Orange Plant Betterment OH Conductor Program-Non 4 kv Crater 66/16 (D) Plateau 35 North Coast Plant Betterment OH Conductor Program-Non 4 kv Wakefield 66/16 (D) Cadway 39 North Coast Plant Betterment OH Conductor Program-Non 4 kv Mayberry 115/12 (D) Stanford 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Modena 66/12 (D) Nepal 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Washington 66/12 (D) Pass 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Vera 66/12 (D) Mercedes 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Tortilla 115/12 (D) Huevos 72 Rurals Plant Betterment OH Conductor Program-Non 4 kv Sullivan 66/12 (D) English 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Ely 66/12 (D) Chile 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Jefferson 66/12 (D) Chimay 34 Metro East Plant Betterment OH Conductor Program-Non 4 kv Riverway 66/12 (D) Mississippi 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Soquel 66/12 (D) Butterfield 34 Metro East Plant Betterment OH Conductor Program-Non 4 kv Yukon 66/4.16 (D) Menlo 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Coffee 33/12 (D) Cafe 79 Desert Plant Betterment OH Conductor Program-Non 4 kv Lynwood 16/4.16 (D) Carlin 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv Shandin 115/12 (D) Northpark 31 Desert Plant Betterment OH Conductor Program-Non 4 kv San Marcos 66/16 (D) Duffer 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Cardiff 66/12 (D) Lena 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Eric 66/12 (D) Dornes 46 Metro West Plant Betterment OH Conductor Program-Non 4 kv West Riverside 33/12 (D) Winnebago 30 Desert Plant Betterment OH Conductor Program-Non 4 kv San Marcos 66/16 (D) Hurst 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Carmenita 66/12 (D) Orchardale 47 Metro West Plant Betterment OH Conductor Program-Non 4 kv Yukon 66/4.16 (D) Menlo 44 Metro West Plant Betterment OH Conductor Program-4 kv Woodruff 12/4.16 (D) Ardis 32 Metro West Plant Betterment OH Conductor Program-4 kv Tapia 66/16 (D) Tuna 35 North Coast Plant Betterment OH Conductor Program-Non 4 kv Farrell 115/12 (D) Volturno 79 Desert Plant Betterment OH Conductor Program-Non 4 kv Walnut 66/12 (D) Chella 26 Metro East Plant Betterment OH Conductor Program-Non 4 kv Rector 66/12 (D) Bush 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Alhambra 66/16 (D) Test 22 Metro East Plant Betterment OH Conductor Program-Non 4 kv El Nido 66/16 (D) Albacore 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Rector 66/12 (D) Linnell 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Tortilla 115/33 (D) Poco 72 Rurals Plant Betterment OH Conductor Program-Non 4 kv Belvedere 16/4.16 (A) Camy 22 Metro East Plant Betterment OH Conductor Program-4 kv Pepper 115/12 (D) Mateus 30 Desert Plant Betterment OH Conductor Program-Non 4 kv Terrace 16/4.16 (D) Rogers 22 Metro East Plant Betterment OH Conductor Program-4 kv Stetson 115/12 (D) Aviator 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv El Nido 66/16 (D) Grizzley 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Trask 66/12 (D) Akron 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Rosemead 66/16 (D) Telstar 22 Metro East Plant Betterment OH Conductor Program-Non 4 kv Apple Valley 115/12 (D) Spear 73 Rurals Plant Betterment OH Conductor Program-Non 4 kv Cady 33/12 (D) Hector 72 Rurals Plant Betterment OH Conductor Program-Non 4 kv Bryan 66/12 (D) Dreyer 29 Orange Plant Betterment OH Conductor Program-Non 4 kv 2016 Page 1 of 7

189 90 Workpaper Southern California Edison / 2018 GRC OCP Scope PIF Substation Circuit District Region BI Category Project Year Mayberry 115/12 (D) Tudor 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Inglewood 66/16 (D) Fairhaven 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Porterville 66/12 (D) Vandalia 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Lynwood 16/4.16 (D) Carlin 32 Metro West Plant Betterment OH Conductor Program-4 kv Eaton 66/16 (D) Baldwin 27 Metro East Plant Betterment OH Conductor Program-Non 4 kv Ivar 16/4.16 (D) Noel 22 Metro East Plant Betterment OH Conductor Program-4 kv Bliss 66/12 (D) Abbey 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Felton 66/16 (D) Deborah 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Coffee 33/12 (D) Blend 79 Desert Plant Betterment OH Conductor Program-Non 4 kv Apple Valley 115/12 (D) Tussing 73 Rurals Plant Betterment OH Conductor Program-Non 4 kv Shandin 115/12 (D) Carmelita 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Rush 66/16 (D) Brookline 22 Metro East Plant Betterment OH Conductor Program-Non 4 kv Cajalco 115/12 (D) Kimdale 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Lucerne 33/12 (D) Sky Hi 73 Rurals Plant Betterment OH Conductor Program-Non 4 kv Apple Valley 115/12 (D) Pahute 73 Rurals Plant Betterment OH Conductor Program-Non 4 kv Calden 66/16 (D) Ardmore 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv Homart 115/12 (D) Urbita 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Carolina 66/12 (D) Connecticut 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Trask 66/12 (D) Omaha 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Carpinteria 66/16 (D) Sheffield 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Woodville 66/12 (D) Flory 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Venice Hill 66/12 (D) Twin Butte 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Tipton 66/12 (D) Callison 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Rector 66/12 (D) Leo 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Liberty 66/12 (D) Freedom 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Tulare 66/12 (D) Paige 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Hanford 66/12 (D) Squadron 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Mascot 66/12 (D) Goldenbear 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Stadler 115/12 (D) Pigskin 88 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Wilsona 66/12 (D) Eastwind 36 North Coast Plant Betterment OH Conductor Program-Non 4 kv Alhambra 66/16 (D) Asteroid 22 Metro East Plant Betterment OH Conductor Program-Non 4 kv Culver 66/4.16 (D) Stevens 42 Metro West Plant Betterment OH Conductor Program-4 kv Redman 66/12 (D) Muroc 36 North Coast Plant Betterment OH Conductor Program-Non 4 kv Fullerton 66/4.16 (D) Ash 48 Orange Plant Betterment OH Conductor Program-4 kv Lennox 16/4.16 (D) Hardy 44 Metro West Plant Betterment OH Conductor Program-4 kv Pierpont 16/4.16 (D) Pearce 39 North Coast Plant Betterment OH Conductor Program-4 kv Octol 66/12 (D) Oakland 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Ojai 66/16 (D) Thacher 39 North Coast Plant Betterment OH Conductor Program-Non 4 kv Eric 66/12 (D) Mapes 46 Metro West Plant Betterment OH Conductor Program-Non 4 kv Marion 66/12 (D) Bertha 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Venida 66/12 (D) Orange Blossom 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Mascot 66/12 (D) Goldenbear 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Levy 66/16 (D) Saratoga 39 North Coast Plant Betterment OH Conductor Program-Non 4 kv Acton 66/12 (D) Bootlegger 36 North Coast Plant Betterment OH Conductor Program-Non 4 kv Telegraph 66/12 (D) Beta 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Hemet 33/12 (D) Green Acres 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Culver 66/16 (D) M.G.M. 42 Metro West Plant Betterment OH Conductor Program-Non 4 kv Placentia 66/12 (D) Coach 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Windsor Hills 16/4.16 (D) Mowder 44 Metro West Plant Betterment OH Conductor Program-4 kv Rolling Hills 66/4.16 (D) Violet 44 Metro West Plant Betterment OH Conductor Program-4 kv Telegraph 66/12 (D) Pete 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Mayberry 115/12 (D) Stanford 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv La Mirada 66/12 (D) Doe 47 Metro West Plant Betterment OH Conductor Program-Non 4 kv Neenach 66/12 (D) Tejon 36 North Coast Plant Betterment OH Conductor Program-Non 4 kv Vegas 66/16 (D) Storke 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Woodville 66/12 (D) Pratt 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv 2017 Page 2 of 7

190 Workpaper Southern California Edison / 2018 GRC 91 OCP Scope PIF Substation Circuit District Region BI Category Project Year Yorba Linda 66/12 (D) Hurricane 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Bayside 66/12 (D) Yellowtail 33 Orange Plant Betterment OH Conductor Program-Non 4 kv Inglewood 66/4.16 (D) Glasgow 44 Metro West Plant Betterment OH Conductor Program-4 kv Railroad 66/12 (D) Brakeman 26 Metro East Plant Betterment OH Conductor Program-Non 4 kv Olinda 66/12 (D) Whipstock 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Mayberry 115/12 (D) Sprague 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Temescal P.T. 33/12 (A) Polymer 34 Metro East Plant Betterment OH Conductor Program-Non 4 kv Bayside 66/12 (D) Tarpon 33 Orange Plant Betterment OH Conductor Program-Non 4 kv Santa Monica 66/16 (D) Albatross 42 Metro West Plant Betterment OH Conductor Program-Non 4 kv Citrus 66/12 (D) Banana 26 Metro East Plant Betterment OH Conductor Program-Non 4 kv Marion 66/12 (D) Mildred 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Cortez 66/12 (D) Vecino 26 Metro East Plant Betterment OH Conductor Program-Non 4 kv Victoria 66/16 (D) Selva 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Oceanview 66/12 (D) Bushard 33 Orange Plant Betterment OH Conductor Program-Non 4 kv Rector 66/12 (D) Nickerson 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Eric 66/12 (D) Semora 46 Metro West Plant Betterment OH Conductor Program-Non 4 kv Bloomington 66/12 (D) Bobber 30 Desert Plant Betterment OH Conductor Program-Non 4 kv Marion 66/12 (D) Annette 48 Orange Plant Betterment OH Conductor Program-Non 4 kv La Palma 66/12 (D) Delco 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Bandini 66/16 (D) Brickyard 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv Hamilton 66/12 (D) Gehrig 33 Orange Plant Betterment OH Conductor Program-Non 4 kv Hanford 66/12 (D) Gee Bee 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Bowl 66/12 (D) Pablo 46 Metro West Plant Betterment OH Conductor Program-Non 4 kv Broadway 66/12 (D) Argonne 46 Metro West Plant Betterment OH Conductor Program-Non 4 kv Barre 66/12 (D) Marigold 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Cortez 66/12 (D) Egmont 26 Metro East Plant Betterment OH Conductor Program-Non 4 kv Parkwood 66/12 (D) Hackberry 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Marion 66/12 (D) Diana 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Newbury 66/16 (D) Lesser 35 North Coast Plant Betterment OH Conductor Program-Non 4 kv Rush 66/16 (D) Peck 22 Metro East Plant Betterment OH Conductor Program-Non 4 kv Cypress 66/12 (D) Clair 46 Metro West Plant Betterment OH Conductor Program-Non 4 kv Walteria 66/16 (D) Blocker 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Washington 66/12 (D) Punt 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Culver 66/4.16 (D) Rimpau 42 Metro West Plant Betterment OH Conductor Program-4 kv Laguna Bell 66/16 (D) Agra 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv Vegas 66/16 (D) Gaucho 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Stadium 66/12 (D) Laws 46 Metro West Plant Betterment OH Conductor Program-Non 4 kv Newcomb 115/12 (D) Presley 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Torrance 66/16 (D) Aspen 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Murphy 66/12 (D) Kilkenny 47 Metro West Plant Betterment OH Conductor Program-Non 4 kv Rector 66/12 (D) Wutchumna 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv La Habra 66/12 (D) Twister 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Camarillo 66/16 (D) Evita 39 North Coast Plant Betterment OH Conductor Program-Non 4 kv Rubidoux 33/12 (D) Flabob 30 Desert Plant Betterment OH Conductor Program-Non 4 kv Alon 66/12 (D) Stocco 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv Liberty 66/12 (D) Blackstone 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Brea 66/12 (D) Cancun 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Cudahy 66/16 (D) Otis 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv Sepulveda 66/16 (D) Interstate 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Fremont 66/16 (D) Reasoner 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv Dalton 66/12 (D) Gravel 27 Metro East Plant Betterment OH Conductor Program-Non 4 kv Hamilton 66/12 (D) Score 33 Orange Plant Betterment OH Conductor Program-Non 4 kv Gage 16/4.16 (D) Priory 32 Metro West Plant Betterment OH Conductor Program-4 kv Culver 66/4.16 (D) Rimpau 42 Metro West Plant Betterment OH Conductor Program-4 kv Telegraph 66/12 (D) Alpha 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Bullis 66/16 (D) Trochu 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv 2017 Page 3 of 7

191 92 Workpaper Southern California Edison / 2018 GRC OCP Scope PIF Substation Circuit District Region BI Category Project Year Hanford 66/12 (D) Burris 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Newcomb 115/12 (D) Harnage 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Yukon 66/16 (D) Adak 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Mascot 66/12 (D) Bullpup 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Ely 66/12 (D) Cuba 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Gilbert 66/12 (D) Putter 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Beverly 66/16 (D) Hilton 42 Metro West Plant Betterment OH Conductor Program-Non 4 kv Bradbury 66/16 (D) Ambrus 27 Metro East Plant Betterment OH Conductor Program-Non 4 kv Lynwood 16/4.16 (D) Tweedy 32 Metro West Plant Betterment OH Conductor Program-4 kv Oak Grove 66/12 (D) Monson 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv San Bernardino 66/12 (D) Unicorn 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Alhambra 66/16 (D) Tri City 22 Metro East Plant Betterment OH Conductor Program-Non 4 kv Yukon 66/16 (D) Alaska 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Orange 66/12 (D) Red 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Vera 66/12 (D) Essex 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Carolina 66/12 (D) Vermont 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Tatanka P.T. 16/4.16 (A) Tatanka 49 North Coast Plant Betterment OH Conductor Program-4 kv Oak Grove 66/12 (D) Shirk 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Cucamonga 66/12 (D) Gurney 34 Metro East Plant Betterment OH Conductor Program-Non 4 kv Windsor Hills 66/16 (D) Stocker 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Porterville 66/12 (D) Hatfield 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Fullerton 66/12 (D) Roma 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Estero 66/16 (D) Tiros 39 North Coast Plant Betterment OH Conductor Program-Non 4 kv Felton 66/16 (D) Kathleen 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Victor 115/12 (D) Beeline 73 Rurals Plant Betterment OH Conductor Program-Non 4 kv Porterville 66/12 (D) Beattie 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Lancaster 66/12 (D) Medallion 36 North Coast Plant Betterment OH Conductor Program-Non 4 kv Cardiff 66/12 (D) Harlem Springs 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Randall 66/12 (D) Colleen 30 Desert Plant Betterment OH Conductor Program-Non 4 kv Walteria 66/16 (D) Tandem 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Francis 66/12 (D) Thor 34 Metro East Plant Betterment OH Conductor Program-Non 4 kv Narod 66/12 (D) Clutch 34 Metro East Plant Betterment OH Conductor Program-Non 4 kv Bridge 66/4.16 (D) Stringer 44 Metro West Plant Betterment OH Conductor Program-4 kv Camden 66/12 (D) Bismuth 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Edinger 12/4.16 (D) Flower 29 Orange Plant Betterment OH Conductor Program-4 kv Octol 66/12 (D) Sargent 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Villa Park 66/12 (D) Rutledge 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Strathmore 66/12 (D) Bowen 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Earlimart 66/12 (D) Doran 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Delano 66/12 (D) Ellington 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Timoteo 66/12 (D) Dental 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Visalia 66/12 (D) Giddings 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Santa Barbara 66/16 (D) Jameson 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Monolith 66/12 (D) Tomahawk 52 North Coast Plant Betterment OH Conductor Program-Non 4 kv Beverly 66/16 (D) Playboy 42 Metro West Plant Betterment OH Conductor Program-Non 4 kv Barre 66/12 (D) Tulip 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Fillmore 66/16 (D) Buckhorn 39 North Coast Plant Betterment OH Conductor Program-Non 4 kv Strathmore 66/12 (D) Prairie 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Villa Park 66/12 (D) Baylor 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Bloomington 66/12 (D) Bobber 30 Desert Plant Betterment OH Conductor Program-Non 4 kv Carson 66/16 (D) Sigma 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv Mentone 115/12 (D) Reed 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Olinda 66/12 (D) Whipstock 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Riverway 66/12 (D) Yellowstone 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Montecito 16/4.16 (D) Swift 49 North Coast Plant Betterment OH Conductor Program-4 kv Goshen 66/12 (D) Manzanillo 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv 2017 Page 4 of 7

192 Workpaper Southern California Edison / 2018 GRC 93 OCP Scope PIF Substation Circuit District Region BI Category Project Year Parkwood 66/12 (D) Pyrus 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Liberty 66/12 (D) Allegiance 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Newcomb 115/12 (D) Oakdale 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Neptune 66/4.16 (D) Keystone 32 Metro West Plant Betterment OH Conductor Program-4 kv Bryan 66/12 (D) Bermuda 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Santa Barbara 66/16 (D) Hot Springs 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Placentia 66/12 (D) Coach 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Orange 66/12 (D) Magenta 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Colorado 66/16 (D) Madison 42 Metro West Plant Betterment OH Conductor Program-Non 4 kv Highland 66/12 (D) Pencil 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Eaton 66/16 (D) Paloma 27 Metro East Plant Betterment OH Conductor Program-Non 4 kv Colorado 66/4.16 (D) Medical 42 Metro West Plant Betterment OH Conductor Program-4 kv Parkwood 66/12 (D) Carob 48 Orange Plant Betterment OH Conductor Program-Non 4 kv Vegas 66/16 (D) Storke 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Rector 66/12 (D) Hart 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Cudahy 66/4.16 (D) Sales 32 Metro West Plant Betterment OH Conductor Program-4 kv Del Rosa 66/12 (D) Mulkey 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Riverway 66/12 (D) Gila 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Downs 33/12 (D) Bowman_old 86 Rurals Plant Betterment OH Conductor Program-Non 4 kv Los Cerritos 12/4.16 (D) Corrine 46 Metro West Plant Betterment OH Conductor Program-4 kv Ojai 66/16 (D) Patricia 39 North Coast Plant Betterment OH Conductor Program-Non 4 kv Rialto 33/4.16 (D) Love 30 Desert Plant Betterment OH Conductor Program-4 kv Broadcom P.T. 12/2.4 (A) Broadcom 29 Orange Plant Betterment OH Conductor Program-4 kv Bunker 115/12 (D) Soldier 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Isla Vista 66/16 (D) Gladiola 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Coffee 33/12 (D) Maxim 79 Desert Plant Betterment OH Conductor Program-Non 4 kv Bryan 66/12 (D) Brazil 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Cypress 66/12 (D) Prida 46 Metro West Plant Betterment OH Conductor Program-Non 4 kv Lark Ellen 66/12 (D) Eleanor 26 Metro East Plant Betterment OH Conductor Program-Non 4 kv Vestal 66/12 (D) Caratan 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Cypress 66/12 (D) Katella 46 Metro West Plant Betterment OH Conductor Program-Non 4 kv Neptune 66/4.16 (D) Fiat 32 Metro West Plant Betterment OH Conductor Program-4 kv Stewart 66/12 (D) Clifford 47 Metro West Plant Betterment OH Conductor Program-Non 4 kv Dalton 66/12 (D) Dameral 27 Metro East Plant Betterment OH Conductor Program-Non 4 kv Topaz 66/4.16 (D) Tourmaline 44 Metro West Plant Betterment OH Conductor Program-4 kv Farrell 115/12 (D) Viminal 79 Desert Plant Betterment OH Conductor Program-Non 4 kv Visalia 66/12 (D) Caldwell 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv El Nido 66/16 (D) Bluegill 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Sullivan 66/12 (D) Calico 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Sixteenth Street 33/12 (D) Rye 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Ditmar 66/16 (D) Valve 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Vegas 66/16 (D) Tecolote 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Sullivan 66/12 (D) Willets 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Alder 66/12 (D) Huff 30 Desert Plant Betterment OH Conductor Program-Non 4 kv Victoria 66/16 (D) Arvana 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Arcadia 66/4.16 (D) Haven 27 Metro East Plant Betterment OH Conductor Program-4 kv Bradbury 66/16 (D) Primrose 27 Metro East Plant Betterment OH Conductor Program-Non 4 kv Brighton 66/16 (D) Chadron 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Vegas 66/16 (D) Wasp 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Venice Hill 66/12 (D) Redbanks 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Strathmore 66/12 (D) Gillette 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Fremont 16/4.16 (D) Mckinley 32 Metro West Plant Betterment OH Conductor Program-4 kv Strathmore 66/12 (D) Prairie 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Tipton 66/12 (D) Surprise 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Los Cerritos 66/12 (D) Mossman 46 Metro West Plant Betterment OH Conductor Program-Non 4 kv Delano 66/12 (D) Radnor 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv 2017 Page 5 of 7

193 94 Workpaper Southern California Edison / 2018 GRC OCP Scope PIF Substation Circuit District Region BI Category Project Year Huntington Park 16/4.16 (D) Walnut Park 32 Metro West Plant Betterment OH Conductor Program-4 kv Slater 66/12 (D) Saints 33 Orange Plant Betterment OH Conductor Program-Non 4 kv Colonia 66/16 (D) Fifth St. 39 North Coast Plant Betterment OH Conductor Program-Non 4 kv Yukon 66/4.16 (D) Fairbanks 44 Metro West Plant Betterment OH Conductor Program-4 kv Pepper 115/12 (D) Cabernet 30 Desert Plant Betterment OH Conductor Program-Non 4 kv Brighton 66/16 (D) Rhumba 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Locust 12/4.16 (D) Lee 46 Metro West Plant Betterment OH Conductor Program-4 kv Oldfield 12/4.16 (D) Matney 46 Metro West Plant Betterment OH Conductor Program-4 kv Tulare 66/12 (D) Aurora 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Neptune 66/12 (D) Storm 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv Bliss 66/12 (D) Serenity 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Laurel 66/12 (D) Ball 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Fremont 66/16 (D) Lantana 32 Metro West Plant Betterment OH Conductor Program-Non 4 kv Mayberry 115/12 (D) Erin 77 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Imperial 66/4.16 (D) James 47 Metro West Plant Betterment OH Conductor Program-4 kv Puente 66/12 (D) Azores 26 Metro East Plant Betterment OH Conductor Program-Non 4 kv Olympic 16/4.16 (D) Wetherly 42 Metro West Plant Betterment OH Conductor Program-4 kv Hanford 66/12 (D) Mussel 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Walnut 66/12 (D) Grazide 26 Metro East Plant Betterment OH Conductor Program-Non 4 kv Venida 66/12 (D) Wells 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Strathmore 66/12 (D) Gillette 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Brewster 16/4.16 (D) Vita 32 Metro West Plant Betterment OH Conductor Program-4 kv Ditmar 16/4.16 (D) Janice 44 Metro West Plant Betterment OH Conductor Program-4 kv Signal Hill 12/4.16 (D) Gundry 46 Metro West Plant Betterment OH Conductor Program-4 kv Mentone 115/12 (D) Hass 31 Desert Plant Betterment OH Conductor Program-Non 4 kv Mt. Tom 55/12 (D) Birchim 85 Rurals Plant Betterment OH Conductor Program-Non 4 kv Bloomington 66/12 (D) Highball 30 Desert Plant Betterment OH Conductor Program-Non 4 kv Redman 66/12 (D) Alfalfa 36 North Coast Plant Betterment OH Conductor Program-Non 4 kv Tipton 66/12 (D) Surprise 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Lindsay 66/12 (D) Cairns 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Venida 66/12 (D) Dungan 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Longdon 16/4.16 (D) Glencoe 32 Metro West Plant Betterment OH Conductor Program-4 kv Laurel 66/12 (D) Linder 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Pioneer 66/12 (D) Settler 47 Metro West Plant Betterment OH Conductor Program-Non 4 kv Lampson 66/12 (D) Cheetah 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Garfield 66/4.16 (D) Cawston 27 Metro East Plant Betterment OH Conductor Program-4 kv Lynwood 16/4.16 (D) Cornish 32 Metro West Plant Betterment OH Conductor Program-4 kv Victoria 66/16 (D) Nelson 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv San Antonio 66/12 (D) Satellite 34 Metro East Plant Betterment OH Conductor Program-Non 4 kv Pepper 115/12 (D) Sauterne 30 Desert Plant Betterment OH Conductor Program-Non 4 kv Edinger 12/4.16 (D) Anahurst 29 Orange Plant Betterment OH Conductor Program-4 kv Bliss 66/12 (D) Serenity 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Inglewood 66/4.16 (D) England 44 Metro West Plant Betterment OH Conductor Program-4 kv Mariposa 66/12 (D) Burum 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Talbert 66/12 (D) Oriole 29 Orange Plant Betterment OH Conductor Program-Non 4 kv Passons 66/12 (D) Jayblue 47 Metro West Plant Betterment OH Conductor Program-Non 4 kv Vegas 66/16 (D) Tecolote 49 North Coast Plant Betterment OH Conductor Program-Non 4 kv Venida 66/12 (D) Wells 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Brookhurst 66/12 (D) Tittle 33 Orange Plant Betterment OH Conductor Program-Non 4 kv Cove P.T. 33/12 (A) Cove 79 Desert Plant Betterment OH Conductor Program-Non 4 kv Skylark 115/12 (D) Stoneman 88 San Jacinto Valley Plant Betterment OH Conductor Program-Non 4 kv Brighton 66/16 (D) Rhumba 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Terra Bella 66/12 (D) Skinkle 51 San Joaquin Plant Betterment OH Conductor Program-Non 4 kv Downs 33/12 (D) Shangrila_old 86 Rurals Plant Betterment OH Conductor Program-Non 4 kv Victoria 66/16 (D) Arvana 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Amador 66/4.16 (D) Tyler 22 Metro East Plant Betterment OH Conductor Program-4 kv 2017 Page 6 of 7

194 Workpaper Southern California Edison / 2018 GRC 95 OCP Scope PIF Substation Circuit District Region BI Category Project Year Walteria 66/16 (D) Blocker 44 Metro West Plant Betterment OH Conductor Program-Non 4 kv Santa Fe Springs 66/12 (D) Laird 47 Metro West Plant Betterment OH Conductor Program-Non 4 kv 2017 Page 7 of 7

195 96 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Cost of Overhead Conductor Rebuilds

196 Workpaper Southern California Edison / 2018 GRC 97 Work paper Title: COST OF OVERHEAD CONDUCTOR REBUILDS WBS Element: CET PD IR OC Witness: Jose Ramon Goizueta This workpaper establishes the average unit cost of rebuilding the overhead circuits Overhead Conductor Rebuild Counts 1,2,& Total Count OH Conductor Rebuilds 5 5 Overhead Conductor Rebuild Expenditures 2015 $ 1& Total OH Conductor Rebuilds $ 2,138,762 $ 2,138,762 Overhead Conductor Rebuild Unit Costs 3& Average OH Conductor Rebuilds $ 436,145 $ 436,145 Unit Cost Used for Forecasting 4& OH Conductor Rebuilds $ 436,145 $ 444,383 $ 453,622 $ 465,047 $ 479,636 $ 494,887 1 Annual expenditures and unit counts based on projects that occurred in Annual expenditures and units counts based on projects that have been closed (construction completed, accounting and financials have been finalized) 3 Unit cost per year is derived by taking total expenditures and dividing by total unit counts 4 Unit Cost used for forecasting includes escalation 5 Last year recorded was selected for forecasting due to limited data because the program was conceived in Units are in circuit miles are equivalent to geographical miles and do not account for multiple phases

197 98 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Cost of Underground Oil Switch Replacements

198 Workpaper Southern California Edison / 2018 GRC 99 Work paper Title: COST OF UNDERGROUND OIL SWITCH REPLACEMENTS WBS Element: CET PD IR SR Witness: Jose Ramon Goizueta This workpaper establishes the average unit cost of replacing a switch UG Oil Switch Replacement Counts 1& Total Count Switch ,409 UG Oil Switch Replacement Expenditures 2015 $ 1& Total Switch $ 13,111,565 $ 12,518,283 $ 18,446,044 $ 18,361,230 $ 21,479,805 $ 83,916,926 UG Oil Switch Replacement Unit Costs ' Average Switch $ 41,890 $ 44,234 $ 49,586 $ 76,825 $ 106,336 $ 59,558 Unit Cost Used for Forecasting 4& Switch $ 59,558 $ 60,683 $ 61,944 $ 63,505 $ 65,497 $ 67,579 1 Annual expenditures and unit counts based on projects that occurred from 2011 to Annual expenditures and units counts based on projects that have been closed (construction completed, finalized accounting and financials) 3 Unit cost per year is derived by taking total expenditures and dividing by total unit counts 4 Unit Cost used for forecasting includes escalation 5 Five year weighted average was selected for forecasting because of significant increase in cost in 2014 and 2015 that are deemed to be outliers

199 100 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Underground Oil Switch Reliability

200 Workpaper Southern California Edison / 2018 GRC Skabodolrka Mfi Qtfq`e Pbif^_fifqv QAC+.1* Tli,.2* Ufqkbpp8 Pldbo Jbb QAC R$B?KQP <NN@O H<I<B@H@IO <I? NTNO@H M@GD<=DGDOT BMJPK M@KJMO H[chfch_ No\mol`[]_ Jcf Nqcn]b M _fc[ \ cfcns Hi^_f -B>9 %#$# 1+0/13 ).21!$#!#' 4" 2" 1?=86> 0A9@6A98 7C( "," *;<9> 0A9@6A98 7C( N^db / lc 3

201 102 Workpaper Southern California Edison / 2018 GRC 54 Ulohm^mbo + Qlrqebok A^ifclokf^ Cafplk - 0./3 EPA +?NNJGA?RGML Skabodolrka Mfi Qtfq`e Pbif^_fifqv QAC+.1* Tli,.2* Ufqkbpp8 Pldbo Jbb QAC R$B?KQP F K GCAOFQC 6N_NUXY J VJRWURWN \^K\^[OJLN XRU \`R]LQ [NURJKRUR]b VXMNU ]QJ] `RUU \QX` ]QN LX[[NUJ]RXW XO ]QN \`R]LQ JPN _N[\^\ [NURJKRUR]b' FF LLMK AE DQN JYY[XJLQ NVYUXb\ [R\T&KJ\NM VN]QXMXUXPb `QRLQ RWLU^MN\ J LXVKRWJ]RXW XO Y[XKJKRUR\]RL% [NURJKRUR]b% JWM NWPRWNN[RWP RWOX[VJ]RXW' ;] RW_XU_N\ ^WMN[\]JWMRWP ]QN JPN XO VJRWURWN \^K\^[OJLN XRU \`R]LQ% LXUUNL]RWP JUU [NUN_JW] MJ]J R'N' L^[[NW] RW_NW]X[b JWM OJRU^[N QR\]X[b% YN[OX[VRWP MJ]J JWJUb\R\ JWM RW]N[_RN`\ `R]Q NWPRWNN[\'?<O< >JGG@>ODJI <I? <I<GTNDN DQN MN_NUXYVNW] XO ]QN VJRWURWN \^K\^[OJLN XRU \`R]LQ [NURJKRUR]b VXMNU KNPRW\ `R]Q ]QN ]J\T XO J\\N\\RWP R]\ RW_NW]X[b% "R'N'% QX` V^LQ XO `QJ] _RW]JPN# JWM R]\ OJRU^[N QR\]X[b' DJKUN ) \QX`\ ]QN \X^[LN\ XO MJ]J OX[ ]QN L^[[NW] YXY^UJ]RXW JWM QR\]X[RLJU MJ]J XO VJRWURWN \^K\^[OJLN XRU \`R]LQ' DJKUN ) & CX^[LN\ XO VJRWURWN \^K\^[OJLN XRU \`R]LQ RW_NW]X[b MJ]J JWM OJRU^[N MJ]J2 )' AJ\\YX[] E9 C`R]LQ MJ]J *' 6R\][RK^]RXW A[XSNL] ;WOX[VJ]RXW Cb\]NV "6A;C# C57 QJ\ JW NZ^RYVNW] VJRW]NWJWLN JWM RW\YNL]RXW MJ]JKJ\N% TWX`W J\ YJ\\YX[]% JWM 6N\RPW C^YYX[] QJ\ 6A;C ]QJ] \]X[N\ Y[XSNL] `X[T MJ]J% \][^L]^[N\% NZ^RYVNW]% JWM X]QN[ O^WL]RXW\ [NUJ]RWP ]X ]QN C57 MR\][RK^]RXW \b\]nv' DQN\N MJ]JKJ\N\ LXW]JRW MN]JRUNM RWOX[VJ]RXW% RWLU^MRWP RW\]JUUJ]RXW MJ]N XO ]QN VJRWURWN \^K\^[OJLN XRU \`R]LQ ]QJ] J[N L^[[NW]Ub RW \N[_RLN "8RP^[N )# JWM [NVX_NM O[XV \N[_RLN' DQR\ JUUX`\ JW ^WMN[\]JWMRWP XO ]QN _XU^VN XO NJLQ JPN P[X^Y XO ]QN VJRWURWN \^K\^[OJLN XRU \`R]LQ' 0 lc 3 N^db 0 lc 3 Cuef_fq Ll, QAC+.1 - Ro^kpjfppflk $ Bfpqof_rqflk - Tli,.2 Ufqkbpp8 P, Jbb

202 Workpaper Southern California Edison / 2018 GRC 103 Ulohm^mbo + Qlrqebok A^ifclokf^ Cafplk - 0./3 EPA +?NNJGA?RGML S kabodolrka Mfi Qtfq`e Pbif^_fifqv QAC+.1* Tli,.2* Ufqkbpp8 Pldbo Jbb M=? N#> ;GML.--- Cdl[djeho e\ GW_db_d[ MkXikh\WY[ I_b Mm_jY^[i Xo S[Wh e\ CdijWbbWj_ed %Wi e\ o[wh*[dz /--6& <JN>G. )'%) -$$,$$ +$$ J *$$ S )$$ 5 < HD G O, e ($$ AJ '$$ L /-- LF.-- T EPE Ǿ = = = = 6 G F F F G G G G G F G F 4 %%%% ~ %-)$ %-)) %-*$ %-*) %-+$ %-+) %-,$ %-,) %--$ %--) &$$$ &$$) S[Wh 8RP^[N ) & 3PN MR\][RK^]RXW XO \^K\^[OJLN VJRWURWN XRU \`R]LQN\ CDNOJMD><G?<O< DIAJMH<ODJI 6A;C \]X[N\ Y[XSNL] `X[T MJ]J% \][^L]^[N\% NZ^RYVNW]% JWM X]QN[ O^WL]RXW\ [NUJ]RWP ]X ]QN C57 MR\][RK^]RXW \b\]nv' 6A;C LXW]JRW MN]JRUNM RWOX[VJ]RXW% RWLU^MRWP JPN XO OJRU^[N% RW\]JUUJ]RXW MJ]N% JWM LJ^\N XO OJRU^[N\ XO ]QN VJRWURWN \^K\^[OJLN XRU \`R]LQ' DQR\ JUUX`\ JW ^WMN[\]JWMRWP XO ]QN [NVX_JU LJ^\N\ JWM ][NWM XO [NVX_JU YN[ bnj[' 8RP^[N * \QX` ]QN VJRWURWN \^K\^[OJLN XRU \`R]LQ OJRU^[N\ OX[ ]QN )1..& /( JPN K[JLTN]' 8RP^[N * & >JRWURWN C`R]LQ 8JRU^[N 9[X^Y INJ[ )1..&)1/( 1 lc 3 N^db 1 lc 3 Cuef_fq Ll, QAC+.1 - Ro^kpjfppflk $ Bfpqof_rqflk - Tli,.2 Ufqkbpp8 P, Jbb

203 104 Workpaper Southern California Edison / 2018 GRC 56 Ulohm^mbo + Qlrqebok A^ifclokf^ Cafplk - 0./3 EPA +?NNJGA?RGML Skabodolrka Mfi Qtfq`e Pbif^_fifqv QAC+.1* Tli,.2* Ufqkbpp8 Pldbo Jbb IKBCH BCQCHKLICJO QAC R$B?KQP DQN MN_NUXYVNW] XO ]QN VJRWURWN \^K\^[OJLN XRU \`R]LQ [NURJKRUR]b VXMNU LJW KN OJLRUR]J]NM Kb ]QN ^\N XO LXVVN[LRJUUb J_JRUJKUN \]J]R\]RL\ \XO]`J[N' C^LQ ]XXU\ J[N Na][NVNUb QNUYO^U RW PNWN[J]RWP J LXW\R\]NW] \N] XO JPN&MNYNWMNW] LXVYXWNW] OJRU^[N Y[XKJKRUR]RN\% _N[RObRWP MJ]J NW][b% ^YMJ]RWP MJ]J% JWM MN_NUXYRWP MXL^VNW]J]RXW' C57 ^]RURcNM XWN XO ]QN\N Y[XM^L]\ JWM \NUNL]NM ]QN GNRK^UU MR\][RK^]RXW J\ ]QN XWN KN\] VNN]RWP R]\ L^[[NW] XKSNL]R_N\' R@D=PGG =<N@? H<DIGDI@ NP=NPMA<>@ JDG NRDO>C M@GD<=DGDOT HJ?@GGDIB GR]Q ]QN N\]JKUR\QVNW] XO JPN MNYNWMNW] VJRWURWN \^K\^[OJLN XRU \`R]LQ OJRU^[N\ JWM ]QN ]X]JU W^VKN[ RW&\N[_RLN RW NJLQ JPN K[JLTN]% ]QN LX[[N\YXWMRWP `NJ[ X^] KJ\NM LXVYXWNW] OJRU^[N [J]N LJW KN N\]JKUR\QNM ;W]NP[J]RWP ]QR\ JPN MNYNWMNW] [N\^U] `R]Q WXW&JPN [NUJ]NM [JWMXV OJRU^[N% ]X]JU VJRWURWN \^K\^[OJLN XRU \`R]LQ OJRU^[N [J]N J\ O^WL]RXW XO JPN LJW KN PNWN[J]NM' DQN GNRK^UU $$/ \XO]`J[N YJLTJPN MN_NUXYNM Kb BNURJ\XO] R\ LQX\NW ]X PNWN[J]N ]QN WNNMNM Y[XKJKRUR\]RL OJRU^[N MR\][RK^]RXW\' DQN \]NY&Kb& \]NY Y[XLN\\ ]X OR] ]QN VJRWURWN \^K\^[OJLN XRU \`R]LQ OJRU^[N [J]N OX[ GNRK^UU YJ[JVN]N[\ "OXUUX`NM `R]Q L[NJ]RXW XO ]QN O^UU GNRK^UU MR\][RK^]RXW# "BNON[NWLN,# R\ WX] RW ]QN ]NVYUJ]N XO GNRK^UU$$/ JWJUb\R\' DQN Z^JW]RORLJ]RXW Y[XLN\\ ]JRUX[NM OX[ ]QN VJRWURWN \^K\^[OJLN XRU \`R]LQ [NURJKRUR]b VXMNU QJ\ KNNW MN_NUXYNM JWM Y[N\NW]NM RW DJKUN +2 DJKUN * & >JRWURWN C`R]LQ GNRK^UU Y[XKJKRUR]b O^WL]RXW MN_NUXYVNW] Y[XLN\\2 )' 6N[R_N JPN MNYNWMNW] OJRU^[N [J]N\ O[XV OJRU^[N MJ]J JWM YXY^UJ]RXW QR\]XP[JV *' E\N GNRK^UU / >XW&URWNJ[ NZ^J]RXW OR] \XU_N[ ]X PNWN[J]N GNRK^UU YJ[JVN]N[\ +' E\N GNRK^UU YJ[JVN]N[\ 4N]J JWM 7]J ]X MN_NUXY WNNMNM GNRK^UU O^WL]RXW\ RWLU^MRWP O"]#% 8"]#% N]L',' ;WY^] ]RVN JWM 8"]# ]X \]JWMJ[M GNRK^UU MJ]J \QNN] ]X MN[R_N [NZ^R[NM Y[XKJKRUR]b O^WL]RXW\ \^LQ J\2 ^W[NURJKRUR]b MR\][RK^]RXW% JWM OJRU^[N [J]N' 2 lc 3 N^db 2 lc 3 Cuef_fq Ll, QAC+.1 - Ro^kpjfppflk $ Bfpqof_rqflk - Tli,.2 Ufqkbpp8 P, Jbb

204 Workpaper Southern California Edison / 2018 GRC 105 Ulohm^mbo + Qlrqebok A^ifclokf^ Cafplk - 0./3 EPA +?NNJGA?RGML 57 Skabodolrka Mfi Qtfq`e Pbif^_fifqv QAC+.1* Tli,.2* Ufqkbpp8 Pldbo Jbb FFF MCNPHON GW_db_d[ MkXikh\WY[ I_b Mm_jY^ L[b_WX_b_jo M=? N#> ;GML DQN [N\^U]\ XO ]QN JWJUb\R\ brnumnm \]J]R\]RLJU _JU^N\ J\ \QX`W RW DJKUN +' DQN VJRWURWN \^K\^[OJLN XRU \`R]LQ JPN MNYNWMNW] OJRU^[N [J]N JWM ^W[NURJKRUR]b O^WL]RXW J[N \QX`W RW 8RP^[N +' DJKUN + & >JRWURWN C`R]LQN\ C]J]R\]RL\ >JRWURWN C`R]LQN\ -+)- 3_N[ 3PN *- >DD8 "INJ[# +- L[b_WMe\j Q[_Xkbb(( 4 * mmm+l[b_wme\j+yec 2<DGOL@ 7<N@ PM 9DH@ 6GJN 8RP^[N + & A[XKJKRUR]b XO 8JRU^[N _\' 3PN OX[ \^K\^[OJLN VJRWURWN XRU \`R]LQN\ FQ MCDCMCJACN )' *()* 9B5 GX[TYJYN[ C57% FXU + 6NY[NLRJ]RXW C]^Mb2 3LLX^W] +./ OX[ 6R\][RK^]RXW E9 C`R]LQN\ Y[NYJ[NM Kb 3YYJ[J]^\ 7WPRWNN[RWP% D 6 C57 *' AJ\\YX[] 6J]J +' 6A;C 6J]J,' GNRK^UU$$/ R\\^NM Kb BNURJCXO] 3 lc 3 N^db 3 lc 3 Cuef_fq Ll, QAC+.1 - Ro^kpjfppflk $ Bfpqof_rqflk - Tli,.2 Ufqkbpp8 P, Jbb

205 106 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Cost of Capacitor Bank Replacements

206 Workpaper Southern California Edison / 2018 GRC 107 Work paper Title: COST OF CAPACITOR BANK REPLACEMENTS WBS Element: CET PD IR CB Witness: Jose Ramon Goizueta This workpaper establishes the average unit cost of replacing a capacitor bank Capacitor Bank Replacements Counts 1& Total Count Capacitor Bank Capacitor Bank Replacement Expenditures 2015 $ 1& Total Capacitor Bank $ 6,965,839 $ 6,847,462 $ 9,465,816 $ 9,503,549 $ 6,833,624 $ 39,616,290 Capacitor Bank Replacements Unit Costs ' Average Capacitor Bank $ 27,533 $ 33,898 $ 30,733 $ 37,563 $ 47,128 $ 34,123 Unit Cost Used for Forecasting 4& Capacitor Bank $ 47,128 $ 48,019 $ 49,017 $ 50,252 $ 51,828 $ 53,476 1 Annual expenditures and unit counts based on projects that occurred from 2011 to Annual expenditures and units counts based on projects that have been closed (construction completed, finalized accounting and financials). 3 Unit cost per year is derived by taking total expenditures and dividing by total unit counts 4 Unit Cost used for forecasting includes escalation 5 Last year recorded was chosen because of increasing cost over the last five years

207 108 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Capacitor Banks

208 Workpaper Southern California Edison / 2018 GRC 109 General Information Account 368 Distribution Capacitor Banks CPR A capacitor is a device that stores energy. Southern California Edison uses switched and fixed capacitor banks to compensate for Var losses (reactive power) and regulate voltages on the distribution system. Fixed capacitor banks are planned for a minimum Var requirement (light-load) and switched capacitor banks are planned for a maximum Var requirement (peak-load) of distribution circuits. Edison primarily utilizes switched capacitor banks for voltage regulation. The majority of the distribution capacitor controls operate based on the primary circuit voltage at its connection point (some controls are strictly temperature or time controlled). The capacitor bank controllers that function based on the primary circuit voltage input follow a bandwidth whereby a capacitor would switch OFF when the primary voltage reaches the upper bandwidth limit and likewise the capacitor bank would switch ON when the primary voltage reaches the lower bandwidth limit. In order to compensate for additional voltage drop during peak conditions in the secondary (120/240 volt) system, some of our capacitor bank controllers function based on a time bias and/or temperature bias. Where the bias will raise or lower the bandwidth during specific times of the day or temperature conditions. The bias is a means to estimate peak conditions and thereby provide additional voltage support. Capacitor banks are found in both overhead and underground circuits. A padmount capacitor bank design is used for the underground applications. There are approximately 12,228 capacitor banks in the distribution system. Of these, approximately 9,293 are switched capacitor banks and 2,935 are fixed capacitor banks. Each fixed capacitor bank is composed of three capacitor units, fuses in holders, rack and mounting hardware. In addition to these elements, a switched capacitor bank requires a potential transformer, fuses for potential transformer, potential transformer mounting bracket, two vacuum capacitor switches and one capacitor controller. See the following figures 1 and 2 for examples of Overhead and Padmount capacitor banks

209 110 Workpaper Southern California Edison / 2018 GRC Figure 1. Overhead Switched Capacitor Bank Figure 2. Padmount Switched Capacitor Bank. Enclosure (left) and Inside enclosure (right)

210 Workpaper Southern California Edison / 2018 GRC 111 Expected Average Life The expected average service life of the capacitor bank elements is estimated as follow: Capacitor Units years Capacitor Switch years Capacitor Controller years The life expectancy of an overhead capacitor bank may be estimated at years with the understanding that some elements such as the capacitor switch or capacitor controller may need to be replaced due to failure within this period of years. A padmount capacitor bank s life expectancy is estimated at years. This shorter life is mainly due to corrosion of the enclosure. It is not practical or cost effective to replace only the corroded enclosure therefore once the enclosure is affected the whole capacitor banks with its elements are replaced. This information is based on manufacturer design estimates, engineering and field crew s experience with the equipment. Major Causes of Retirement The inspection program of the capacitor banks is a part of our preventive maintenance program. Once every five years, each capacitor bank in our system is inspected for proper operation, corrosion, oil leakage and loose connections. Typically, capacitor banks are replaced once they fail and have reached over 20 years of service. The impact of a single capacitor bank failure is usually only a small drop in circuit voltage. This drop becomes consequential only when there are several failed capacitor banks in the same circuit. When failures do occur, newer capacitor banks are typically repaired while older capacitor banks are typically replaced. These older capacitor banks may contain obsolete components which cannot be cost-effectively repaired or the parts required to repair it are not longer available, therefore the entire capacitor bank would be replaced. In addition, in the case of padmount capacitor banks, the retirement of the unit may be dictated based on the condition of the enclosure. If the enclosure is severely affected by corrosion the whole capacitor bank will be replaced. There are still many capacitor banks with obsolete capacitor switches in service. In the event that one of the switches fails and the capacitor bank is in good condition, all switches in this capacitor bank will be replaced

211 112 Workpaper Southern California Edison / 2018 GRC All the switches on a capacitor bank must be of the same design and operation to guarantee proper operation. In some cases, as the circuit needs change and the demand for more Var increases, many capacitor banks may be replaced by a bigger size capacitor bank. Southern California Edison has been actively replacing and/or installing approximately 400 to 500 capacitor banks per year. Many capacitor banks have already exceeded their life expectancy and others are soon approaching their 20 years of service. In order to continue to sustain the steady growth in demand for Var support with new installations and continue to support the replacement of failed or retired capacitor banks, the estimate of 400 to 500 new banks per year is expected to remain consistent thought out the next few years. Retirement Frequency by Age As previously discussed, the retirement or replacement of capacitor banks is a result of mainly failures or obsolete elements. Age is not a factor that will dictate the need to replace a capacitor bank but simply a guide to estimate the end of the service life of a capacitor bank. It is not unusual for a capacitor bank to exceed its life expectancy and continue to function for several years before failure. Accordingly, reasonable estimate or determination to establish retirement frequency by age would be impractical to achieve, although failures most likely accelerate in proportion to their years of service. There is no pattern or trend for which retirement of capacitors based specifically on age could be ascertained. Impacts of Dynamic Environment Southern California Edison has been proactively looking for ways to improve capacitor bank design and evaluating new equipment to extend their life in service. Within the last 10 years several improvements have been accomplished in an effort to prolong the frequency of retirement and impact service life of a capacitor bank over the next 20 to 30 years. Some of these improvements are as follow: In 2002, Edison standardized on capacitor vacuum switches for both overhead and padmount capacitor banks. The vacuum switch is expected to last longer than oil switches due to its higher

212 Workpaper Southern California Edison / 2018 GRC 113 In 2001, Edison standardized the 2-switch capacitor bank design instead of the original 3-switch design. By reducing the number of switches in a capacitor bank, the possibility of switch failures is reduced, therefore improving the reliability of the bank. Switch failure is one of the most common causes of failure in a cap bank. In 2001, Edison standardized on dead-front padmount capacitor banks. This improves reliability by reducing circuit interruptions and bank failures caused by animal intrusion. It also increases safety by reducing the exposure of live line components in the bank. In 2008, Edison standardized on a Heavy Duty capacitor unit design as a reaction to the increase in premature failures of regular capacitor. Factors such as Edison s standard Y- ungrounded capacitor bank design and thin dielectric films in regular capacitor design are believed to be the cause of premature failures. The Heavy Duty capacitor unit although more expensive than the regular capacitor unit, is expected to improve reliability, reduce maintenance cost and prevent premature failures of capacitor units. Except for the technological advances meant to improve capacitor s reliability and performance, there are no other significant technological advances over the past 15 years, nor technological changes that are anticipated in the near future that would influence the frequency of retirement and impact the service life of capacitor banks. Similarly, there are no policy, legal, procedural, or regulatory changes over the past 15 years or similar changes expected in the future that would influence the frequency of retirement and impact the service life of capacitor banks

213 114 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Cost of Automatic Recloser Replacements

214 Workpaper Southern California Edison / 2018 GRC 115 Work paper Title: COST OF AUTOMATIC RECLOSER REPLACEMENTS WBS Element: CET PD IR AR Witness: Jose Ramon Goizueta This workpaper establishes the average unit cost of replacing an automatic recloser Automatic Recloser Replacements Counts 1& Total Count Automatic Recloser Automatic Recloser Replacement Expenditures 2015 $ 1& Total Automatic Recloser $ 1,413,542 $ 1,084,972 $ 989,787 $ 1,701,173 $ 2,221,022 $ 7,410,497 Automatic Recloser Replacements Unit Costs ' Average Automatic Recloser $ 64,252 $ 67,811 $ 89,981 $ 58,661 $ 74,034 $ 68,616 Unit Cost Used for Forecasting 4& Automatic Recloser $ 74,034 $ 75,432 $ 77,001 $ 78,940 $ 81,417 $ 84,005 1 Annual expenditures and unit counts based on projects that occurred from 2011 to Annual expenditures and units counts based on projects that have been closed (construction completed, accounting and financials have been finalized) 3 Unit cost per year is derived by taking total expenditures and dividing by total unit counts 4 Unit Cost used for forecasting includes escalation 5 Last year recorded was selected because of increasing cost over last five years with the exception of 2014 and last year recorded is a more accurate portrayal of current pricing

215 116 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Automatic Reclosers

216 Workpaper Southern California Edison / 2018 GRC 117 Account 365 OH Circuit Breakers (OH Automatic Reclosers) CPR General Information: Automatic Reclosers are used in the distribution system to sectionalize circuits and isolate faults. We typically install reclosers on overhead circuits but in some situations padmounted units are applied. The automatic reclosers are used in the 4kV, 12kV, 16kV and 33kV levels. The majority of existing field installations are the oil filled design. We have approximately 1050 automatic reclosers installed in the distribution system. The latest designs of these reclosers have undergone extensive changes with solid dielectric insulation and vacuum interrupters along with advanced protective features for the electronic controllers. Oil Filled Recloser Solid Dielectric Recloser Installation Type (3/2010) Installations Overhead Recloser installations 998 Pad-mounted Recloser installations 32 Applications of automatic reclosers vary thoughout the Edison service area. The vast majority of reclosers are installed outside of the major metropolitan areas. Typical applications include sectionalizing circuits for reliability in high wind areas or vegetation areas such as the desert or canyons, breaker for large customer facilities, distributed generation interface, large transformer protection, end of line protection where the substation circuit breaker may not be capable of detecting faults, and increased protection

217 118 Workpaper Southern California Edison / 2018 GRC and sectionalizing for fire hazard areas. Our automatic reclosers are maintained by specialized crews. A failed or inoperable automatic recloser may present high risk for damages. As a primary protective device for critical installations, allowing automatic reclosers to remain in service until failure is not a common practice. While field failures do occur with the equipment, the detailed inspection criteria and replacement strategies should allow proper maintenance and replacement planning to avoid in-service internal equipment failures. For replacement and maintenance purposes, standard recloser installations include additional switches to provide the capability to bypass and electrically isolate the recloser. A 3-phase gang operated switch provides the means for bypassing the recloser during maintenance. The bypass switch can also be operated closed to maintain the powerflow in the event the automatic recloser is faulty. Line side and load side single phase disconnect switches are also installed to provide means to disconnect and isolate the recloser during maintenance intervals. The disconnect switches also provide means for isolating a recloser due to a failure or for replacement. Expected Average Service Lives by Major Component: The expected average service life for automatic reclosers and recloser controllers is approximately years. This is based on engineering judgment and the experience and input from Apparatus crews who maintain and repair automatic reclosers. The new technologies offer reduced overall maintenance on the breaker portion of the automatic recloser. Although with the increased electronic components for the controllers we are seeing a change in the components required for repairing automatic reclosers. The Apparatus test technicians are now required to update firmware setting files and test electronic components as part of many of the inspection cycles.

218 Workpaper Southern California Edison / 2018 GRC 119 Previous controllers were repairable by the component design, where the current standard recloser controllers are essentially computers. Harmonics and surges are very damaging to electronic components, while care is taken to protect the devices from harmful effects of surges, we may see a decrease in service life of the electronic controllers with the increasing electronics within the recloser controllers. For example battery chargers for recloser controllers have been a serviceable and replaceable component, though with new controllers the battery charger is integral to the controller. A failure of the battery charger requires the entire recloser controller to be replaced and can only be repaired by the manufacturer. Major Causes of Retirement: Major causes of retirement for reclosers are failure and obsolescence. Over the past 15 years we have replaced many of the obsolete reclosers that were installed in the 1960 s. These reclosers were specifically identified as the oldest reclosers that were no longer supported by any manufacturers for replacement components. In additional the recloser controllers were not capable of being automated with our radio system for remote monitoring. With the exception of a small number of reclosers that are part of added facilities with specific customers, all of these legacy reclosers have been replaced. The remaining vintage oil reclosers installed in the 1970 s have limited interrupting ratings and continuous current capabilities. Some circumstances require these reclosers to be replaced due to increased fault current levels; however it is more common for the limited continuous current to be inadequate for increased circuit loading. When these reclosers fail they either cannot be repaired or cannot be repaired cost-effectively. As with many oil-filled components, in-service failures can be violent. Our present infrastructure replacement program provides replacement priority to the vintage oil-filled reclosers. The condition of the recloser, age, circuit loading, and relative significance of the circuit are all factors for determining replacement priority. Retirement Frequency by Age: Typical retirement or failure reasons range from excessive corrosion, animal contact, and internal components out of tolerance. The automatic recloser internal components are tested during the maintenance intervals. These tests include trip tests to verify the operating mechanism integrity, oil testing and megger testing for insulating properties, and various tests on the electronic controls to verify proper operation. Since 2002, our standard automatic recloser is the G&W Viper AR. The AR is equipped with a Schweitzer SEL-351R electronic control with remote control and monitoring capabilities. The control provides the ability to monitor line loading and voltage levels,

219 120 Workpaper Southern California Edison / 2018 GRC and is setup for alarming for internal problems such as battery failures or for indicating when the recloser has tripped open from a line fault. Most reclosers on the SCE distribution system are 3-phase interrupting devices. A small 25kV system exists where the protection scheme utilizes single phase reclosers for increased reliability and protection of the long lines in the high wind areas surrounding Palm Springs. Many of these single phase reclosers were installed in the 1960s and 1970s, and we are now noticing an increase in the failure rates for these installations. Around installations exist and our goal is to proactively replace all the existing oil filled hydraulic operated reclosers. The replacement design will employ solid dielectric and vacuum interrupting technology along with a microprocessor based electronic controller that is capable of remote monitoring. To date, there is no available data and information that can be appropriately applied to establish and/or quantify retirement frequency by age for reclosers. Replacement are generally sporadic therefore, reasonable determination to establish retirement frequency by age would be impractical and difficult to achieve. Impacts of Dynamic Environment (economical, political, environmental, etc): The latest recloser designs have undergone extensive changes from the oil filled recloser designs. Vacuum interrupters along with a robust magnetic actuator operating mechanism provide the increased ratings. The solid dielectric insulation help reduce the environmental impact should a unit fail in-service. New recloser installations employ solid dielectric insulation eliminating the insulating oil tank. Oil filled reclosers house approximately 40-gallons of mineral oil, along with an oil filled control power transformer 5-10 gallons. The new solid-dielectric reclosers require less operating power due to mechanism changes, thus allowing the use of smaller solid dielectric controller power transformers. These new designs create an oil-less installation drastically reducing clean-up and containment costs due to oil spill situations. The magnetic actuator mechanism allows for the equipment to be configured in a compact padmounted design. The compact design is being reviewed for new applications to further advance the distribution grid by providing additional sectionalizing and monitoring capabilities. Many recloser applications are in the hot climate of the desert and rural regions. The increased temperatures and increased housing in these areas may contribute to an increase in recloser replacement upgrades and new installations for reliably servicing the customer base in these areas. Many field reclosers are now approaching their expected end of service life. Although we initially installed reclosers in the 1960 s, recloser installations sites increased

220 Workpaper Southern California Edison / 2018 GRC 121 throughout the 1970 s. These 1970s vintage model reclosers are no longer supported by the manufacturer, and repair costs when possible are typically not economical. We can expect to see an increase in the replacement base of automatic reclosers in the upcoming years as these initial vintage reclosers are inspected by our Apparatus Crews and evaluated by our Infrastructure replacement team. Another potential increase in automatic recloser installations follows with the building of a smarter grid. Automatic restoration and fault isolation schemes utilize automatic recloser type devices for interrupting fault current, testing circuits, and transmitting grid information for determining the best switching procedures for restoring the system. There are no policy, legal, or procedural changes, and/or any regulatory mandates that have been or anticipated to be enacted in the near future that would influence the frequencies of retirement and impact the service life of automatic reclosers.

221 122 Workpaper Southern California Edison / 2018 GRC Workpaper Title: Cost of PCB Transformer Replacements

222 Workpaper Southern California Edison / 2018 GRC 123 Work paper Title: COST OF PCB TRANSFORMER REPLACEMENTS WBS Element: CET PD IR PC Witness: Jose Ramon Goizueta This workpaper establishes the average unit cost of replacing a PCB transformer PCB Transformer Replacement Counts 1& Total Count PCB Transformer PCB Transformer Replacement Expenditures 2015 $ 1& Total PCB Transformer $ 816,313 $ 564,389 $ 992,291 $ 1,405,167 $ 1,325,884 $ 5,104,044 PCB Transformer Replacement Unit Costs ' Average PCB Transformer $ 9,604 $ 5,588 $ 4,333 $ 4,475 $ 6,314 $ 5,436 Unit Cost Used for Forecasting 4& PCB Transformer $ 5,436 $ 5,538 $ 5,653 $ 5,796 $ 5,978 $ 6,168 1 Annual expenditures and unit counts based on projects that occurred from 2011 to Annual expenditures and units counts based on projects that have been closed (construction completed, finalized accounting and financials) 3 Unit cost per year is derived by taking total expenditures and dividing by total unit counts 4 Unit Cost used for forecasting includes escalation 5 Five year weighted average was selected for forecasting due to fluctuations in cost over the last five years

223 124 Workpaper Southern California Edison / 2018 GRC Workpaper Title:

224 Workpaper Southern California Edison / 2018 GRC 125 Sub IR Program - Process to Develop 5 th Year IR (and Adjustments to 3 rd and 4 th Year for all Sub IR programs and to PMWIF/Pre-Engineering Job Walk Phase April 1.1 Sub IR Tech Team Produce & Review Monthly Priority List Based on Health Index & Eco Life Model for CB's & Transformers Develop 5 Year Plan for Each Program & Adjust 3 to 5 Year As Needed Review/Approve or Deny & Adjust 3-5 Year Plan & Approve 5 th Year & Rename as Revision 2 Sub IR Plan 2.1 CBM Risk & Strategy PM Send Approved Revision 2 Sub IR Plan to Stakeholders/TIP & Non Represented Stakeholders Review/Approve & Rename as Revision 3 Sub IR Plan/ & Document Input & Decision for Future GRC Testimony & Data Revise Sub IR Plan with Stakeholders Adjustments & Send Revision 3 Sub IR Plan to Sponsor Grid Operations Electric System Planning Protection Engineering Apparatus Engineering Substation Construction & Maintenance (SC&M) Sponsor When developing/adjusting/approving IR plan: Identify recent O53/emergency replacements Identify bundled projects from other sponsoring organizations Identify affected/potential licensed projects Document justification for all IR adjustments for future GRC testimony and data requests April 1 -April Sub IR Tech Team 1.3 Sub IR Tech Lead SC&M SME Meet with Sub IR Tech Team & Stakeholders & Determine Scope Perform Sizing Analysis for CB & Transformer 1.5 SC&M Sub IR Tech Team SC&M SME Receive & Coordinate Tasks Results & Forward to Sub IR Tech Team 1.4 Sub IR Tech Team & Stakeholders Assign Tasks If Required & Return to SUB IR Technical SC&M SME 1.6 Sub IR Tech Team May 1 -May 31 June 1 June Sub IR Tech Team Communicate Approved Revision 2 Sub IR Plan to Apparatus Engineering to Determine Bundle Work &/or Request Adjustment if Required 2.2 Sub IR Tech Team Communicate Approved Revision 2 Sub IR Plan to Protection Engineering to Determine Bundle Work &/or Request Adjustment if Required 2.4 Sub IR Tech Team SC&M SME July 1 July 31 Gather & Organize All Revisions 2 Sub IR Plan & Recommendations 2.3 Sub IR Tech Team Approve Rev 2 Sub IR Plan Review /Approve or Deny Revision 2 Sub IR Plan & Send Recommendations to Tech Team SC&M SME Deny Rev 2 Sub IR Plan 2.5 Sub IR Tech Team August 1 August 20 When adjusting/approving IR plan: Identify recent O53/emergency replacements Identify bundled projects from other sponsoring organizations Identify affected/potential licensed projects Document justification for all IR adjustments for future GRC testimony and data requests If Denied No Further Action 3.1 Sub IR Tech Team 3.2 Sub IR Sponsor Approve Sub IR Plan Revision 3 & Send to Apparatus PM July 1 July Sub IR Tech Team Communicate Approved Rev. 2 Sub IR Plan to ESP to Determine Bundle Work &/or Request Adjustments, if Required FE and Grid Modernization/ Reinforcement July 1 July Sub IR Tech Team Communicate Approved Revision 2 Sub IR Plan to Grid Operations & Determine Bundle Work &/or Request Adjustment if Required 1 Thursday, July 28, 2016

225 126 Workpaper Southern California Edison / 2018 GRC (Cont d) Sub IR Program - Process to Develop 5 th Year IR Plan (and Adjustments to 3 rd and 4 th Year) for all Sub IR Programs to PMWIF/Pre-Engineering Job Walk Phase Note: Depending on Asset Type, Issuance of PMWIF is 2-4 Years Prior to Operating Date (OD) of Project 3.2 Sub IR Sponsor August 1 August 20 Sub IR Sponsor approves Technical Team(s) recommended plan. (Rev. 3) 3.6 Sub IR Sponsor & Tech Team 3.1 Approve or Deny Changes from Scoping Job Walk & Rename as Final Sub IR Plan Final Sub IR Denied No If Denied No Action Substation Engineering Grid Operations Apparatus PM s (David and Angel) TS&O Programs & Strategy Project & Planning Information Management Substation Construction & Maintenance (SC&M) Sponsor Substation Construction & Maintenance (SC&M) Sponsor 4.2 Project & Planning Information Management Review & Validate Project Summary for T&D Visibility & Resource Allocations 4.3 Project & Planning Information Management March 1 March 31 Conduct Regional Project Coordination Meetings to Optimize Scheduling & Bundling Opportunities 4.4 Project & Planning Information Management Issue Project Management Work Initiation Forms (PMWIF) 4.5 Project & Planning Information Management Issue PMWIF s for IR Projects to Programs & Strategy (P&S) Yes 4.10 SC&M Sponsor Approve or Deny Scope Change Request (SCR) Yes Deny Scope Change Request No Submit Sub IR Request Form for Future Year 4.6 TS&O Program Strategy Receive & Sign PMWIF s/ Execute & Project Manage Scope, Schedule & Budget. 4.7 TS&O Program Strategy Issue Authorization to Proceed (ATPs) to Substation Engineering TS&O Programs & Strategy Modify & Document Project Scope, Schedule & Budget as Necessary 3.3 August 21 September 30 Apparatus PM Modify Project Summary for Approved Revision 3 & Input to Project Summary Interface (PSI) 3.4 Apparatus PM October 1 December 31 (Year Ahead) Schedule Scoping Job Walks for Approved Projects 3.5 Apparatus PM Review Findings of Job Walk & Submit Revision of Sub IR Plan to Sponsor 4.1 Apparatus PM Modify Approved Project Summary of Final Sub IR Plan & Forward to Project Planning Yes 4.9 Apparatus PM Receive & Track SCRs & Communicate SCR Decision to Sub IR Tech Team & Appropriate 4.8 Substation Engineering Conduct Pre-Engineering Job Walk & Determine & Submit Scope Change Request (SCR) if Applicable SCR Required? No 4.12 Substation Engineering Engineer & Procure Material 2 Thursday, July 28, 2016

226 Workpaper Southern California Edison / 2018 GRC 127 Workpaper Title: Cost of Transformer Bank Replacements

227 128 Workpaper Southern California Edison / 2018 GRC Work paper Title: COST OF TRANSFORMER BANK REPLACEMENTS WBS Element: CET ET IR TB Witness: Jose Ramon Goizueta This workpaper establishes the average unit cost of replacing a transformer bank Total Count 12kV kV kV kV kV kV TR Replacement Expenditures 2015 $ 1& Total 12kV $ 707,027 $ 583,764 $ 5,191,051 $ 1,091,138 $ 5,100,040 $ 12,673,020 16kV $ 1,114,426 $ 3,599,292 $ 5,403,893 $ 5,750,720 $ 3,302,049 $ 19,170,379 33kV $ 6,167,639 $ 2,871,788 $ 4,182,477 $ 5,661,022 $ 2,341,879 $ 21,224,805 66kV $ 16,729,216 $ 14,296,972 $ 23,116,183 $ 16,755,097 $ 24,005,796 $ 94,903, kV $ 15,189,611 $ 9,050,424 $ $ 6,116,493 $ 8,004,890 $ 38,361, kV $ 15,629,315 $ 9,528,486 $ 20,988,081 $ 23,173,009 $ 10,934,782 $ 80,253,672 TR Replacements Unit Costs ' Average 4kV $ 707,027 $ 583,764 $ 1,297,763 $ 1,091,138 $ 1,275,010 $ 1,152,093 12/16kV $ 557,213 $ 1,199,764 $ 1,350,973 $ 1,437,680 $ 1,100,683 $ 1,198,149 33kV $ 1,027,940 $ 957,263 $ 2,091,239 $ 943,504 $ 1,170,939 $ 1,117,095 66kV $ 1,286,863 $ 1,299,725 $ 1,284,232 $ 1,117,006 $ 1,200,290 $ 1,232, kV $ 1,518,961 $ 1,508,404 $ $ 1,529,123 $ 2,001,223 $ 1,598, kV $ 3,907,329 $ 4,764,243 $ 4,197,616 $ 7,724,336 $ 5,467,391 $ 5,015,855 Unit Cost Used for Forecasting 4& kV $ 1,152,000 $ 1,173,760 $ 1,198,164 $ 1,228,340 $ 1,266,874 $ 1,307,158 12/16kV $ 1,198,000 $ 1,220,628 $ 1,246,008 $ 1,277,388 $ 1,317,461 $ 1,359,354 33kV $ 1,117,000 $ 1,138,098 $ 1,161,762 $ 1,191,021 $ 1,228,384 $ 1,267,444 66kV $ 1,233,000 $ 1,256,290 $ 1,282,410 $ 1,314,708 $ 1,355,951 $ 1,399, kV $ 1,598,000 $ 1,628,184 $ 1,662,037 $ 1,703,895 $ 1,757,348 $ 1,813, kV $ 5,016,000 $ 5,084,771 $ 5,197,762 $ 5,325,943 $ 5,466,215 $ 5,609,480 1 Annual expenditures and unit counts based on projects that occurred from 2011 to Annual expenditures and units counts based on projects in various states which include closed projects (construction completed, accounting and financials have been finalized) and projects that are in service (construction completed but accounting and financials have not been finalized) 3 Unit cost per year is derived by taking total expenditures and dividing by total unit counts for voltage classes 4 Unit Cost used for forecasting includes escalation 5 Five year weighted average was selected for forecasting due to fluctuations in cost over the last five years

228 Workpaper Southern California Edison / 2018 GRC 129 Workpaper Title: Substation Transformer Reliability Model

229 130 Workpaper Southern California Edison / 2018 GRC Substation Transformer Reliability SCE T&D System and Reliability Strategy System & Reliability Strategy Engineering & Technical Services Transmission & Distribution Substation Transformer Reliability Model July 2016 REPORT SRS Prepared by: (Z. S. Roldan) Page 1 of 11

230 Workpaper Southern California Edison / 2018 GRC 131 Substation Transformer Reliability SCE T&D System & Reliability Strategy I OBJECTIVE Develop a risk-informed methodology reliability model for substation transformers A and B class in forecasting the wear-out rate and the relationship between age and reliability. II APPROACH The approach employs risk-informed methodology which includes a combination of probabilistic technique, life data analysis and engineering information. The initial stage of the engineering analysis was an evaluation of the historic removals and failures, and of the transformers currently in-service in order to develop a relationship between age and the probability of in-service failure. This process also include data analysis and interviews with engineers and maintenance experts. DATA COLLECTION AND ANALYSIS The development of a substation transformer reliability model begins with the task of assessing its inventory, (i.e., how much of what vintage) and its failure history by collecting data and gathering information from transformer experts. Primarily the data comes from the SCE SAP Master Database which house the equipment inventory and removal history. The Substation Construction and Maintenance (SC&M) has an event site called the PCR (Preliminary Cause Reports) that records events that occur in the substation e.g. transformers, circuit breakers, relays, capacitors, and other equipment inside the fence. The PCR site is valuable in finding the equipment probable cause analysis reports and in determining transformer failure cause. The Transformer Oil Analysis (TOA) database is valuable in validating the transformer data currently in-service, and its failure history. Review of the transformer failure data is conducted routinely with the engineering group and the SC&M group experts. 2 of 11

231 132 Workpaper Southern California Edison / 2018 GRC

232 Workpaper Southern California Edison / 2018 GRC 133

233 134 Workpaper Southern California Edison / 2018 GRC Substation Transformer Reliability SCE T&D System & Reliability Strategy HISTORICAL DATA INFORMATION Engineering and SC&M groups have both maintained information on substation failure historical data. In addition the removal data and failure events are available from SCE s SAP Master Database and SC&M PCR site. The review and interview conducted by the SRS (System & Reliability Strategy) group with Engineering and SC&M experts concluded that the SAP data is useful base reference for the model development in addition to the corroboration with the engineers and maintenance experts. The SRS, Apparatus Engineering and SC&M, developed a single cohesive substation transformer historical failure database. The records from engineering and SC&M were combined, reviewed in detail and were validated for the use of this analysis and also for future reference. This historical database will be updated as events occur. Table 1 and 2 Show the Removal History of B and A-Bank Transformers Table 1 Transformer Removal Data B-Transformers Row Labels Breakdown CUTOVER 3 Load Growth Repaired Replacement Retirement Temporary Bank 1 1 UNKN Xfmr IR REPL Grand Total Table 2 Transformer Removal Data A-TRANSFORMERS Row Labels Breakdown Load Growth 3 Repaired Replacement 1 UNKN Xfmr IR REPL Grand Total of 11

234 Workpaper Southern California Edison / 2018 GRC 135 Substation Transformer Reliability SCE T&D System & Reliability Strategy MODEL DEVELOPMENT The development of this substation transformer reliability model can be facilitated by the use of commercially available statistics software. Such tools are extremely helpful in generating a consistent set of age-dependent component failure probabilities, verifying data entry, updating data, and developing documentation. SCE utilized one of these products and selected the Weibull distribution as the one of the best in meeting its current objectives. WEIBULL BASED SUBSTATION TRANSFORMER RELIABILITY MODELLING With the establishment of age dependent substation transformer failures and total number of transformer in-service in each age bracket, the corresponding wear out based component failure rate can be established. Integrating this age dependent result with non-age related random failure, total transformer failure rate as function of age can be generated The Reliasoft Synthesis 10, Weibull ++10 software package developed by Reliasoft is chosen to generate the needed probabilistic failure distributions. The approach in the life data analysis is based on transformers that are in-service, the breakdown history and the transformers that are considered to be in imminent failure (or near term) based on its Health Index assessment. The following are the composition of the life data analysis: "F" state or failure group 1) Transformers removed from service due to breakdown. 2) Transformers removed from service due to infrastructure replacement (IR) with health index that considered poor condition (or < 50% HI) or imminent failure 3) Transformers removed from service due to load growth with health index that considered poor condition (or < 50% HI, imminent failure) 4) Active transformers with health index less than 50% (imminent failure or poor condition) "S" state or suspension or ACTIVE group 5) Active transformers as per the histogram. 6 of 11

235 136 Workpaper Southern California Edison / 2018 GRC Substation Transformer Reliability SCE T&D System & Reliability Strategy Table 3 Weibull Analysis Assumptions B-Bank Analysis Assumptions Count Ave Age Breakdown ( ) IR with HI <50% Imminent Failure Active TX with HI <50% Imminent Failure Load Growth with HI <50% Imminent Failure Table 4 Weibull Analysis Assumptions A-Bank Analysis Assumptions Count Ave Age Breakdown IR with HI <50% Active TX with HI <50% of 11

236 Workpaper Southern California Edison / 2018 GRC 137 Substation Transformer Reliability SCE T&D System & Reliability Strategy III RESULTS The results of the analysis yielded statistical values as shown in Table 5. The A and B Bank Transformer s age dependent failure rate and unreliability function are shown in figures 3 and 4. Table 5 - SCE s Substation Transformers and Statistics. Substation Transformers YE2015 A B Spares Ground Reactors Number of Xfmrs Aver Age Eta Beta MTTF(mean time to failure), Year NOTES: 1) Reactor inventory is as of YE2014 2) Average ages for spares, ground and reactors are based on YE Phase TX Phase TX of 11

237 138 Workpaper Southern California Edison / 2018 GRC Substation Transformer Reliability SCE T&D System & Reliability Strategy Figure 3 - Probability of Failure vs Age for B-Bank Transformers ReliaSoft Weibull Failure Rate vs Time Plot Failure Rate Data 1: B TX Weibull-2P RRX SRM MED FM F=45/S=263 Failure Rate Line F a ilu re Ra te, f( t) / R( t) ZOILO ROLDAN SOUTHERN CALIFORNIA EDISON, CO. 10/14/ :29:45 PM Time, ( t) 9 of 11

238 Workpaper Southern California Edison / 2018 GRC 139 Substation Transformer Reliability SCE T&D System & Reliability Strategy Figure 4 - Probability of Failure vs Age for A-Bank Transformers ReliaSoft Weibull Failure Rate vs Time Plot Failure Rate Data 1 Weibull-2P MLE SRM MED LRB F=5/S=13 Failure Rate Line F a ilu re Ra te, f( t) / R( t) ZOILO ROLDAN SOUTHERN CALIFORNIA EDISON, CO. 3/3/2016 6:52:48 AM Time, ( t) 10 of 11

239 140 Workpaper Southern California Edison / 2018 GRC Substation Transformer Reliability SCE T&D System & Reliability Strategy Table 6 and 7 shows the 10 year Forecast of Transformer Wear-Out Rate. By multiplying the probability of failure by the number of transformers in each age group, the number of transformers expected to reach the end of their service lives over the next ten years can be determined. Table 6 Forecast B-Bank Transformer Wear-Out Rate B-Transformers Voltag e Class KV kV kV kV kV Total Table 7 Forecast A-Bank Transformer Wear-Out Rate A Transformers Voltage Class A-Bank of 11

240 Workpaper Southern California Edison / 2018 GRC 141 Workpaper Title: A-Bank Health Index

241 142 Workpaper Southern California Edison / 2018 GRC Health Index for Transformer A Banks Workpaper The table below is a summary of A bank transformers that are included in SCE s 2018 GRC application. This includes a condition code from a recent inspection. Each condition code is a result of a specific test that is a characteristic of the underlying degradation process that test measures. Provided below the table is a description of each column heading and where applicable a brief description of the specific test and associated condition level per SCE Maintenance and Inspection Manual (MIM) for Substation Construction and Maintenance (SC&M). Health IWP Position Effective Oil LTC LTC Winding Substation TOA Index OD Number Age Quality OTA DGA %PF SME Parts Springville 2A-A Springville 1A-A Springville 1A-C Etiwanda 6A Springville 2A-B Springville 1A-B San Bernardino 2A Springville 2A-C Villa Park 1A Villa Park 3A Center 2A Center 3A La Fresa 2A La Fresa 4A Alamitos 1A-A Alamitos 1A-B Alamitos 1A-C Alamitos 2A-A Alamitos 2A-B Alamitos 2A-C Hinson 1A Ellis 2A La Fresa 3A Mirage 1A Ormond Beach 2A Laguna Bell 1A Laguna Bell 3A Villa Park 2A Eagle Mountain 3A Lighthipe 1A Lighthipe 2A El Nido 3A Rector 4A Santiago 4A Health Index for A Bank Transformers.docx Page 1 of 6

242 Workpaper Southern California Edison / 2018 GRC 143 Health Index Definition: The aggregated estimate of deterioration based on condition. Condition is estimated based on test result and assigned a condition code level between 1 and 4, with 4 being the most critical. The health index score is not a relative ranking of all A bank but a weighted score of condition. More than one asset can have the same health index score. Additional Health Index Information The Health Indexing of an asset class represents the quantification of a condition assessment. Results from inspections, test and maintenance activities are categorized and weighted based on confidence factors and the underlying degradation process that they measure. The contribution of each poor test result to deterioration is aggregated multiplicatively to estimate the deterioration of the underlying degradation process. Process Deterioration = Process Confidence * (1-(1-Deterioration_Test1)*(1-Deterioration_Test2)*(1-Deterioration_TestN ) The deterioration of each degradation processes (Solid insulation, liquid insulation winding, core, bushings, LTC etc.) is then aggregated multiplicatively in a similar manner to estimate the overall deterioration of the asset. In the example below, Transformer Oil Analysis (TOA) analysis is indicating that there is high amounts of hot metal gas (Ethane C2H6) deterioration in the Winding degradation process. This test has 80% confidence and suggests 80% deterioration based on the confidences shown in pink and the overall winding deterioration. Additional test result and confidence % for the overall condition assessment is below in figure: 1 Alamitos No.1A Bank B Phase example degradation inputs. IWP OD Integrated Work Plan Operating Date or date construction is expected to be completed. Substation The name of the substation where the transformer is located. Health Index for A Bank Transformers.docx Page 2 of 6

243 144 Workpaper Southern California Edison / 2018 GRC Position Number The name designation for the transformer at a specific substation location. Effective Age This is the calendar age of the transformer plus any age adjustments base on condition assessments. Effective Age Adjustments based on the Health Index formulation The Health Index is compared with the Calendar Age of the asset to determine whether this asset is aging as expected. If it is found to have worse condition than expected based on the Calendar Age, the model uses an Effective Age for the asset that is adjusted upwards. However, while poor condition is generally a strong predictor of asset failure for most asset types, good condition is not typically a strong predictor of non-failure. In the Alamitos No.1A Bank B Phase example degradation inputs below, the calendar age of the asset is 59 years old, the health index formulation using actual conditions age adjust of 37 years for an effective age of 96. TOA Transformer Oil Analysis (TOA) analysis, Dissolved gas analysis (DGA), provides comprehensive data to help determine the health of the transformer while in service. Gasses are formed in oil when the insulation system is subjected to thermal, electrical, and mechanical Health Index for A Bank Transformers.docx Page 3 of 6