2018 General Rate Case

Transcription

1 Application No.: A.1-0- Exhibit No.: SCE-0, Vol. 1 (Appendix) Witnesses: R. Woods (U -E) 01 General Rate Case Transmission & Distribution (T&D) Volume 1 Appendix to Operational Overview and Risk-Informed Decision Making Before the Public Utilities Commission of the State of California Rosemead, California September 1, 01

2 SCE-0: Transmission & Distribution Volume 1 Appendix Table Of Contents Section Page Witness A. Review of Key Risk Model Formulas, Terms, and Concepts...1 R. Woods 1. Glossary of PRISM Terms...1. Risk Scoring Formula.... Current Residual Risk and Mitigation Risk Reduction.... Worst Reasonable Direct Impact (WRDI).... Data & Modeling... B. Risk Modeling Analysis Details Overhead Conductors... a) Risk Identification... b) Current Residual Risk Evaluation... (1) Triggering Event Frequency (TEF)... () Consequence Percentage (CP) and Consequence Impact (CI)... c) Mitigation Alternatives Identification... d) Mitigation Risk Reduction Evaluation... (1) Tranche Analysis... () TEF and CP Effectiveness... e) Risk Spend Efficiency...1 f) Key Takeaways...1 (1) Data...1 () Modeling...1 () Observation(s)...1 -i-

3 SCE-0: Transmission & Distribution Volume 1 Appendix Table Of Contents (Continued) Section Page Witness. Poles...1 a) Risk Identification...1 b) Current Residual Risk Evaluation...1 (1) Triggering Event Frequency...1 () Consequence Percentage (CP) and Consequence Impact (CI)...1 c) Mitigation Alternatives Identification...0 d) Mitigation Risk Reduction Evaluation...1 (1) Pole Replacement...1 () Steel Stubbing...1 () Guy Wire Repair / Replacement... () E-Truss... e) Risk Spend Efficiency... f) Key Takeaways... (1) Modeling... () Observation(s).... Underground Structures... a) Risk Identification... b) Current Residual Risk Evaluation... (1) Triggering Event Frequency (TEF)... () Consequence Percentage (CP) and Consequence Impact (CI)... c) Mitigation Alternatives Identification... d) Mitigation Risk Reduction Evaluation... -ii-

4 SCE-0: Transmission & Distribution Volume 1 Appendix Table Of Contents (Continued) Section Page Witness (1) Tranche Analysis... () TEF and CP Effectiveness...0 e) Risk Spend Efficiency...1 f) Key Takeaways...1 (1) Data...1 () Modeling... () Observation(s).... Circuit Breakers... a) Risk Identification... b) Current Residual Risk Evaluation... (1) Triggering Event Frequency (TEF)... () Consequence Percentage (CP) and Consequence Impact (CI).... c) Mitigation Alternatives Identification... d) Mitigation Risk Reduction Evaluation... (1) Tranche Analysis... () TEF and CP Effectiveness... e) Risk Spend Efficiency... f) Key Takeaways... (1) Modeling.... Transformers... a) Risk Identification... b) Current Residual Risk Evaluation... -iii-

5 SCE-0: Transmission & Distribution Volume 1 Appendix Table Of Contents (Continued) Section Page Witness (1) Triggering Event Frequency (TEF)... () Consequence Percentage (CP) and Consequence Impact (CI)... c) Mitigation Alternatives Identification... d) Mitigation Risk Reduction Evaluation...0 (1) Tranche Analysis...0 () TEF and CP Effectiveness...0 e) Risk Spend Efficiency...0 f) Key Takeaways...1 (1) Modeling...1. Underground Cables...1 a) Risk Identification... b) Current Residual Risk Evaluation... (1) Triggering Event Frequency (TEF)... () Consequence Percentage (CP) and Consequence Impact (CI)... c) Mitigation Alternatives Identification... d) Mitigation Risk Reduction Evaluation... (1) Tranche Analysis... () TEF and CP Effectiveness... e) Risk Spend Efficiency... a) Key Takeaways... (1) Modeling.... kv Circuits... -iv-

6 SCE-0: Transmission & Distribution Volume 1 Appendix Table Of Contents (Continued) Section Page Witness a) Risk Identification... b) Current Residual Risk Evaluation... (1) Triggering Event Frequency (TEF)... () Consequence Percentage (CP) and Consequence Impact (CI)... c) Mitigation Alternatives Identification... d) Mitigation Risk Reduction Evaluation... (1) Tranche Analysis... () TEF and CP Effectiveness... e) Risk Spend Efficiency...1 f) Key Takeaways... (1) Modeling.... Vegetation Management... a) Risk Identification... b) Current Residual Risk Evaluation... (1) Triggering Event Frequency (TEF)... () Consequence Percentage (CP) and Consequence Impact (CI)... c) Mitigation Alternatives Identification... d) Mitigation Risk Reduction Evaluation... (1) Tranche Analysis... () TEF and CP Effectiveness... e) Risk Spend Efficiency... f) Key Takeaways... -v-

7 SCE-0: Transmission & Distribution Volume 1 Appendix Table Of Contents (Continued) Section Page Witness (1) Data... () Modeling.... Conclusion... -vi-

8 I. APPENDIX A. Review of Key Risk Model Formulas, Terms, and Concepts Prioritized Risk Informed Strategic Management (PRISM) leverages an event based methodology. Events can lead to a number of negative outcomes, which can be classified in consequence dimension. Outcomes are stratified into different levels of consequence impact. The following provides a summary overview of key terms, formulae, and concepts related to the PRISM framework that are referenced throughout this volume. 1. Glossary of PRISM Terms Acronym Term Definition Asset T&D physical assets (poles, wires, substations, etc.) AST Asset Strategy Teams formed with subject matter experts to identify risks, drivers, mitigations, and data sources for a specific asset. Team (AST) CS (1,,) Challenge Sessions held to calibrate and validate scoring within (CS1) and across CI CP CRR MRR Sessions Consequence Consequence Impact Consequence Percentage Current Residual Risk Driver Event Impact Dimension Mitigated Risk Reduction Mitigation Outcome (CS) Asset Strategy Teams, as well as with senior leadership (CS) The specific Impact Dimension within the Risk Evaluation Tool (e.g. Safety, Reliability, Financial, etc.) that is associated with an Outcome (e.g. injury, wildfire, property damage, etc.) Magnitude or severity of impact measured on 1 to scale Probability that a consequence occurs when triggering event occurs Measure of risk units without the proposed mitigation in place. CRR inherently includes the benefits of past mitigations. A higher CRR means more risk Contributing factors causing an event or consequence Occurrence or change of a particular set of circumstances Categories created to capture the different types of consequences that a risk event may have (Safety, Reliability, Environmental, Financial, and Compliance) Measure of risk units that are reduced associated with a given mitigation project or program. Also known as the benefit of a proposed mitigation project; a higher MRR means more risk mitigated. An action that would result in lowering an asset s Risk Score, either through lowering the frequency of triggering events (TEF), the likelihood of the consequence to occur (CP) or the consequence impact s severity (CI) A negative consequence that occurs as a result of a risk event or triggering event (e.g. injury, wildfire, property damage, etc.). Outcomes can have consequences in one or more impact dimension 1

9 PRR PST Planned Residual Risk Program Strategy Team Risk (e.g. safety, reliability, financial, etc.) Measure of risk units after implementation of the proposed mitigation Team formed with subject matter experts to identify the proper technique to score current risks and mitigations for programs that cross multiple asset classes. Potential for an adverse event that can impact company's ability to achieve its objectives RS Risk Score A numerical representation of a risk, defined as a measure of frequency and impact measures & resulting in a measure of risk units ; Mathematically defined as (TEF x CP) x I CI (Currently, I = ) RSE TEF WRDI Risk Spend Efficiency Risk Statement Risk Units Tranche Triggering Event Triggering Event Frequency Worst Reasonable Direct Impact A numerical measure of the efficiency with which a mitigation mitigates against a risk; Mathematically defined as the ratio of expected risk reduction (MRR) per $1,000. A higher RSE the more efficient a program or asset is at mitigating risk An event that leads to an outcome (e.g. Cable failure leading to an outage ) The measure of risk scores used in PRISM, which is a function of risk frequency & impact A subgroup of assets, grouped together based on similar risk profiles A risk event that could potentially lead to one or more negative outcomes, each of which could have one or more consequences Number of times a triggering event is expected within a one year timeframe The risk score associated with the highest risk score within a given risk statement, separated by various CP/CI values (e.g. for the risk statement Cable failure leading on an outage, the WRDI risk score splits out the observed outages by CI level and takes the CP* CI combination that leads to the highest risk score 1. Risk Scoring Formula The formula used to compute risk is: = TEF is the triggering event frequency, CP is the Consequence Percentage, and CI is the Consequence Impact. Triggering events relate to circumstances that can lead to undesirable outcomes. TEF is measured on a continuous frequency scale and represents that annual event frequency. The consequence percentage refers to the conditional probability that an undesirable outcome occurs when the triggering event occurs. Lastly, the Consequence Impact measures the impact severity on a one to seven scale. For example, an overhead conductor down in service can be a triggering event with a forecast annual frequency. When a wire down event does occur, some percentage of the time (CP value) it results in an

10 injury outcome. When the injury outcomes occur the worst reasonable impact severity is represented by the Consequence Impact (CI).. Current Residual Risk and Mitigation Risk Reduction The Current Residual Risk (CRR) represents the computed risk score prior to implementation of a proposed mitigation. The CRR value is often determined based on inputs that rely on historical event data, which inherently includes the risk reduction benefits of past mitigations. Therefore, CRR values also include the benefits of past mitigations. CRR values can be represented as 1-year CRR (CRR1) scores or present value CRR scores (CRRPV). One-year CRR scores only account for the risk in a single year while present value CRR scores represent the entire stream of yearly CRR values over a 1-year window discounted to present value. For example, if a risk has a constant CRR value of 0 risk units over a 1 year window, the CRR1 value for the present year would be 0 but the CRRPV of the entire stream of risk discounted back to present would be 1. 1 Mitigation Risk Reduction (MRR) represents the risk reduction benefit associated with a mitigation. Risk reduction can stem from a mitigation s ability to reduce the triggering event frequency, the conditional probability the event would result in an outcome, the severity of an outcome, or some combination of the three. For example, replacing overhead conductor with new and larger wire can reduce wire down events (TEF) but new conductor does not impact whether the wire stays energized if it does come down. On the other hand, advanced relays that can detect a hard to detect high-impedance fault can de-energize a downed line and reduce the probability that someone is injured when a line comes down (CP) but may not be able to prevent the original wire down event. Lastly, fire walls between substation transformers is an example of a mitigation that can reduce the impact (CI). The walls will not prevent an in-service failure of a transformer (TEF) or the likelihood it will result in an outage (CP), but it can prevent the fire from spreading to neighboring transformers that could increase the size of the outage (CI). Current Residual Risk and Mitigation Risk Reduction are both measured over time as discussed in the volume. The present value of mitigation benefits over the 1-year window are denoted by the term MRRPV. 1 Using a risk over time discount rate equal to percent. Refer to SCE-0, Vol. 1, Section III.D..b).

11 Worst Reasonable Direct Impact (WRDI) PRISM currently uses a Worst Reasonable Direct Impact (WRDI) assumption that selects the highest scoring combination of Consequence Percentage and Consequence Impact for each Risk Statement. Consider the following risk statement as an example: overhead conductor down leading to an outage. Historical data shows that wire down events lead to outages that range in severity, but the Consequence Impact Level (CI) outage outcome yields the highest risk score, so the WRDI assumption excludes scores for other CI levels. This approach provides a consistent method for scoring across areas with varying levels of data. For example, scoring risks with Reliability outcomes is supported by the robust Outage Database & Reliability Metrics System (ODRM) that provides a comprehensive record for every customer outage. On the other hand, scoring risks with Safety outcomes is supported by CPUC Reportable Incident records, which may not include close call outcomes. The WRDI assumption is intended to reflect scores in both areas, without under-valuing scores in areas like Safety that have less comprehensive datasets.. Data & Modeling We used a variety of data sources to build the PRISM risk scores. The sources we used for the scoring areas covered in the testimony are below in Table I-1. Rule 1 and 1 require an incident to be recorded as a CPUC Reportable Incident if the outcome results in any of the following criteria: a significant injury leading to hospitalization or a fatality; an outage for % or more of SCE customers, greater than $0k in damage, or a newsworthy event.

12 Table I-1 Data Sources by Scoring Area 1 B. Risk Modeling Analysis Details As discussed in the SCE-0 Vol 1, SCE has made significant efforts to advance its risk evaluation capabilities by developing a quantitative risk-based scoring methodology. This section provides analysis details for selected scoring areas, and is intended to illustrate how SCE s risk framework has been applied to specific assets and activities. Risk models are continuously improving through each cycle and the information being presented is based on work through March Overhead Conductors The risk analysis for overhead conductors covers all distribution primary conductor and Overhead Conductor Program (OCP) existing mitigations and mitigations in consideration for future

13 1 use. Overhead Conductor Program mitigations focus on reducing the frequency and impact of wire down events; please see SCE-0, Volume for more details. SCE first described its proposed methodology for assessing and quantifying overhead conductor risks in its Safety Model Assessment testimony filed in May of 01. Overhead conductor risk statements have not changed, but the methodology to estimate the impact of OCP mitigations has been updated. a) Risk Identification The risk event-outcome-impact combinations for overhead conductor risks are summarized in Table I-. We evaluate overhead conductor risks in relation to three possible events. Conductor down in service relates to conductor falling the ground. Intact conductor failure in service relates to incidents where the conductor fails due to faults or external causes, but remains intact overhead. Finally, human contact with intact conductor refers to incidents in which people such as tree trimmers or construction workers come in contact with in-service conductor. Table I- Overhead Conductor Risk Statements RS# Triggering Event Outcome Impact Dimension 1 Injury Safety Financial Overhead conductor down in service Wildfire Safety Financial Environmental Property Damage Safety Financial Outage Reliability Freeway/Road Closure Financial Intact conductor failure in service Outage Fatality Reliability Safety Human contact with intact conductor Injury Safety Financial A , SCE s Safety Model Assessment Proceeding Application filed May 1, 01.

14 These risks are intended to capture the consequences of risks triggered by overhead conductor events even if the outcome is not directly related to overhead conductor failure. For example, in a human contacting intact conductor incident, the outcome is associated with overhead conductor. In contrast, the outcomes resulting from car hit pole events are triggered by the pole asset and are accounted for in pole risks. b) Current Residual Risk Evaluation We determined Risk Score inputs from historical data, prior to completing any OCP work. As an initial assumption we estimate CRR for conductor down events with the assumption that historical values represent the asset risk without mitigation programs in place. Overhead conductor CRR scores are summarized in Table I-. CRR score tables like the one below show single-year CRR scores by risk statement. Each risk statement represents a unique triggering event outcome pair. The table also shows these scores broken out by risk dimension to facilitate comparisons of scores by risk statement and risk dimension. For example, in Table I- below, it is possible to determine that human contact with intact conductor leading to injury is the largest risk contribution for the overhead conductor asset, and that the safety risk dimension drives this score. Table I- Overhead Conductor CRR Scores 1-Year Current Residual Risk (CRR) Triggering Event Outcome Safety Reliability Financial Environmental Compliance Subtotal % of Total Injury,000 -, ,0 1% Overhead conductor down Wildfire 1, - 1, 1, -, 1% Property Damage 1, ,1 0% Outage - 0, ,1 % Freeway/Road Closure - -,1 - -,1 0% Intact conductor failure Outage 1,000, ,1 % Human contact with intact conductor Injury,0,000,00 - -,,00 % Subtotal,0,,0,0 1, -,1,1 % of Total % 1% 1% 0% 0% (1) Triggering Event Frequency (TEF) Overhead conductor analysis accounts for three Triggering Events, each of which relies on different data sources and estimates. - Overhead conductor down (aka wire down): SCE started systematically recording wire down events across all regions in May 01 with Troubleman repair orders as the primary source. TEF

15 is the annualized average of wire down incidents documented from May 01 to December 01 for lines classified as primary conductor. - Intact conductor failure: The Outage Database & Reliability Metrics system (ODRM) is the primary data source used to estimate TEF for intact conductor failures. ODRM data from 00 to 01 is filtered to outages with overhead equipment cause code descriptions. It is assumed that all wire down events result in an outage, so the difference in annual overhead conductor caused outages and wire down events is assumed to be the number of intact conductor failures per year. - Human contact with intact conductor: A review of CPUC Reportable Incidents from 00 to 01 is relied on as the primary source to estimate the number of injuries per year that occur from people coming in contact with in-service conductor. () Consequence Percentage (CP) and Consequence Impact (CI) Data sources and methodologies for CP and CI estimates vary by outcome and impact dimension. - Injury: CPUC Reportable Incidents from 00 to 01 provide counts for the number of incidents in which all three overhead conductor triggering events result in serious injury or fatality. - Wildfire: CPUC reportable incidents and Cal Fire Data provide records for wildfires with information on acreage damaged. Recent fires in SCE territory align with Consequence Impact Level (CI) and CI severity. Subject Matter Experts (SMEs) considered the 00 to 01 CPUC report time frame as relatively short for estimating probability for a potential high consequence wildfire. Given that SCE and SDG&E territories are adjacent to one another, it is assumed that a wildfire of similar severity to the 00 Witch/Guejito fire in SDG&E territory is possible in SCE territory. Public press showed safety, environmental, and financial damages for this 00 fire were consistent with a CI severity, and SMEs assume it is possible to anticipate a fire of this severity once every years in SCE service territory. The risk for a CI wildfire is evenly distributed between distribution overhead conductor, transmission overhead conductor, and poles. - Property Damage: CPUC Reportable Incidents from 00 to 01 provide counts for the number of incidents that have led to property damage.

16 Outage: ODRM data from 00 to 01 provides Customer Hours of Interruption (CHI) data to estimate severity overhead conductor caused outages. Event IDs are used to link ODRM and wire down repair order data to provide CHI estimates for wire down events. - Freeway or Road Closure: Wire down data includes a field that tracks this outcome, which provides the information needed to estimate a CP for this outcome when there is a wire down event. CPUC reportable incidents included records of two such incidents consistent with a CI level financial outcome. c) Mitigation Alternatives Identification The Overhead Conductor Program began in 01 with two primary mitigations: Reconductoring and Branch Line Fusing. In addition to evaluating the two existing mitigations, PRISM 01 reviewed other advanced mitigations capable of reducing overhead conductor risk in the future. The process to determine which advanced mitigations we should evaluate started at the end of 01; SMEs met to determine the range of mitigation options with the potential to reduce public safety of wire down events. SMEs originally suggested over 0 mitigations to consider as part of future OCP work. In PRISM 01, SMEs selected four advanced mitigations pilots to score: High Impedance Relays, Single Phase Automatic Reclosers, Analog Radio Detection, and Clampstars. In PRISM 01, we also scored four alternative mitigations: Aerial Bundled Cable (underground cable use in overhead lines), tree wire, Undergrounding, and Steel Poles. This section provides a brief explanation on the expected benefits of each mitigation. - Reconductoring is intended to target and replace small conductor to a minimum of 1/0 ACSR for tap and ACSR for main lines. While in the process of reconductoring, we also replace deteriorated or corroded conductor, crossarms, poles, connection hardware, and other damaged equipment. - Branch Line Fusing (BLF) tap (branch) lines is intended to de energize conductor when a fuse detects fault currents sufficiently large to potentially cause conductor annealing or melting. - High Impedance (HiZ) Relay Detection devices are installed at substation(s) for distribution circuits. The relays will detect a high impedance fault, which could indicate a possible energized wire down condition that would otherwise go undetected by other existing protection devices. - Single Phase Automatic Reclosers (ARs) improve circuit reliability by eliminating outages on branch circuits caused by temporary faults and protect smaller conductor from arcing damage.

17 Analog Radio Detection senses depressed voltage to indicate a possible load side energized wire down condition on single phase radials using the SCE Netcomm radio system and voltage relays. If successful, the detection system may be adopted by existing AMI smart meters to avoid additional hardware installations. - ClampStar Splice Applications are shunt connection devices used to improve automatic splice installation capabilities on all SCE distribution wire sizes. - Aerial Bundle Cable is insulated conductor wrapped around a guy messenger wire and is intended to prevent faults caused by contact from animals, small tree limbs and other vegetation, mylar balloons, and could include other types of contact. Preventing faults would reduce wire downs and intact conductor failures. Aerial cable's insulation also provides significant protection against safety incidents associated with humans contacting overhead lines. - Tree wire is bare conductor enclosed in a high density polyethylene, which may provide partial insulation. Compared to aerial bundle cable, the benefits are very similar. But since aerial cable is fully insulated, the risk reduction benefits would be higher in comparison to tree wire. Tree wire is intended to prevent momentary faults caused by contact from animals, small tree limbs and other vegetation, mylar balloons, and could include other types of contact. Preventing faults would reduce wire downs and intact conductor failures. Tree wire s partial insulation also provides some protection against safety incidents associated with humans contacting overhead lines. - Undergrounding overhead lines would mitigate all overhead risk but would inherit the risk of a new underground line. Compared to the other five mitigations being evaluated, undergrounding overhead lines would provide the most risk reduction, but at a very high cost. - Steel Poles were evaluated as a fire mitigation measure, and are intended to provide the same benefits as wood poles, plus the benefits resulting from greater fire resistance. Steel poles could also potentially reduce the number of injuries to emergency personnel by reducing the amount of pole and conductor debris on the ground in the event of a wildfire. d) Mitigation Risk Reduction Evaluation (1) Tranche Analysis Overhead conductor mitigations were organized into groups with similar risks, called tranches. Overhead conductor data is currently available at the circuit level. Since circuits

18 are a common reference point for conductor, SMEs elected to evaluate mitigations at a circuit level for existing mitigations. SCE s service territory includes, circuits, of which approximately,000 contain some portion of overhead conductor. PRISM 01 analysis ranks all circuits based on what SMEs consider to be leading indicators of future wire down events. This includes historical wire downs, circuit CB operations, fault duty, miles of small conductor, and conductor-related maintenance notifications. Circuit ranks have been relied on to make 01 and 01 OCP scoping decisions. Circuit rankings are the direct input to calculate a TEF value by circuit. An individual circuit s attributes relative to average circuit attributes in these indicators determines its predicted wire down count, or TEF value. The sum of all predicted wire down counts accounts for these leading indicators and is made to equal the historic annual observed TEF. Risk analysis also allocates consequence risk according to CP attributes for the circuit. Wildfire risk is allocated based on an analysis of high fire miles, high wind miles, and proximity to canyons for a given circuit. Injury risk is allocated according to customer density for a circuit relative to the maximum in the system. Outage risk is allocated according to number of outages for a circuit relative to the maximum in the system. () TEF and CP Effectiveness TEF and CP effectiveness are quantified to estimate the effectiveness of overhead conductor mitigation programs. Overhead conductor TEF effectiveness refers to the ability of a mitigation to reduce the probability of wire down events. Overhead conductor CP effectiveness generally refers to the ability of a mitigation to de energize downed wires, given that critical negative consequences occur as a result of downed wires remaining energized. The CI is generally not impacted, since the severity of the negative consequence is not mitigated (e.g., if a person contacts a downed wire that is still energized, the assumption is the impact severity would still be a fatality). Overhead Conductor SMEs collaborated to estimate TEF and CP effectiveness during working sessions for each mitigation. A common exercise in this effort involved asking SMEs how effective a certain mitigation would be at preventing wire downs for a given driver. Wire down data detailing how frequently certain wire downs occurred as a result of melting or breaking was considered by SMEs in this exercise. SMEs were then asked if a mitigation could de energize downed wires. The following points provide brief explanations on how each mitigation is expected to reduce TEF or CP.

19 Reconductoring TEF Effectiveness: New, larger conductor reduces TEF through its improved ability to withstand melt faults that lead to wire down events. Reconductoring also replaces supporting structures and provides reductions in crossarm failures, connector failures, other apparatus hardware, and splice failures. - Branch Line Fuse TEF Effectiveness: Branch line fuses prevent melt failures from occurring by minimizing the amount of time a segment of conductor is exposed to fault current. SCE does not expect Branch line fuses to be effective at preventing arc failures. - High Impedance Relay CP Effectiveness: Relays can detect high impedance faults and control existing protection devices to de-energize downed wires. - Single Phase AR TEF Effectiveness: Single Phase ARs can reduce outages on tap (branch) lines caused by temporary faults; limiting outages on those lines helps maintain the conductor s ability to withstand melt and arc faults. - Single Phase AR CP Effectiveness: Single Phase ARs can de energize a conductor, thus reducing safety risks associated with energized conductor down. - Analog Radio Detection CP Effectiveness: Analog Radio Detection can detect depressed voltage and be integrated with existing protection devices to de energize downed wires and reduce public safety risks associated with energized wire down events. - ClampStar TEF Effectiveness: ClampStars can prevent automatic splices from failing, mitigating wire down splice failures by automatic splices. - Aerial Bundled Cable TEF Effectiveness: Aerial bundled cable can reduce TEF through its improved ability to withstand melt and arcing faults that lead to wire downs. Aerial cable can also reduce intact conductor failures due to faults caused by contact. - Aerial Bundled Cable CP Effectiveness: Aerial cable can reduce outcomes associated with down wire such as injury / fatalities, wildfires, and property fires. Aerial cable can also reduce the safety consequences associated with humans contacting intact conductor. - Tree wire TEF Effectiveness: Tree wire can reduce TEF through reducing the likelihood of melt failure under fault conditions. Tree wire is only partially insulated due to the cover. Therefore, benefits for tree wire will be less than the benefits for aerial cable. - Tree wire CP Effectiveness: Tree wire can reduce outcomes associated with down wire such as injury / fatalities, wildfires, and property fires. Tree wire can also reduce the safety consequences associated with humans contacting intact conductor. 1

20 Undergrounding TEF Effectiveness: Undergrounding would eliminate all overhead triggering events (wire down, intact conductor failure, humans contacting intact conductor) and inherit underground triggering events (UG cable failure, UG structure failure, and humans contacting UG cable). - Undergrounding CP Effectiveness: Consequences associated with overhead lines are eliminated but consequences with underground lines are inherited. - Steel Pole CP Effectiveness: In the event of a wire down or pole failure, steel poles are not expected to reduce the CP of a wildfire starting. Steel poles may however reduce conditional probabilities of potential financial, reliability, or safety impacts occurring. e) Risk Spend Efficiency Table I- shows select risk spend efficiency results. As noted, only the highest or maximum individual circuit score for any given mitigation is shown in the table below. It is also important to point out that subject matter experts are relatively more confident in the assumptions underlying scores for mitigations currently in use (i.e. reconductoring and BLFs) than they are for inputs to scores for alternative mitigations (i.e. aerial cable, tree wire, undergrounding, and steel poles). Assumption for alternative mitigations require additional validation to better understand how these mitigations could scale if they are to be deployed across the service territory. 1

21 Table I- Overhead Conductor Mitigation Risk Spend Efficiencies Highest RSEPV Mitigation (MRRPV/$1k) Reconductoring & BLF 1 Reconductoring BLF Aerial Cable - Small Conductor Aerial Cable - All Conductor Tree Wire - Small Conductor Tree Wire - All Conductor Undergrounding Steel Poles The results shown in Table I- represent circuits that have some level of risk and are inexpensive to mitigate. The circuits are selected based on the core risk being addressed for each mitigation. Comments below provides brief explanations of RSE scores for several mitigation options. - Branch line fusing targets circuits that are generally in dense customer areas, have wire down risk, and are inexpensive to mitigate. This means circuits that have few locations for branch lines but protect a fair amount of conductor. - Aerial bundled cable and tree wire also target circuits of short length that are in dense customer areas; these circuits are less expensive relative to circuits that cover long distances, given costs are estimated per foot. The top circuits being selected may not have had wire down incidents occur, but were selected due to aerial cable and tree wire s ability to reduce fatalities associated with people touching intact conductor. Two scenarios are being shown above: (1) assuming small conductor is replaced; and () assuming all conductor is replaced. - Undergrounding is more cost-effective in targeting circuits that are smaller than 0. miles and located in high customer density areas. Scores for mitigations evaluated in 01 (i.e. High Impedance Relays, Single Phase Automatic Reclosers, Analog Radio Detection, and Clampstars) have not been scored at a circuit tranche level and therefore do not appear in Table I-. We performed OCP scoring by circuit, so the numbers in this column are for the circuit with the highest RSE for a given mitigation. 1

22 f) Key Takeaways (1) Data It is important to note that a lack of data on conductor condition and age led SMEs to rely on proxy data that they assume are leading indicators for conductor failures. Continuing to develop data sources will improve the risk analysis. For example, wire down data is a critical input, but system wide data only exists for approximately two years. We continue to collect data, and with a longer time period of data we should be able to improve the degree to which modeling can assess trends with leading indicators. In the future, it may also be beneficial to expand data to more granular levels than the circuit. () Modeling Continue to develop overhead conductor analysis to account for secondary conductor risks and dependencies between overhead conductor mitigations and mitigations for other assets. () Observation(s) The triggering event of human contact with intact conductor presents the greatest absolute risk for overhead conductor. SCE currently invests in public outreach programs to warn the public against getting near any intact conductor. In addition to these efforts, it may be beneficial in the future to consider OCP mitigations focused on reducing the risk associated with human contact with intact conductor events.. Poles The risk analysis for poles encompasses all distribution and subtransmission wood poles, and relates to the inspection and remediation activities associated with the Deteriorated Pole Program and the Pole Loading Program described in SCE-0, Volume. SCE first described its proposed methodology for assessing and quantifying pole-related risks in its Safety Model Assessment testimony filed in May of 01. The risk statements used to evaluate pole-related risks have not changed, but the methodology to estimate the impact of SCE programs on pole-related risks have been updated and enhanced. a) Risk Identification The risk event-outcome-impact combinations used to describe pole-related risk are summarized in Table I-. We evaluate pole risks in relation to a common triggering event of a pole falling in service. The risk statements evaluated are pole falls in service leading to injury, pole falls 1

23 in service leading to property damage, pole falls in service leading to outage, and pole falls in service leading to wildfire. Table I- Pole Risk Statements These risks are intended to capture the consequences of poles falling even if the resulting hazardous event is not directly related to poles. For example, wildfires are not expected to be caused by a pole falling, but rather by energized wire falling when the pole it is attached to falls. If the pole did not fall, the consequences would not have been triggered, and hence the risks associated with the wires in such cases are attributed to pole risks. The risks associated with overhead wire failing independent of a pole falling are accounted for in overhead conductor risks. Inspections and assessments provide information pertaining to the in-situ condition of poles and were used in the risk modeling to inform the probability of a pole falling. Pole loading assessments and intrusive inspections both identify poles for replacement, but the criteria for replacement differs for each program. The risk scores take this into account by quantifying both the overall risks presented by all poles and the specific level of risks presented by specific tranches of poles grouped together based on the results of their intrusive inspections and pole loading assessments. 1

24 b) Current Residual Risk Evaluation Table I- summarizes the 1-year current residual risk scores for poles, broken down by risk statement and impact dimension. The following sections describe the inputs and assumptions used to compute these scores. Table I- Poles CRR Scores 1-Year Current Residual Risk (CRR) Triggering Event Outcome Safety Reliability Financial Environmental Compliance Subtotal % of Total fatality 1,1-1,1 0% fatality,1,1 0% Pole Failure property damage 1 1 0% outage,1,1 1% wildfire,1,1,1 0, 1% Subtotal 0,0,1,,1-1,0, % of Total % 1% % % 0% (1) Triggering Event Frequency We forecast Triggering Event Frequency using a model of pole behavior which incorporated the impact of pole decay and pole loading on a pole s capacity to withstand extreme winds. This model created a virtual representation of approximately 1.M wood poles in SCE s distribution and subtransmission system, including installation date, maintenance schedules, safety factors, and levels of decay. The model includes several components which simulate the progression of natural processes within the pole population over several decades, including the initiation and spread of pole decay and the frequency of extreme wind events. By combining forecast prevalence and severity of decay with the expected frequency of extreme winds, the model estimates the frequency of poles falling as the pole population ages. Using this framework, the model was used to estimate the impact of the intrusive inspection program and pole loading assessment program on the frequency of poles falling in service. A second model forecasts the quantity of poles that begin decaying. This model uses pole attributes such as age and installation era to estimate the likelihood that decay will begin. It relies on two years of inspection data on over 00,000 poles, and produces a fit to the observed data as shown in Figure I-1. 1

25 Figure I-1 Modeled vs. Actual Inspection Failure Rates by Age A simulation combining these models with data on the entire pole population provides estimates of the number of poles that will be found for each pole tranche with the current inspection standards and frequencies. A pole tranche represents a group of poles that share the same characteristics across three attributes: decay level ( levels depending on intrusive pole inspection result), Pole Loading Calculation result (pass or fail), and whether pole resides in high fire zone (yes or no). The simulation incorporates a virtual representation of every pole on SCE s system. When possible, specific age, inspection result, and high fire status data are merged together based on pole equipment numbers to obtain the variables used to estimate the likelihood of an overwhelming wind event using the methods described above. When data is not available, we use population-level data to assign values which represent the expected distribution of values at the population level. The simulation models the progression of decay and likelihood of overwhelming wind events on a year-by-year and pole-by-pole basis for the fifteen year period. At the same time, the simulation models the progress of Intrusive Inspection Program and Pole Loading Programs through the virtual population of poles, and the flagging of poles that are detected for removal according to current program standards. Removed poles are replaced with new, sound poles that are free of decay and loaded for the applicable safety factor. 1

26 () Consequence Percentage (CP) and Consequence Impact (CI) Table I- summarizes the consequence percentage and consequence impact values for each risk statement along with the justification for the CI levels chosen. Table I- Consequence Percentage and Consequence Impact Values When possible, CP were based on claims data and pole notification counts. For example, over a seven-year period two poles falling in service led to fatalities a rate of 0. incidents per year. An average of 1,1 pole notifications were generated annually from late 0 to early 01; the rate of notifications was assumed to be relatively constant over the timeframe of the claims data. This indicates that poles resulted in fatalities at a rate of 0.0%. Similarly, one property damage claim equivalent to a level impact occurred, a rate of 0.01% per notification. We derived Wildfire CP based on a level CI with a recurrence interval of 1 in years. This reflects the belief that the drivers of wildfire risk specifically, the extent and intensity of SCE s operations in high fire areas have not dramatically changed over the past years. The age distribution of high fire poles in SCE s current inventory supports this hypothesis: high-fire poles comprise about 0% of those installed before, and % of those installed after. The absence of a level fire during this time period indicates that SCE s current practices have succeeded in reducing the risk of such a fire to some level less frequent than 1 in years. Therefore, the consequence percentage for wildfire reflects the 1 in year ceiling frequency (0.0 per year) divided 1

27 by the estimated number of notifications in high-fire areas (1,1 * % of poles in high-fire areas = 1) or 0.0% per triggering event. This risk is allocated equally across poles, distribution overhead conductor, and transmission overhead conductor. c) Mitigation Alternatives Identification Inspections are activities that support mitigations that can reduce risk, but are not mitigations in and of themselves. The Intrusive Pole Inspection (IPI) Program detects interior decay in poles, which reduces pole strength. Intrusive inspections determine the percentage of a pole s original strength that remains, given a certain level of interior decay. Poles failing intrusive inspections can result in a mitigation such as pole replacement or restoration. The Pole Loading Program (PLP) combines assessments and software modeling to determine a pole s ability to support expected equipment and wind loads. While intrusive inspections can determine the percent of the pole s original strength remaining, pole loading assessments can determine the as-built capacity of the pole to support load. Pole loading failures can result in a mitigation such as pole repair or replacement. This section provides a brief explanation on the expected benefits of each mitigation. - Pole Replacements involve wholesale replacement of an in-situ pole with a new pole meeting the current design standards. Replacing a pole mitigates many, but not all, of the drivers that can cause the triggering event of a pole falling in service. First, the new pole material has not experienced strength reduction due to deterioration. Second, the replacement pole will be designed to meet the current pole loading criteria, which will account for the actual loading on the pole. Please note that pole replacement will not directly mitigate potential overloading due to additional loads being applied to the pole following replacement. Finally, pole replacement can, to some degree, mitigate weather driven triggering events. For example, poles designed to a lower design wind speed that was later updated to a higher wind speed will necessarily have a greater ability to withstand high wind and, thus, a lower probability of falling in the future. - Steel Stubbing can potentially restore poles with decay above a minimum RSM threshold or that fail pole loading calculations. Steel Stubbing can mitigate deterioration by reinforcing the pole around the area of decay and restoring the pole to its original strength. Like pole replacements, mitigations that improve the in-place strength of a pole can reduce the probability that the pole would fall during a weather event and thus indirectly address the weather driver. 0

28 Guy Wire Replacement can remediate cases of pole loading failures. Like pole replacements, mitigations that improve the in-situ strength of a pole can reduce the probability that the pole would fall during a weather event; thus, these mitigations can indirectly address the weather driver. - E-Trussing can potentially reinforce poles that fail pole loading calculations to increase the strength of the pole. E-Truss restorations have not previously been applied as a pole loading solution and require additional engineering analysis to determine the appropriate scope and requirements for implementation. d) Mitigation Risk Reduction Evaluation The mitigation risk reduction for each of the pole mitigations is computed based on a forecast reduction to TEF. For all mitigations, the CP and CI values are assumed to be unchanged; the conditional probability that a pole falling would lead to an outcome (CP) and the associated magnitude of that outcome (CI) are a function of the physical location of the pole and not the condition of the pole itself. The following sections summarize the approach and assumptions used to compute the MRR values for each mitigation. (1) Pole Replacement We determine the risk reduction for pole replacements by the severity of decay and pole loading calculation results. The risk reduction stems from a reduction in triggering event frequency, and we computed it using the pole model described in section I.B..b) of this appendix. The likelihood of a pole experiencing an overwhelming wind load depends upon its current level of decay and its pole loading status. The MRR of a pole replacement is calculated based on the reduction in risk gained by replacing a pole with some level of decay, overloading, or both, with a pole that has no decay and is properly pole loaded. () Steel Stubbing The steel stubbing mitigation is assumed to have the same risk-reduction benefit as a pole replacement over the 1-year horizon used to calculate MRRpv. As a result, the MRR values for steel stubbing and pole replacement are equal, and determined by the level of deterioration and pole loading calculation results of the pole replaced or mitigated. Over longer time periods, steel stubbing poles are not expected to have the same longevity as a pole replacement. 1

29 () Guy Wire Repair / Replacement The MRR for guy repairs are based on an average MRR associated with remediating a pole failing pole loading assessment. () E-Truss E-Truss MRR calculations use the same assumptions as those stated above for steel stubbing. e) Risk Spend Efficiency The risk spend efficiencies for tranches defined by mitigation, decay level, pole load calculation (PLC) status and high fire status are summarized in Table I- and Table I-. The Estimated Units in these tables reflect both the 0% PLC reject rate which was estimated prior to the SpidaCalc client file update, and the.% reject rate estimated for the new SpidaCalc client file. Refer to WP SCE-0, Vol. (August_01_SCE_SPIDACalc_Impact_Assessment).

30 Table I- Mitigation risk spend efficiencies (RSEpv) by tranche - Replacements Unit Cost** RSEpv Estimated Units # Mitigation Tranche Description Decay Level PLC Result High Fire ($000) % PLC reject 1 Replacement Highest Fail Y $ 1,1 0 Replacement Highest Pass Y $ 1,1 Replacement High Fail Y $ 1,1 Replacement High Pass Y $ 1,1 0 0 Replacement Highest Fail N $ 1,1 1 1 Replacement Highest Pass N $ 1,1 0 Replacement Medium Fail Y $ 1,1 Replacement Medium Pass Y $ 1,1 0 Replacement High Fail N $ 1,1 Replacement High Pass N $ 1,1 1,0 Replacement Low Pass Y $ 1, Replacement Medium Fail N $ 1,1 0 1 Replacement Medium Pass N $ 1, Replacement Low Fail Y $ 1, Replacement Low Pass N $ 1,1 1,0 1 Replacement Low Fail N $ 1,1 1 1 Replacement None Fail Y $ 1,1,1 1 Replacement None Fail N $ 1,1, **Unit costs are weighted average of transmission & distribution unit cost

31 Table I- Mitigation Risk Spend Efficiencies by Tranche (RSEpv) Repairs Unit Cost** RSEpv Estimated Units # Mitigation Tranche Description Decay Level PLC Result High Fire ($000) % PLC reject 1 Steel Stub Low Pass N $,,0 0 Repair (Guy) None Fail Y $ 1,1 1 Repair (Guy) None Fail N $ 1,1 E Truss* None Fail Y $,00 1,1 E Truss* None Fail N $,00 1, *This application is currently under review, unit cost and RSE are preliminary **Unit costs are weighted average of transmission & distribution unit cost f) Key Takeaways (1) Modeling Incremental risk associated with pole failure in high-fire areas is reflected in the risk scores of otherwise similar poles. High-wind areas were not treated as a separate tranche. () Observation(s) Risk Spend Efficiency results for pole mitigations shows that when repairs are possible, they represent more cost efficient mitigation relative to replacements. However, repair options cannot be applied in all cases, and the time horizon of PRISM analysis does not include the full life cycle of both repair and replacement options. For example, a repair lasting 0 years and a replacement lasting 0 or more years are not treated differently, as long as the risk reduction achieved over the next 1 years is equivalent given the current MRR forecast window extends 1 years as discussed in SCE-0, Volume 1, Section III.D..b).. Underground Structures Underground structures includes activities performed by the Underground Structure Program. This program utilizes Underground Detailed Inspection (UDI) findings and detailed Field

32 Investigations to assess and mitigate deteriorated underground vaults and manholes. SCE first introduced the program in the 00 General Rate Case as an Infrastructure Replacement program, and expanded and refined the program as described in the 01 and 01 rate cases. GO 1 requires that UDIs be performed on underground equipment. The Underground Structure Program expands the scope of UDIs to include a preliminary assessment of the condition of the structure containing the equipment. A UDI inspector may recommend that a structural engineer perform a detailed Field Investigation assessment. The program manages the Field Investigation process and evaluates findings to determine the appropriate method and timeframe for addressing deterioration in the structure. Up until 01, vault replacement provided the only mitigation option for the Underground Structure Program. Temporary shorings were also constructed, but only as a safety measure until replacements could be completed. Late in 01 the program began testing a long-term shoring option which provides a lower-cost alternative to replacement in some cases, called adjustable vault reinforcement. The program plans to scale up adjustable vault reinforcement in 01 and to focus conventional replacements on the highest risk structures. Please refer to SCE-0, Volume for more details the Underground Structure Program. a) Risk Identification Risks identified for underground structures are summarized below in Table I-.

33 Table I- Underground Structure Risk Statements RS# Triggering Event Outcome Impact Dimension 1 Traffic Accident Financial External Collapse of Nonstandard Structures with High Deterioration (Tranche 1) Equipment Damage Financial Outage Reliability External Collapse of Standard Structure with High Deterioration (Tranche ) Internal Collapse of Standard Structure with High Deterioration (Tranche ) Internal Collapse of Standard Structure with Moderate Deterioration (Tranche ) Traffic Accident Equipment Damage Outage Equipment Damage Outage Injury Financial Financial Reliability Financial Reliability Safety b) Current Residual Risk Evaluation We developed CRR scores using a two-stage model of structure decay. The first stage of the model forecasts the quantity of structures which are expected to fall within given location, design, and deterioration levels. To forecast this quantity we used the age of the structure and level of ceiling decay obtained from recent field investigations. The second stage of the model forecasts the likelihood that a structure will actually collapse once it reaches a certain level of decay. This stage relied on SME judgement on the median number of years that structures with a given level of decay could last before experiencing a collapse. Both stages of the model are described in greater detail in the following sections. (1) Triggering Event Frequency (TEF) The underground structure statistical model estimates the count of structures with different levels of deterioration using installation dates, current ages of the underground structure inventory, and field investigation results. Triggering event frequency forecasts rely on a statistical model we developed using a subset that encompasses about 1% of the over,00 field investigations that we have performed under the Underground Structure Program since its inception.

34 This data included structure number, design, general and specific location, and level of decay for that subset of structures. SMEs compared the statistical model against engineering models which directly simulate the progress of decay within concrete. Engineering models suggest decay can progress more rapidly than the statistical model, particularly for structures with certain designs and concrete mixtures. Future refinements to the UGS model may rely more on a ground-up engineering approach which identifies particular decay patterns for particular types of structures. () Consequence Percentage (CP) and Consequence Impact (CI) Consequence percentage estimates relied heavily on outage data from the Outage Database and Reliability Metrics System (ODRM) and recent individual reports provided by Underground Structure program management and stakeholders. Personal Injury Reports (PIR) provided information on injuries occurring in conjunction with work in underground structures. CP and CI calculations were estimated for each of the four triggering events. These triggering events also correspond to the tranches in this risk analysis, which are covered in detail in Planned Residual Risk section below. - External Collapse of Nonstandard Structures with High Deterioration (Tranche 1): Data sources relied on for underground structures indicate two external collapses in years, or a TEF of 0. events per year. The estimated average population of structures falling into this tranche was over a year period. Dividing the annual rate of 0. incidents by the average tranche population of structures yields a CP of 0.%. Historically, these incidents have led to vehicle damage, equipment damage, and outages. The impact level for a traffic accident was set to CI based on the assumption that vehicle and road damage would fall between $0,000 and $00,000. The consequence impact for equipment damage was also set to CI assuming a similar level of incremental equipment replacement costs. The consequence impact for outages was set to a CI level based on ODRM data for a past structure collapse. - External Collapse of Standard Structure with High Deterioration (Tranche ): The structures in Tranche were assumed to be four times less likely as structures in Tranche 1 to be responsible for external collapses based on their designs, as explained in the previous section. As a result, 0% of the historical rate of incidents was attributed to Tranche structures. Since the historical rate was calculated to be 0. events per year, 0. events per year * 0% or 0.0 events per year were attributed to Tranche structures. The estimated average population of structures falling

35 into this tranche was over a ten year period. Dividing the annual rate of 0.0 incidents by the average tranche population of structures yields a consequence percentage of 0.01%. The consequence impacts for Tranche equal those of Tranche 1. - Internal Collapse of Standard Structure with High Deterioration (Tranche ): Two events were observed in years, resulting in an annual rate of 0. incidents per year. The estimated average population of structures falling into this tranche was 00 over a year period. Dividing the annual rate of 0. incidents by the average tranche population of 00 structures yields a consequence percentage of 0.0%. The reliability impact was set at a CI level, based on the same justification used for the reliability consequences for Tranches 1 and. Financial impacts due to damaged equipment were estimated to fall in the $0,000 to $00,000 range associated with CI outcome. - Internal collapse of Standard Structure with Moderate Deterioration (Tranche ): Personal injury report (PIR) data contains one debris-related injury in years of data from the beginning of 0 through the end of 01, or an annual rate of 0. incidents per year. The population of structures falling into this tranche was estimated at 1,00 for this year period. Dividing the annual rate of 0. incidents by the average tranche population of 1,00 structures yields a CP of 0.0%. The consequence severity was set to a CI level based on the description of the injury in the PIR report. c) Mitigation Alternatives Identification Three existing mitigations are currently in use and were considered in this risk analysis. - Conventional construction: Conventional structure replacements require a number of steps to execute, including excavation of the site, removal of the structure in place, temporary removal of equipment in the structure, installation of a complete new structure, and the transfer of equipment into the new structure. At times, we may need to complete additional work, such as repaving large sections of road surface. - Adjustable vault reinforcement: Adjustable vault reinforcement improves on the shoring methods we previously used to temporarily stabilize decaying surfaces in vaults. Unlike other shoring, adjustable vault reinforcement provides a durable improvement to structures. This improvement restores load-carrying capacity without significantly reducing the internal space available to perform repairs and operations. Shoring methods and procedures designed to reduce the risk of

36 injury in structures may have negative consequences that are not captured in this analysis. For example, temporary shoring may provide protection from debris, but can also make switch operation difficult or impossible. Temporary shoring may also prevent employees from performing thermal scans on equipment that is obscured by shoring materials. - In-Place Vault Reinforcement: In-Place Vault Reinforcement is a technique that installs permanent reinforcement to the interior of an existing decaying vaults. We use prefabricated materials to install a smaller vault inside of the decaying vault. This significantly shores up the structural integrity of the decaying vault. This option is much less expensive than conventional replacement, because the work can be done with no excavation. This has the potential to save costs on traffic control, permitting, and civil construction costs. d) Mitigation Risk Reduction Evaluation (1) Tranche Analysis Tranches correspond to risk statements. The type and magnitude of risk varies depending on a number of asset attributes. Each tranche represents a unique combination of three attributes: location, design, and deterioration level. - Tranche 1, External Collapse, Nonstandard Structures: Structures in this tranche have high deterioration, nonstandard tunnel designs, and are located in streets. Due to their location and design, these structures present a risk of collapse under surface loads. Tranche 1 structures reflect a nonstandard tunnel design, which leaves the tunnel segment containing the access opening particularly vulnerable to ceiling decay. This can rapidly undermine the capacity of the ceiling to support load. Structures with a standard tunnel design are less vulnerable to ceiling decay, and subject matter experts assume they represent a lower level of risk relative to nonstandard tunnel design structures. - Tranche, External Collapse, Standard Structures: Structures in this tranche have high deterioration, standard designs, and are located in streets. Due to their location, these structures present a risk of collapse under surface loads. - Tranche, Internal Collapse, High Deterioration: Structures in this tranche have high deterioration, standard designs, and are not located in streets. Due to their location, these structures present a risk of internal collapse.

37 Tranche : Internal Collapse Moderate Deterioration: Structures in this tranche have moderate deterioration, standard designs, and are located in streets or other areas. Due to their deterioration level, these structures present a risk to employees entering the structure. () TEF and CP Effectiveness Unlike other models that rely on TEF and CP Effectiveness estimates to calculate MRR values, the underground structures analysis estimates MRR values with a series of probability density functions. These functions are intended to project failure rates by tranche over time. We calculate the MRR for each type of mitigation based on the tranche of assets that we apply the mitigation to. SMEs assume that each mitigation discussed above will mitigate 0% of the risks in the 1-year analysis timeframe and MRR values vary by tranche. The per-unit MRR for each tranche was calculated based on the second stage of the underground structure model as discussed in Section b above. Based on the attributes of structures within each tranche, we developed a different probability density function for failure over time. The per-unit MRR can be calculated using probability density functions as a forecast of the likelihood of collapse for that individual asset. The probability of failure in each future year is estimated as the product of TEF and CP values. The MRR for each tranche is calculated by multiplying the probability density function for each tranche (which is the equivalent of the TEF and CP product for the next 1 years) by the applicable consequence impact terms for the failure of an asset in that tranche. The observed rate of collapses reflect the lower end of the probability density functions for each tranche. This represents the relatively small amount of time that structures remain in this high deterioration state before we discover the issue and repair the structure. Under the current program of UDIs and Field Investigations, this should be five years at most the three-year interval between UDI s and a two-year time period from planning to completing the replacement. The expected rate of collapse for each tranche depends on the observed historical rate and an assumed point in time in the future at which half the structures in that tranche would have collapsed. The highest risk structures were assumed to reflect a -year median life, with the remaining tranches experiencing 1., 1, and 0 year median remaining lifespans. The consequence rates are consistent with observed historical rates of collapse, and the assumed median remaining lifespans are comparable to the expected remaining lifespans of these structures. Structures with the greatest rates of decay, installed from the mid-s through mid-s, have been in service for 0 to 0 years and originally were designed to last sixty years. 0

38 e) Risk Spend Efficiency SMEs evaluated the current residual risk (CRR), mitigated residual risk (MRR), and planned residual risk (PRR) scores for underground structure tranches with two different mitigations: conventional structure replacement and adjustable vault reinforcement. The resulting risk scores are shown below in Table I-. Table I- Underground Structure MRR & RSE Tranche Consequence RSE - RSE - RSE - Observed Conventional Shoring Voltek CP CI MRRpv Traffic accident 0.% 1, 1) External Collapse, Equipment Damage 0.% 1, Nonstandard Outage % 1,0 Traffic accident 0.01%,00 ) External Collapse Equipment Damage 0.01%,00 Outage % 0,01 ) Internal Collapse, Equipment Damage 0.0%, High Deterioration Outage %,0 ) Internal Collapse, Moderate Deterioration Injury % Risk Spend Efficiency (RSE) scores indicate that replacement of structures located in streets generally results in higher risk benefits (Tranches 1 & vs Tranches & ). The low cost of shoring relative to conventional replacement produces a high RSE even for structures not located in streets. It also provides a high risk value when shoring is technically feasible. Shoring and replacement of moderately deteriorated structures produces a low RSE, but does not capture operational benefits of maintaining access to equipment in structures. f) Key Takeaways (1) Data The data sources we relied on in this analysis may understate the actual rate of incidents caused by underground structures deteriorating. Evidence of debris-driven equipment failures and outages may be destroyed by the subsequent violent failure of equipment. The energy released by electrical equipment failure could also generate debris, which can be difficult to distinguish from pre-existing structural decay. As awareness of the importance of this data increases, the number of incidents reported will also likely increase. Since 01, three events related to structure failure have been reported. 1

39 () Modeling As mentioned in Mitigation Alternatives, shoring methods and procedures designed to reduce the risk of injury in structures, particularly those with deterioration, may also have consequences which are not captured in this analysis. Future analysis for underground structures may consider incorporating risks associated with restrictions in operations and inspections that stem from certain shoring practices. Personal Injury Report (PIR) data includes injuries occurring in underground structures, which cannot be directly linked to structure decay. However, conditions within the structure may have played a role in these injuries. Risk analysis may produce different results if the analysis incorporates data from the entire set of field investigations. We may explore incorporating more data into the analysis to develop a more robust model, or to develop several models which project different patterns of decay based on the attributes of the structure. () Observation(s) A substantial 00x increase in CRR from year 1 to year is forecast for underground structures. This trend is driven by the slow pace of concrete decay relative to the current inspection cycle, as well as the near-term impact of structures with less-than-ideal concrete mixtures reaching their end of service life.. Circuit Breakers Circuit breakers (CBs) perform the critical function of turning off the flow of electricity to a circuit which experiences an abnormal system condition. One such condition is an electrical fault. If left unmitigated, electrical faults can draw a massive amount of energy through the upstream portion of the circuit which can damage or destroy overhead conductor, underground cable, distribution switches, substation buses, and transformers. In order to prevent this from occurring, circuit breakers must operate quickly. A properly functioning circuit breaker is expected to detect an overcurrent condition and isolate the circuit in less than one-tenth of a second. Distribution circuit breakers are typically located in residential and commercial area substations ( B substations). At these substations, electricity is transformed from a sub transmission level voltage (usually kv but sometimes kv) down to a distribution level voltage of either kv, 1 kv, 1 kv, 1 kv, kv, or. kv. The distribution (non-bulk) circuit breaker replacement program identifies and replaces breakers that are approaching the end of their service lives and therefore becoming increasingly unreliable. Such breakers may also contain parts prone to failure or that are no

40 longer readily available for purchase making them costly to replace. Sub transmission and Transmission circuit breakers are typically located inside transmission substations, and serve the transmission voltages (0 and 00 kv). More detail on CBs and CB infrastructure replacement is found in SCE-0, Volume. a) Risk Identification Table I-1 Circuit Breaker Risk Statements b) Current Residual Risk Evaluation We developed separate risk models for each type of circuit breaker classification based on the 01 age distribution of breakers. We used these models to calculate and forecast the current residual risk and mitigated risk reduction units over the planning horizon (01-0). Table I-1 and Table I-1 show Current Residual Risk (CRR) scores for circuit breaker triggering events.

41 Table I-1 CB <=kv CRR Scores Table I-1 CB 0 & 00kV CRR Scores 1 (1) Triggering Event Frequency (TEF) The circuit breaker triggering event is described as circuit breaker being unavailable, which includes CB failure, CB explosion, and a non-operational CB due to potential faults upstream or downstream. We evaluated and scored the triggering event for each two categories of CBs: less than or equal to kv and 0/00 kv CBs. Based on a driver analysis, we evaluated specific mitigation plans that would reduce risks over the planning horizon (01-0). The PRISM 01 risk modeling framework pairs a Weibull curve with the chronological age distribution of in service circuit breakers. A Weibull, or failure curve, provides an estimated probability of failure for a given asset based on asset age; it is derived through a statistical analysis of historical failures. The risk analysis assumes that assets removed preemptively were close to

42 in-service failure had action not been taken. This assumption is reasonable, given that we frequently test and monitor these assets. That testing and monitoring provides an up-to-date indication of asset health, which informs our decision to preemptively replace the asset. The circuit breaker risk model calculates the end-of-life leading to potential failure of circuit breakers over the planning horizon (01-0) based on the 01 age distribution and using the Weibull failure rate coefficients as described above. The model calculates the end-of-life count of assets leading to in-service failures for that particular year based on the age distribution of SCE s portfolio of CBs in that current year. When a circuit breaker fails in a particular year, the model replaces that breaker with a new unit with an age of 1 for the model to age for future year projections. The age of the entire circuit breaker population is then increased by one year for the subsequent year to calculate the failure rates. The process is repeated over the planning horizon (01-0) to forecast the current residual TEF units. () Consequence Percentage (CP) and Consequence Impact (CI). We perform detailed analysis, using multiple data sources, to map TEF to the appropriate risk statements and calculate the CP. The Worst Reasonable Direct Impact (WRDI) approach is used to score risks. In cases where there was limited or no data, SMEs relied on judgement to evaluate risks. c) Mitigation Alternatives Identification The SCE Substation Infrastructure Replacement (SIR) program preemptively replaces major pieces of aging or obsolete substation equipment in order to minimize the negative effect of aging equipment on system reliability, safety, and operability/maintainability. The SIR program includes two functions: Transformer Replacement and Circuit Breaker Replacement. This circuit breaker risk model evaluates the current SIR Circuit Breaker Replacement process. d) Mitigation Risk Reduction Evaluation (1) Tranche Analysis For the purpose of this risk analysis, the following voltages are used to divide SCE circuit breakers in two classes, or tranches of assets for analysis: - Circuit Breakers less than and equal to kv Refer to SCE-0, Vol. for more detail on the SIR program.

43 1 1 - Circuit Breakers 0 kv & 00 kv () TEF and CP Effectiveness Subject Matter Experts (SMEs) evaluated mitigations by quantifying their impact on TEF, assuming that CP and CI remained unchanged over the planning horizon (01-00). To forecast the MRR, we assume that we replace the oldest circuit breakers first. Since the CP and CI remain unchanged, the MRR can be calculated for each CB in each tranche using a Weibull probability density function alone. The probability density function is discounted at % to calculate the Net Present Value (NPV) of avoided TEF. The MRR is then calculated as the product of the NPV of TEF, CP, and CI. e) Risk Spend Efficiency MRR per unit and unit costs are used to calculate the Risk Spend Efficiency (RSE) over the planning horizon, which is shown below in Table I-1. RSE decreases over time because assets forecast to be replaced in the future are not as old as the assets replaced in the first year. Table I-1 Circuit Breaker Mitigation Risk Spend Efficiencies by Year of Expected Replacement Year RSE PV (CB<=kV) RSE PV (0kV & 00kV) f) Key Takeaways (1) Modeling SMEs identified multiple potential improvements for the risk model related to how Consequence Impact is measured for substation risks. A cross-functional group of SMEs will continue to improve the circuit breaker risk analysis model by considering scoring issues related to grid reliability, financial impact, and the safety risk dimension. SMEs will also consider the possibility

44 of measuring circuit breaker risks with additional tranche classifications that account for the insulating medium (i.e. Oil, SF, Gas, Vacuum and Air CBs).. Transformers A substation power transformer controls the continuity of the power supply and when it is unavailable through failure or other circumstance a stoppage or reduction in power supply is sustained. Risk scoring for this asset covers B-Bank (< kv), A-Bank (0 kv), and AA-Bank (00 kv) transformers. a) Risk Identification Table I-1 Transformer Risk Statements b) Current Residual Risk Evaluation Table I-1, Table I-1, and Table I-1 show Current Residual Risk (CRR) scores for transformer triggering events for B, A, and AA-Bank transformers.

45 Table I-1 B-Bank Transformer CRR Scores Table I-1 A-Bank Transformer CRR Scores Table I-1 AA-Bank Transformer CRR Scores AA Bank Transformers 1-Year Current Residual Risk (CRR) Triggering Event Outcome Safety Reliability Financial Environmental Compliance Subtotal % of Total Curtailments/Generation/Contingencies 0,0 0,0 1% Derating of Equipment 0% Outages 0 0 0% Transformer Unavilibility Power Quality - - 0% Property Damage,01,01 % Regulatory Fines % Revenue Loss 0 0 0% Safety Incident Leakage/Spills Subtotal,0 0, % of Total 0% 0% 0% 0% 0%