8. Adopted Elements of Cornerstone
Our rejection of Cornerstone as proposed does not mean that current reliability must remain static or that it is unnecessary to address the needs of certain customers who may receiving a lower level of reliability than other PG&E customers. Nor does it mean that we should ignore identified problems with PG&E's electric distribution system that may affect reliability.
Also, while we did not find the use of reliability index comparisons to be a compelling reason for adopting Cornerstone, we do recognize the value to PG&E's customers in improving PG&E's own SAIDI and SAIFI measures over time by implementing needed projects in a cost-effective manner. In considering the different elements of the Cornerstone proposal, we are looking to see if such value can be achieved by adopting a scaled down version of PG&E's proposal.
8.1. Distribution Automation
In this proceeding, PG&E is proposing to spend $605 million in capital to install a distribution automation system referred to as fault location, isolation and service restoration, or "FLISR," on 1,200 circuits in urban and suburban areas between 2010 and 2016. More specifically, PG&E plans to automate all of its 17-kilovolt (kV) and 21-kV circuits in urban and suburban areas (approximately 400 feeders) and approximately 800 12-kV feeders, which represents 56% of all of PG&E's urban and suburban 12-kV feeders. In addition to the actual line devices installed as a part of Cornerstone, PG&E will also automate circuit breakers, update substation equipment and purchase or use a communication system to allow the automated devices to transmit information and receive instructions. PG&E states that its proposal to install an estimated 6,000 line devices and pieces of communication equipment will provide significant automation on its system, similar to what other utilities are doing across the United States for their customers. PG&E adds that automating these circuits in urban and suburban areas will have the greatest impact on the largest number of customers and will create a significantly smarter grid. Also, while the distribution automation proposal is a significant undertaking, it will provide a substantial increase in PG&E's reliability performance and will provide long-term, concrete reliability benefits for PG&E customers.
PG&E states that it elected to automate 1,200 circuits, including most of the 17-kv and 21k-v circuits and 800 12-kv circuits, because it recognized that performance of certain circuits varied over time and that automating all of the identified circuits provided the most significant reliability improvements for customers and created a much more robust distribution system. This may all be true, but we have already determined that PG&E has not demonstrated the need for a program of Cornerstone's magnitude. At this point, there is insufficient justification for authorizing capital expenditures amounting to over $600 million to maximize reliability improvements or create a more robust distribution system. At some future point in time, this may change. PG&E will have the opportunity to justify the need to do so in future GRCs when it evaluates its reliability requirements more comprehensively and in more detail.
What is of interest now is the identification of the worst-performing circuits. We consider worst-performing circuits to be a problem simply as a matter of equity. While it will always be the case that some customers, for various reasons, will receive higher or lower levels of reliability than other customers, it is important that the needs of customers who continually receive significantly poorer service be addressed.
In its testimony, DRA includes graphs that show, for the 12kv, 17kv and 21kv circuits in urban and suburban areas, there are distinctly identifiable worst circuits in each of the categories.11 DRA argues that PG&E should be addressing the worst-performing circuits as part of the Company's regular business activities in the GRC, even without its Cornerstone proposal.
As part of its second alternative recommendation, TURN recommends that the Commission adopt funding of $129.7 million in capital expenditures and $7.7 million in expense for automating PG&E's worst 400 circuits, regardless of voltage level, over the time period 2010 through 2013. This recommendation would capture the identifiable worst performing circuits shown on DRA's graphs, as well as for 4-kv circuits. TURN notes that automating the 400 worst circuits would cost $7.2 million per minute of SAIDI saved, while the additional 817 circuits not in the group of 400 worst would cost $28.6 million per minute of SAIDI reduction. TURN also bases its estimate of cost assuming three automated zones per feeder,12 and notes that automating more than three zones over all of the 1,217 circuits PG&E wants to automate only nets a SAIDI reduction of 1.17 minutes, but adds $133.4 million in costs.
While TURN's proposed budget is based on implementing distribution automation on the 400 worst-performing circuits, TURN recognizes that what gets done may be somewhat different from that. TURN states that PG&E needs to do what is cost-effective and even some of the 400 worst circuits may not be cost-effective if the cost of automating them and building additional connectivity is prohibitively expensive. TURN indicates that the purpose of its proposed budget is to require PG&E to prioritize its program so that it obtains the most reliability benefit for the least cost (recognizing operational constraints).
If certain circuits fall within this worst-performing group, but the combined cost of automating them and building additional connectivity is prohibitively expensive (e.g., because a project requires five miles of underground line or a new transformer bank, etc.), then it may be that PG&E skips some circuits and implements automation on the 406th worst-performing circuit instead. Similarly, TURN states that it does not mean to suggest that PG&E automate only three zones on every circuit. There may be a few cases where automating four zones would net a much higher return than it would on the average circuit.
TURN's alternative recommendation for distribution automation is a reasonable means for addressing our reliability concerns with respect to poorly performing circuits. We will adopt its recommendations as described above, but with a slightly modified cost as described below.
TURN adjusted PG&E's costs in a number of ways. For purposes of determining a reasonable budget we will use TURN's three zone assumption. Also, TURN escalates PG&E labor by forecasted factors for the years 2012 and 2013. This is preferable and comparable to what is done in GRCs, as opposed to PG&E's use of the last negotiated rate for years beyond the labor contract timeframe.13 Also TURN used a value of $75,000 for underground devices as opposed to PG&E's use of $100,000. PG&E indicates that the cost is uncertain and both estimates are reasonable. We will use TURN's estimate.
We will adopt $136.341 million in capital expenditures for automating the 400 worst-performing circuits. This reflects TURN's recommendation but allows PG&E to recover limited costs for pole replacements and vehicles. TURN does not assert that these costs are not necessary or related, and to the extent possible we prefer to consider related costs together in one proceeding. Today's adopted pole replacement and vehicle costs are commensurate with the reduced level of distribution automation capital funding and add $6.648 million, $5.678 million for pole replacement costs and $0.970 million for vehicles.
We will adopt TURN's estimate of $7.7 million as the distribution automation expense that is related to the adopted capital proposal.
PG&E proposes to upgrade feeders in urban/suburban areas to have more conductor capability, ties and associated equipment in place to transfer customers within three-to-five switching steps when restoring service. According to PG&E, this work directly supports its proposal to install FLISR systems on approximately 1,200 circuits, adding that the estimated reliability benefits of distribution automation cannot be achieved without an appropriate level of feeder interconnectivity. In addition, PG&E proposes to enhance the level of interconnectivity on circuits it does not automate. PG&E forecasts capital costs of $697.167 million related to these efforts.
To determine the amount of work necessary to improve feeder interconnectivity, PG&E states that its engineers analyzed each feeder within urban/suburban distribution planning areas (DPAs) and identified instances where five or more switching steps are necessary to restore all customers following the failure of a feeder. As explained it its testimony, PG&E used the analysis associated with the loss of a feeder and number of transfers as a proxy to estimate the cost to provide the level of feeder interconnectivity necessary to support the distribution automation project proposal and other interconnectivity enhancements. After identifying the work necessary to resolve feeder deficiencies, a multiplier of 2.8 based on a weighted average of estimated zones per circuit14 was applied to estimate the amount of work.
Using this analysis, PG&E has estimated an expenditure level it believes is appropriate but has not developed a specific list of projects pending a detailed circuit-by-circuit analysis. PG&E states that if the Commission approves its request, engineers would then perform the detailed analysis necessary to identify the specific projects. Their primary focus would be to identify the work necessary to support the deployment of distribution automation and their secondary focus would be to identify work that will enhance the interconnectivity of circuits that are not automated.
We have adopted a certain level of distribution automation as discussed earlier in this decision and agree that a certain amount of feeder interconnectivity is necessary to accommodate that level. However, with a minor exception, we will not adopt further proposed expenditures to enhance the interconnectivity of circuits that are not automated. Again, the need for a broadly based program has not been justified. PG&E may request such enhancements in future proceedings provided it can justify the need and costs for such a program.
As part of its alternative recommendation, TURN recommends that the Commission approve an escalated capital expenditure of $45.538 million for connectivity to support the distribution automation portion of its alternative recommendation. This includes $7.4 million of "low-hanging fruit" identified in its testimony plus some additional circuitry to support distribution automation. TURN's recommended budget is based on the average unit cost per circuit of the lowest 60% of connectivity projects to obtain four transfers, a multiplier of 2.0 (the maximum for its proposal to automate three zones), and 32% of total circuits being completed in three years.
We previously adopted TURN's proposal for automating the 400 worst-performing circuits. We will also adopt TURN's recommendation with respect to the related needs for circuit connectivity. It will provide a reasonable amount of capital for improving connectivity, recognizing that the existing distribution system has a significant amount of existing connectivity and that while connectivity work related to distribution automation may be needed, PG&E should prioritize the projects it undertakes by skipping over very expensive projects with limited reliability benefits.
Additionally, we will adopt the TURN recommendation to invest approximately $7.4 million in the cheapest of the capacity connectivity projects that actually have reliability benefits.15 TURN notes that PG&E should be able to realize over 1 minute of SAIDI by such investment. In general, we support specific reliability improvement programs or projects that are cost-effective.
PG&E is proposing to change its planning process related to emergency transformer deficiencies to ensure there is adequate back-up for these critical pieces of equipment, reduce the risk of customers experiencing long-duration outages and ensure there is available capacity to fully implement the distribution automation proposal. Because utilization factors are high, lead times and costs for replacement units have increased and mobile transformers can take up to 24 hours to install, PG&E believes it is inappropriate to rely on mobile transformers to resolve substation transformer emergency deficiencies. Instead, PG&E proposes that urban/suburban DPAs have adequate emergency capacity installed within the DPA to cover the loss of a substation transformer bank without having to rely on a mobile transformer to restore service. Under PG&E's Cornerstone proposal, PG&E will plan for adequate capacity to restore all customers using the remaining capacity in the DPA following the failure of any transformer bank. This means PG&E will rely on the emergency ratings of the remaining transformers and other facilities as necessary to restore customers. Mobile transformer would instead be used to return the remaining transformers and facilities to their normal ratings.
PG&E's states that continued use of mobile transformers as defined in the current emergency capacity criteria would negate the benefits associated with Cornerstone proposal. Specifically continued reliance on mobile transformers means: (1) extended customer outage risks would not be reduced; (2) emergency capacity would not be available during all loading periods, especially the higher loading periods; (3) FLISR systems may not operate during the higher loading periods affecting customers as well as the estimated SAIFI and SAIFI improvements; (4) there would be an increase in the overall risk around critical equipment that is rising in cost and experiencing longer lead times; (5) there would be no life extension of current substation transformers and equipment; and (6) Cornerstone would not ultimately correct a system with known emergency substation transformer deficiencies which increases the likelihood of single failures cascading into larger events affecting thousands of customers.
As substation transformers, new feeder breakers and associated distribution facilities are installed to provide the emergency capacity PG&E is recommending, PG&E states it will coordinate that installation work with the feeder interconnectivity work to ensure a fully integrated distribution system. Together, these new facilities will support PG&E's distribution automation proposal by providing enough capacity to allow automated feeders to operate year-round.
Based on its proposal to phase out the use of mobile transformers to provide substation transformer emergency capacity, Cornerstone includes 95 specific transformer bank projects over the years 2011 through 2016 at a total capital cost of $610 million.
The specifics of PG&E's proposal were addressed by TURN and DRA in their opposition to the request.
In rebuttal PG&E states:
...Again, PG&E stresses that [Cornerstone] is fundamentally a policy question of whether PG&E should undertake the work to move to a new level of reliability performance, distribution system flexibility, and robustness. If so, because distribution substation transformers are the single most important piece of equipment in the electric distribution system, PG&E believes it is appropriate for the Commission to include this part of the company's proposal, which includes a higher level of emergency substation emergency capacity in urban/suburban areas to reach the new goal.16
While reliability performance, distribution flexibility, and robustness are laudable goals, the need to move to higher levels of each has not been demonstrated. Hence, as discussed earlier, we have rejected PG&E's Cornerstone proposal. Consequently, it would not be appropriate to make the wholesale changes proposed by PG&E with respect to substation transformer emergency capacity. In addition, we are somewhat troubled by PG&E's specific proposal to make the change to rely less on mobile transformers.
In opposing PG&E's substation transformer proposal, TURN makes the following important points, most of which were not rebutted by PG&E:
· PG&E's distribution planning guidelines have historically required a cost-benefit analysis prior to the installation of substation emergency capacity projects, in order to reduce costs while maintaining reliability.
· The outages the emergency capacity project would address are extremely rare.
· PG&E's Substation Asset Management Program has significantly improved transformer reliability.
· Substation transformers are extremely reliable.
· Peak periods are quite narrow in most locations, making the likelihood of a transformer failure during peak load small.
· PG&E's emergency capacity program would not have affected most of the 33 transformer outages with associated customer outages that occurred from 2002 - 2008.
· The emergency capacity project is clearly unreasonable when subjected to TURN's elementary value of service analysis.
· The emergency capacity program will contribute miniscule benefits to SAIDI and SAIFI for $610 million.
· With respect to the danger of long outages, the outages that have been observed total 0.82 minutes of outage per customer per year.
· By fragmenting Cornerstone from GRC costs, PG&E is creating a misleading and incomplete analysis. PG&E is already taking several other steps to rate its system more conservatively and build more "normal" capacity to reduce normal overloads. Such steps are already funded or proposed for 2011 GRC funding.
· The purpose of mobile transformers is to maintain flexibility to use them in emergencies, not to use them as temporary replacements for aging equipment as proposed by PG&E. TURN submits that if the Commission believes there is a small reliability problem with PG&E's transformer fleet, it should encourage PG&E to acquire a few more mobile transformers instead of spending half a billion dollars on new substation transformers. TURN notes the Energy Policy Act of 2005 encouraged the use of mobile transformers.
· With respect to distribution automation benefits, the bulk of such benefits come from speeding up what PG&E has always done with its system manually. Only incremental and second order benefits would come from adding any capacity elsewhere on the system. Also, emergency capacity does not provide normal capacity, according to PG&E itself. If these benefits were truly significant, PG&E would have spent time and effort quantifying them.
· The potential for life extension of transformers is an economic benefit. Engineering literature suggests that the optimal loading of distribution transformers can be analyzed using cost-benefit and value of service analysis - the very techniques that PG&E has eschewed in this application.
First of all, not only has the overall need for PG&E's Cornerstone proposal not been shown, the specific need for the substation transformer emergency capacity proposal has also not been demonstrated. It is not clear what the problems are that PG&E's proposal would solve. TURN's evidence suggests that, with respect to substation transformer emergency capacity, the extended customer outage risk is low and the bulk of FLISR benefits associated with distribution automation do not depend on the more redundant system that PG&E's proposal would provide.
Even if there were a problem with respect to substation transformer emergency capacity, there is not a preponderance of evidence that indicates that PG&E's proposal to not rely on mobile transformers is the optimal solution. For instance, TURN suggests the possibility of incorporating more mobile transformers rather than less, noting that the Electric Power Research Institute is conducting research into cheaper emergency mobile transformers with 20% lower cost, 25% less weight and 50% faster installation.
At this point, we suggest that as part of future reliability related analyses, PG&E take into consideration TURN and DRA criticisms of its proposal in this case in justifying the need for improving substation transformer emergency capacity and choosing and prioritizing the optimal solution to address that need.17
While we will not adopt PG&E's approximate $610 million substation transformer emergency capacity proposal, and expect that at least in the short term PG&E will rely on mobile transformers to address related problems, we do note that PG&E's individual bank loss deficiency studies indicate there are 191 substation emergency deficiencies and the proposal for 95 specific projects addresses deficiencies that range from slightly more than 0.1 megawatt (MW) to 38.6 MW. Even though the risk of outages related to substation emergency capacity appears to be low, we do not wish to overextend the mobile transformers. We believe it is prudent to keep the potential number of deficient transformer banks in check by adopting a limited number of specific projects for those substations that have the largest MW deficiencies. This will help ensure the continued viability of mobile transformer use for emergency capacity purposes. The list of projects provided by PG&E for the period 2010 through 2013 shows that there are 23 substations with deficiencies greater than 15 MW. 15 MW appears to be a reasonable cut off point to address our immediate concern to not overextend the use of mobile transformers. If spread over the time period 2011 to 2013, based on PG&E's cost estimate, the capital expenditures total $108.220 million for substation transformers.18 We will include an additional $5.329 million for associated feeder breakers, $0.962 million for distribution substation planning, $1.763 million for distribution capacity project management, and $1.487 million for substation maintenance expenses.19 In future proceedings, assuming the continued use of mobile transformers, we expect that more detailed analyses of need will provide a more appropriate delineation of whether such a cut off point based on MW deficiency or number of deficiencies is necessary, and, if so, what an appropriate value would be.
PG&E states that it is including a rural component to this filing for a number of reasons. First, the performance of PG&E's electric distribution system in rural areas is noticeably worse than other utilities. Second, the Company believes that rural circuits are likely less suitable for distribution automation because circuit ties may be inadequate and/or establishing communication between all the devices may be problematic. Third, rural circuits are generally longer than suburban/urban circuits. According to PG&E, this makes them good candidates for devices that will mitigate the number of customer minutes and interruptions due to faults on tap-lines or sections of mainline that are far away from Company service centers.
PG&E's proposes to install approximately 500 reclosers and 5,000 fuses on 16 rural circuits between 2010 and 2016. PG&E estimates a total capital expenditure of $62.4 million to install these devices which, when fully installed and operational, will improve annual SAIDI and SAIFI values by approximately five percent and seven percent, respectively, over the 2004 to 2008 average. PG&E has also estimated and requested $0.236 million in expenses associated with the rural reliability program.
PG&E states that its testimony from the 2007 GRC included information regarding reliability and the installation of protective devices, and its proposal in this proceeding simply represents more of the same type of work described in that rate case.
If nothing else, the reliability comparison information provided by PG&E reveals that among PG&E customers, those in low density areas, principally rural, receive less reliable service than those in high density areas. Similar to the worst circuit program that is adopted for distribution automation, the rural reliability program addresses problems for certain PG&E customers who, as a matter of course, receive worse service than other PG&E customers. Additionally, TURN indicates that, while it did not analyze the rural reliability proposal in detail, it seemed on the surface to be cost-effective. For that reason, TURN indicated that despite its primary recommendation that the entire Cornerstone project be rejected, it would not object if the Commission saw fit to reject everything other than the rural reliability component. In its alternative recommendation, TURN accelerated PG&E's 2010 to 2016 proposal over its recommended 2010 to 2013 time period. CUE also recognized the cost-effectiveness of the rural reliability proposal and recommended that the fuse and recloser program be expanded where cost-effective, not only for rural areas but for urban and suburban areas as well.
Because we see a need for improving rural reliability and PG&E's proposal appears to be cost-effective, we will adopt PG&E's proposal to install 5,000 fuses and 500 reclosers on rural circuits. Because this decision's authorization for reliability improvements extends only through 2013, we will accelerate the installation over the 2010 through 2013 time period, as recommended by TURN. We will not expand the proposal further as recommended by CUE, because it is not clear what PG&E has or has not included in its 2011 GRC for related improvements. The need to further expand the program can be appropriately evaluated in future GRC filings. For this proceeding, we adopt $59.294 million in capital expenditures and $0.079 million in expenses for rural reliability, as discussed below.
In response to DRA's claims that there is insufficient justification for the estimated capital and expense estimates, PG&E indicates that the number of installations was determined by reviewing data provided by distribution engineers and performing an analysis of primary circuit connectivity, and the unit costs were based on judgment after reviewing what was included in the 2007 GRC, a sampling of project authorizations and discussions with other engineering management personnel. With respect to the number of poles that would need to be replaced, PG&E states that judgment was used because there are no data that provide better values. While it may not be based on detailed calculations, PG&E has sufficiently explained its estimating methodology. In general, there is no prohibition against using subjective judgment of experienced personnel in making an estimate. We evaluate such methodologies along with whatever else is proposed and determine what is most reasonable. In this proceeding, with respect to the rural reliability proposal, there are no other estimating methodologies on the record and there is no evidence that PG&E's results are unreasonable. Therefore, for the most part, we accept PG&E's cost estimates for rural reliability as being reasonable. They have been adjusted only for the cost per pole recommended by DRA ($7,750 versus $8,200 estimated by PG&E) and accepted by PG&E and the labor escalation factors for 2012 and 2013 as recommended by TURN and discussed earlier for distribution automation.
8.5. SAIDI and SAIFI Benefits
PG&E's Cornerstone request includes $1.997 million in capital expenditures over the 2010 through 2016 timeframe. PG&E claims that the benefits of its Cornerstone proposal are a reduction of 39.3 minutes per customer per year for SAIDI and a reduction of 0.405 interruptions per customer per year for SAIFI. In authorizing $357 million in capital expenditures over the 2010 through 2013 timeframe, we have selected the most cost effective parts of the project consistent with the recommendations of TURN. Consequently, the SAIDI and SAIFI benefits associated with this decision correspond to those claimed by TURN, namely, 26.6 minutes per customer per year for SAIDI and 0.265 interruptions per customer per year for SAIFI. Therefore, we estimate that this decision will capture approximately 68 percent of PG&E's claimed SAIDI benefit and 65 percent of PG&E's claimed SAIFI benefit for 18 percent of the
capital expenditures requested by PG&E.20
11 See Exhibit 503, at 6-10.
12 The number of automated zones on a feeder is the primary driver of reliability benefits. PG&E assumes three to five zones per feeder in its analysis.
13 The adopted labor escalation factor for 2008 to 2012 is 1.1449 (as opposed to PG&E's estimate of 1.1587) and the adopted labor escalation factor for 2008 to 2013 is 1.1778 (as opposed to PG&E's estimate of 1.2021).
14 Multiplier equals weighted average of estimated zones per circuit minus one because the zone served by the breaker does not need a tie.
15 In Exhibit 121, at 79-80, TURN shows that in dividing the projects into quintiles, the first two quintiles provide 43% of the total reliability benefits for all of the 194 projects for only $7.4 million or 6.5% of the $114.6 million total cost.
16 Exhibit 2, at 2-3 to 2-4.
17 DRA's criticism of the proposal included, among other things, that the capacity costs are not justified by the small benefits; the small reliability improvements are not all due to Cornerstone; PG&E did not update its analysis; the details necessary to conduct an analysis of specific projects were not provided to DRA; and changes in the California economy need to be reflected in the DPA analysis. We do note that PG&E did rebut much of what DRA said. However, with respect to analysis of specific projects, we expect that much of DRA's concerns will be alleviated if PG&E's follows the directives of this decision in choosing and documenting optimal solutions after considering the severity of the problem, reasonable alternatives, appropriate cost-effectiveness analyses, and non-quantifiable benefits.
18 TURN claims that PG&E's unit cost for substation transformers is high. However, we will use PG&E's estimates, noting that its costs are based upon an analysis of all projects over a full year and compared to prior year's unit costs. We prefer this to TURN's analysis that used only three data points in its analysis.
19 These amounts reflect 11 breakers, reduced distribution substation planning and distribution capacity project management costs based on the ratio of adopted capital expenditures to PG&E requested capital expenditures for those categories, and substation maintenance expenses associated with 8 transformers installed in 2011 and 8 transformers installed in 2012.
20 Actual SAIDI and SAIFI benefits will depend on which projects PG&E implements through its prioritization and selection process. Also, the cited TURN and decision SAIDI and SAIFI benefits do not include benefits associated with the $60.446 million in authorized capital expenditures for substation transformer emergency capacity. Such benefits are likely to be relatively small. PG&E estimated SAIDI and SAIFI reductions of 3.6 minutes per customer per year and 0.037 interruptions per customer per year in conjunction with its $517 million substation transformer emergency capacity capital expenditure request.
