SCE estimates the cost of the SGRP at $680 million (100% level).10 This includes $569 million for replacement steam generator installation, and $111 million for removal and disposal of the original steam generators. These estimates include a contingency amount to address uncertainties associated with the fact that the estimates are preliminary and conceptual. Of the $680 million, $141 million (26%) is for contingencies.11 SCE states that the level of contingencies in its estimate is sufficient to cover all known risks.
TURN states that there is a significant risk of cost overruns. Therefore, unless the Commission adopts a cost cap, TURN recommends that a 20% cost overrun be modeled.
Aglet Consumer Alliance (Aglet) states that the costs for removal and disposal of the original steam generators are especially uncertain due to the costs of cutting large holes in the containment structures for removal of the original steam generators and installation of the new ones, and the lack of documentation from the intended disposal contractor concerning its ability to accept the original steam generators.
SDG&E represents that SCE historically has been unable to reliably forecast its SONGS capital budget. For example, in January 2000, SCE forecasted its capital additions for 2004 at $37 million, whereas actual additions were $143 million. SDG&E states that, while SCE's first capital additions forecasts for 2005 and 2006 were $50 million and $80 million respectively, SCE's most recent forecasts are $114 million for each of these two years. SDG&E does not allege that SCE is imprudent in its estimates. SDG&E represents that such forecasts of capital costs for nuclear projects, such as the SGRP, are inherently unreliable due to the exposure to events beyond the utility's control.
California Earth Corps (CEC) states that SCE has inadequately considered the costs associated with transportation of the replacement steam generators, and transportation and disposal of the original steam generators. As an example, CEC states that SCE did not consider whether there will be a sufficient work force to perform the SGRP given the SGRP planned by Pacific Gas and Electric Company (PG&E) for 2008 and 2009. CEC also states that SCE's cost estimate for steam generator removal and transportation is based upon the estimates for Unit 1 which involved one trip across the beach at a cost of $4 million (2002 dollars), whereas the SGRP will involve at least four trips, for which SCE has estimated a cost of only $8.9 million.
CEC states that, regarding environmental mitigation, SCE uses a $6.4 million contingency to cover all costs associated with mitigating environmental impacts without having had discussions with the permitting agencies or identifying what the mitigation measures might be. CEC recommends that SCE be required to redo its estimates of environmental analysis and mitigation once the Final EIR is done, and rerun its cost-effectiveness analysis based on those results.
No party has represented that SCE's SGRP cost estimate is too high. As discussed later in this decision, we impose a cap on SGRP costs, including removal and disposal of the original steam generators, based on SCE's estimates. Since SCE may not recover SGRP costs in excess of the cap, we need not address specific increases to the SGRP cost estimate above the cap. Therefore, SCE's SGRP cost estimate is reasonable for use in determining the cost-effectiveness of the SGRP. Notwithstanding the above, we will include the effect of a 10% increase in SGRP costs in our cost-effectiveness analysis to determine the sensitivity of the cost-effectiveness of the SGRP to such increases.
SCE's O&M costs have two components; the base O&M costs, and the costs for steam generator repairs and inspections (RFO O&M costs). SCE forecasted SONGS base O&M costs and RFO O&M costs for 2004 through 2022 based on 2004 recorded costs, the 2005 budget and the test year 2006 general rate case (2006 GRC) estimates.12
For RFO O&M costs in the SGRP case (if the SGRP is performed), SCE assumed that it would not incur steam generator inspection costs during RFO 16 because the original steam generators will be replaced at that time rather than inspected. Thereafter, SCE assumed that steam generator inspections would occur in every other RFO rather than in every RFO as is currently the case. SCE also assumed that if the SGRP is performed, there would be no steam generator repair costs in subsequent RFOs. In the shutdown case (if the SGRP is not performed), SCE used the same methodology for 2004 through 2009, at which point SONGS would shut down.
SCE states that it developed a high O&M cost estimate that is 10% above its 2006 GRC estimate. It states that the high O&M cost estimate reasonably bounds most unforeseeable regulatory and extraordinary operating expenses.
TURN states that due to the uncertainties relating to future operations, the Commission should direct SCE to use the 2006 GRC estimate for base O&M costs as the base case, and to include a high case that is 20% above the 2006 GRC estimate.
TURN states that recorded RFO O&M costs for RFOs 12 and 13 were higher than the amount included in SCE's 2006 GRC forecast. In addition, TURN points out that SCE's witness Perez testified that one-time costs for Unit 3's RFO 14 will exceed SCE's forecast. Therefore, TURN recommends that SCE be required to model RFO O&M costs 10% above the amount included in its 2006 GRC as the base case, and to model a high case that assumes a 20% increase over the base case.
SDG&E states that SCE's inability to accurately forecast SONGS O&M costs is comparable to its inability to forecast its capital expenditures. SDG&E represents that SCE's SONGS O&M budget for 2004 was $40 million above the amount forecasted in its 2003 GRC.13 SDG&E points out SCE's base O&M forecast based on its 2006 GRC is about 12% higher than its initial forecast in this proceeding due to previously unforeseen NRC compliance requirements.
As discussed above, SCE's history with respect to O&M costs demonstrates that O&M costs are likely to exceed its estimates, and that there will likely be future O&M costs that are not currently known. Since SCE states that its high O&M cost estimate, 10% above the 2006 GRC estimate, bounds most unforeseeable regulatory and extraordinary operating expenses, and we are estimating expenditures up to 17 years in advance in this proceeding, we find SCE's high O&M estimate (base O&M plus RFO O&M) reasonable and will use it in our base case. We will also consider the effect of a 10% increase above this level to determine the sensitivity of the cost-effectiveness of the SGRP to such increases.
For the SGRP case, SCE forecasted its capital additions based on 2004 recorded capital additions, the forecast in its 2006 GRC, and an averaging methodology for 2009-2022. It then reduced the forecast amounts by 20% in 2019, 40% in 2020, 60% in 2021, and 80% in 2022, to reflect the fact that SONGS would not be operated beyond 2022. For the shutdown case, SCE used the base capital additions scenario for 2004-2009. It then reduced the forecast amounts by 40% in 2007, 60% in 2008, and 80% in 2009, to reflect the fact that SONGS would not be operated beyond 2009.
SCE states that it developed a high capital additions estimate that is 22% above the 2006 GRC estimate. It represents that a review of data for 1987-2004 shows that actual costs exceeded forecasts developed 5 years before by approximately 25%. It also says that excluding post-911 security measures in 2004 from the analysis yields a 15% variance. SCE states that its high capital additions estimate reasonably bounds the uncertainty inherent in its capital additions forecast.
TURN points out that SCE originally modeled a high case that assumed capital additions 50% over its estimate. Since then, it filed its 2006 GRC with higher estimated capital additions. In addition, it presented to the SONGS Board of Review a still higher number.14 Therefore, TURN recommends the Commission direct SCE to revise its modeling of capital additions to include a high case with capital additions 50% over those included in the 2006 GRC.
Aglet states that SCE's capital additions forecasts do not fairly incorporate the risks of major additions due to increased security requirements, seismic requirements, or environmental mitigation requirements. Aglet also argues that SCE's forecasts do not reflect changing regulatory incentives regarding capital additions. For example, Aglet states that from the beginning of commercial operations through 1995, capital additions increased. From 1996-2003, SCE was under incentive ratemaking for SONGS, during which period capital additions declined. In 2004, when cost-of-service ratemaking resumed, capital additions increased. For these reasons, Aglet states that capital additions will likely increase over time, and SCE's high capital additions estimate is more likely to occur than its base case.
SDG&E states that, for the reasons discussed above, SCE's inability to control costs calls into question SCE's capital additions estimates.
CEC states that SCE failed to include in its capital additions estimates the effects of ageing power plant components. CEC represents that industry experience shows that nuclear power plants experience increased costs for repair and replacement of components in the early and the late years of operations, and that steam generator degradation is an example of the effect of ageing. CEC cited as an example a fire in 2001 that was started by a worn-out circuit breaker, and resulted in $100 million in unanticipated costs. CEC also cites as an example the $64 million reactor head replacement project, included in SCE's cost-effectiveness evaluation, the need for which was not known a year previously. CEC further states, based on industry experience at the Davis-Bessee Nuclear Power Plant, that there is at least a 1.4% chance that one of the SONGS units will require $630 million in additional capital expenditures if the SGRP is performed.
For the above reasons, CEC recommends that the Commission require SCE to rerun its cost-benefit analysis using a reasonable estimate of the costs for repair and replacement of ageing components. In the alternative, CEC recommends that the Commission factor in an additional $630 million in future capital additions. CEC also asserts that additional costs due to uncertainties related to unforeseen events and the uncertainty of how the NRC would react to them should have been included in SCE's cost-effectiveness analysis.
Data for 1987-2004 shows that actual costs exceeded the forecast developed five years before by approximately 25%. SCE states that its high capital additions estimate (22% above the 2006 GRC estimate) reasonably bounds the uncertainty inherent in its capital forecast. The 25% historical variance between estimates five years in advance and actual expenditures reflects the fact that there will be unanticipated costs due to ageing, changing NRC requirements, or some other reason. In this proceeding, we are estimating expenditures up to 17 years in advance rather than five years. Therefore, we find that a capital additions estimate of 25% above the 2006 GRC estimate is reasonable and appropriate for use in our base case. We will also consider the effect of a 10% increase above this level to determine the sensitivity of the cost-effectiveness of the SGRP to such increases.
CEC believes that there is a high probability that the NRC will impose more stringent security requirements on SONGS. To illustrate its point, it states that on September 17, 2004, the United States Court of Appeals for the District of Columbia issued an order that states that the NRC will commence a rulemaking proceeding to consider revisions to the design basis threat that forms the basis for the NRC's security requirements at nuclear power plants. CEC also says that the Government Accountability Office (GAO), in testimony before a House of Representatives subcommittee, said that the NRC could not assure that commercial nuclear power plants were safe from terrorist attack. CEC says the GAO reported that the Department of Energy is reviewing the security requirements for its nuclear power plants. CEC notes that the current requirements do not include defense against terrorist attacks by airplanes. CEC contends that additional security requirements will be imposed by the NRC to address defense against a terrorist attack by airplane that will result in increased capital and O&M costs that should be included in the cost-effectiveness evaluation of the SGRP. CEC provided three scenarios to illustrate its estimates of the increased security costs:
· The first scenario assumes that SONGS stays in operation. CEC estimates that additional security requirements would result in additional capital costs of $374 million spread over the first two years, and $12.5 million per year thereafter until the reactors are shut down. Annual O&M costs would increase by $54.5 million until the reactors are shut down, and $16 million per year after shutdown.15
· The second scenario assumes that SONGS is permanently shut down when the requirements are put into effect. It also assumes that a lesser level of enhanced defenses would be put in place only to safeguard spent fuel. CEC estimates that the additional capital costs would be $154 million spread over the first five years after shutdown, and $3.4 million per year thereafter. The additional annual O&M costs would be $16 million per year after shutdown.
· The third scenario assumes that SONGS continues in operation for three years after initiation of the security requirements, and then is shut down.16 CEC estimates that the additional capital costs will be $201.5 million spread over the first three years. The additional O&M costs would be $35 million per year for the three years the reactors are operating, and $16 million per year after shutdown.
Based on the above, CEC recommends that SCE be required to rerun its model with the above cost estimates, and perform a sensitivity analysis.
CEC's first scenario corresponds to continued operation more than three years after the enhanced requirements are put into effect. This corresponds to both the case where the SGRP is performed and the case where the SGRP is not performed, unless it is known at the time the security requirements are put into effect that neither SONGS unit will continue in operation for more than three years. Given the uncertainty as to when SONGS will shut down if the SGRP is not performed, this appears to be the most likely of the three scenarios both with and without the SGRP.
CEC's second scenario has both SONGS units permanently shutting down when the enhanced security requirements are put into effect. Since the replacement energy cost for one unit is substantial, it could be cost-effective to implement security requirements even if SONGS has only a few years of life remaining. Therefore, this scenario is unlikely.
CEC's third scenario assumes that the NRC would exempt SONGS from some of the new security requirements because it will not continue in operation for more than three years. Since it is uncertain when either of the SONGS units will shut down without the SGRP, it appears unlikely that the NRC would impose lesser security requirements. As a result, this scenario is unlikely.
CEC appears to believe that enhanced security requirements will be imposed within the next few years. In that case, its first scenario would apply whether or not the SGRP is performed. The only effect on the cost-effectiveness analysis would be that the reduction in the increased O&M costs from $54.5 million to $16 million due to shutting SONGS down would occur at a later date.
We have no basis in the record for estimating the probability of the occurrence of future increased security requirements or their timing. As discussed above, CEC's assumption that lesser additional security requirements would be imposed if SONGS is shut down at the time of imposition is unlikely. Based on CEC's representations most, if not all, of any new security requirements would be imposed on SONGS with or without the SGRP. In addition, the costs estimated by CEC are illustrative examples rather than estimates based on known requirements. For the above reasons, we will not adopt CEC's cost estimates. However, the possibility of future increased security requirements supports our earlier conclusion that some increase in future O&M costs and capital additions above the amount forecast by SCE is appropriate.
SCE did not specifically model the impact of an unplanned outage at SONGS over its remaining lives. However, it states that the difference between an 88% capacity factor and an 84% capacity factor is 350 days of operation. Therefore, SCE represents that the effect of a one-year outage is approximately equal to the effect of a reduction in the capacity factor from 88% to 84%.
TURN states that 16 domestic nuclear plants have experienced outages over 12 months since January 1, 1990, and at least six other plants have had outages of nine to twelve months. Therefore, TURN says that it is reasonable to consider the potential for a year-long outage in evaluating the cost-effectiveness of the SGRP. TURN notes that a reduction of the capacity factor from 88% to 84% would be a partial proxy for such an outage.
The effect of a one-year outage on the SGRP's cost-effectiveness will vary, depending on when it occurs due to the time value of money. Therefore, since the discount rate generally exceeds the escalation of the cost components in the cost-effectiveness analysis, the effect decreases over time. Utilizing a 4% reduction in the capacity factor as a proxy spreads the outage over the remaining life of the plant, which means that the actual costs of a one-year outage could be a greater or lesser amount depending on when it occurs. The record does not demonstrate that a one-year outage is likely.17 Therefore, we will not include a one-year outage in our base case. However, we will include a one-year outage in our cost-effectiveness analysis to determine the effect of such an outage.
SCE based its 88% capacity factor on the nine-year average for 1996-2004.
TURN agrees that an 88% capacity factor is reasonable for a base case. It recommends that an 84% and an 80% capacity factor be used in the cost-effectiveness analysis to reflect the fact that SONGS has had lower capacity factors in the past (SONGS averaged an 84.89% capacity factor between 1988 and 2003).
Aglet finds SCE's use of an 88% capacity factor reasonable.
Since an 88% capacity factor reflects the average for 1996-2004, and the parties have no objection to it, we find it reasonable and will use it in our base case. We will also include 92% and 84% capacity factors in our analysis to examine the effect of variations in the capacity factor on cost-effectiveness.
There are two components to SCE's replacement energy costs: replacement generation capacity, and electricity production costs. Replacement generation capacity measures the benefit of deferring construction of 2,150 MW of replacement generation from 2009 to 2022. Electricity production costs are the costs to operate the replacement generation.
SCE assumed that replacement generation would consist of combined-cycle gas turbines (CCGT) with a heat rate of 7,250 Btu/kWh.18
SCE used models to forecast electricity production costs that require a gas price forecast as an input. SCE's gas price forecast assumed that liquefied natural gas would be imported to southern California. SCE represents that its forecast is conservative because it assumes lower gas prices than are presently occurring in the market.
TURN points out that, in Decision (D.) 03-12-059 and D.04-06-011, the Commission approved CCGTs with heat rates of 7,100 Btu/kWh for the Mountainview Power Project (Mountainview), and 6,971 Btu/kWh for Calpine Corporation's Otay Mesa Power Plant (Otay Mesa), respectively. As a result, TURN recommends the use of 7,100 Btu/kWh for this proceeding.
TURN states that there is a possibility that SCE's Mohave Generating Station (Mohave) will return to service in the 2009-2010 time frame, and recommends that this possibility be considered.
In D.03-12-059, the Commission authorized SCE to acquire Mountainview Power Company, LLC (MPC) as a wholly owned subsidiary of SCE and to enter into a power purchase agreement with MPC for the purchase of electricity from Mountainview. Mountainview is a CCGT with a target heat rate of 7,100 Btu/kWh. The target heat rate is the heat rate for the new plant after no more than 100 hours of operation. It is not the heat rate that would be expected over the life of the plant. The heat rate over the life of the plant would likely be higher due to the effects of ageing. SCE's forecast heat rate is approximately 2% higher than the Mountainview target heat rate.
By D.04-06-011, the Commission authorized SDG&E to execute the Otay Mesa Power Purchase Agreement (PPA) and recover the costs through commodity rates. Otay Mesa is a CCGT currently under construction by Calpine Corporation. It will have a nominal output of 585 MW, with guaranteed base load and peak heat rates of 6,971 and 7,230 Btu/kWh, respectively. This means that the actual heat rates achieved by Otay Mesa will be between 6,971 and 7,230 Btu/kWh, or within 4% of the value used by SCE. Otay Mesa has water available for cooling, which helps it achieve its low heat rate. Additionally, the PPA is only for the first 10 years of the plant's life. The heat rate over the life of the plant will likely be higher due to the effects of ageing. Therefore, the guaranteed heat rates are for the first ten years of the plant's life, not its entire operating life.
The heat rate used by SCE is for the life of a CCGT, as opposed to the first 100 hours or ten years of operation. It is within 4% of the heat rates used in D.03-12-059, and D.04-06-011. For these reasons, we find the 7,250 Btu/kWh heat rate used by SCE reasonable.
In D.04-12-016, the Commission authorized SCE to make expenditures on Mohave to allow continued operations after 2005, and to study options regarding continued operation or replacement of Mohave's power generation. SCE was also directed to evaluate other viable procurement options to be used in conjunction with Mohave. The Commission's stated goal is to return Mohave to service with as short a shut-down period as possible. At this time, it is unknown whether Mohave will be in service after 2005, and at what cost. Therefore, there is no way to include potential Mohave generation in the cost-effectiveness evaluation. As a result, we will not adopt TURN's recommendation that it be considered.
SCE states that the shutdown of SONGS would cause transmission system degradation that could lead to blackouts. Therefore, it represents that significant transmission mitigation would be necessary. SCE assumes that such mitigation would not already be in place and, therefore, included it in its cost-effectiveness evaluation. SCE further represents that such mitigation will need to be in place prior to shutdown to avoid thermal overloading of the transmission system or low voltage conditions that could lead to blackouts. SCE evaluated three transmission mitigation scenarios, and assumed that the SGRP would defer the transmission mitigation costs to 2022.
SCE's first scenario is a reinforced SCE/SDG&E 230 kilovolt (kV) interface (Barre-Ellis 230kV transmission line upgrade). This would consist of a 230kV transmission line upgrade, and the installation of voltage support equipment installed in substations.19 This is the lowest cost scenario.
SCE's second scenario includes an Imperial Valley-Ramona 500kV transmission line project, and an upgrade to the Path 49 Upgrade Project scope. In addition, additional voltage support equipment would have to be installed at substations. This is the highest cost transmission mitigation scenario.
SCE's third scenario includes the Valley-Rainbow 500kV transmission line project. Additional voltage support equipment would also have to be installed at substations.
TURN argues that SCE's assumption that the SGRP would defer transmission upgrades to 2022 is flawed. It states that, since SDG&E says it will build a 500kV transmission line whether the SGRP is performed or not, it is unreasonable to assume that the SGRP would defer the need for a 500kV transmission line to 2022, even if SDG&E's 500kV transmission line is not completed by 2010 as SDG&E represents.
TURN states that SCE's lowest cost transmission mitigation scenario should be considered the worst case. TURN estimates, based on SDG&E's representation that a 500kV transmission line will be built regardless of whether the SGRP is performed, that only additional voltage support equipment would be needed as transmission mitigation.
Aglet states that since SDG&E will add a 500kV interconnection to its transmission system, SCE has overstated the transmission mitigation benefits of the SGRP.
SDG&E states that SCE's transmission mitigation scenarios are incorrect because SCE fails to recognize that SDG&E will construct a 500kV transmission line by 2010 to meet SDG&E's own needs regardless of whether the SGRP is performed. SDG&E acknowledges that it has not yet identified the preferred transmission project, and has not sought formal approval from the Independent System Operator (ISO). However, SDG&E has performed studies that identified several 500kV project alternatives since the fall of 2003 with the ISO's involvement. SDG&E states that the completion of such a project by 2010 is aggressive, but feasible. SDG&E points out that in D.04-12-048, the Commission found its long-term resource plan, which included a 500kV transmission line in the 2010 time frame, to be reasonable to satisfy transmission grid reliability needs. SDG&E states that if it files its transmission line project application in April 2006, and the Commission utilizes a full 18-month review period, a certificate of public convenience and necessity could be issued at the beginning of October 2007. That would leave 33 months to complete the project by 2010. SDG&E states that this schedule is comparable to the schedule SCE contemplates for its Devers-Palo Verde 2 Transmission Project.
SDG&E states that, while its 500kV transmission line will eliminate the need for such a line to mitigate the effect of shutting down SONGS, some voltage support equipment would be needed (598 MVARs). SDG&E estimates the cost of this equipment to be at most $61.25 million, of which $36.25 million (354 MVARs) is attributable to SCE.
SDG&E's long-term resource plan, which includes a 500kV transmission line in the 2010 time frame, was found in D.04-12-048 to be reasonable to satisfy SDG&E's transmission grid reliability needs. If the SGRP is not performed, it appears that Unit 3, and likely Unit 2, will continue in operation beyond 2010. As a result, there appears to be sufficient time for SDG&E to construct such a line. Therefore, we agree with SDG&E that such a transmission line will be built by SDG&E regardless of whether the SGRP is undertaken, and will be available to mitigate the effect of SONGS shutdown. For this reason, and the fact that we are addressing only SCE's costs for transmission mitigation if SONGS shuts down, we need only address the amount of voltage support equipment needed by SCE.
Both SCE and SDG&E agree that additional voltage support equipment will be needed to mitigate the effects of shutting down SONGS. However, they do not agree on the amount. SCE's three scenarios indicate an average of 1,136 MVARs of voltage support equipment to be installed by SCE. SDG&E estimates the total requirement for it and SCE at 598 MVARs, of which 354 MVARS would be attributable to SCE's transmission system. Given the uncertainty as to the exact nature of the transmission line SDG&E will build, we believe that the voltage support equipment requirements proposed by SCE and SDG&E provide a reasonable range (354-1,136 MVARs) of voltage support equipment to be installed on SCE's transmission system. The midpoint of this range is 745 MVARs at a cost of $78.8 million, which we find reasonable and will use in our base case.
Steam generator tube degradation forecasts are expressed as the percent probability that a unit (its steam generators) will exceed the plugging limit.20 SCE's tube degradation forecasts are based on statistical forecasts by DEI. SCE adjusted the DEI forecasts (SCE increased the probability that the units would exceed the plugging limit) to include subjective components to account for changes in industry experience, and NRC guidance and requirements.
TURN recommends that SONGS should be assumed to run, in the absence of the SGRP, until the probability of exceeding the plugging limit is 50%. It states that this would be one to two refueling outages later than 2009, based on SCE's information.
Aglet points out that DEI stated that there is a large amount of uncertainty about the forecasts and that there is a new mode of degradation operating at Unit 2. Aglet also states that an SCE file report states that SCE's forecasts of steam generator repairs cannot be used with confidence more than a few refueling cycles into the future. Therefore, Aglet concludes that the rates at which degradation will occur are uncertain.
SDG&E states that SCE's adjustments are 6% for Unit 2, and 12% for Unit 3. SDG&E says that over the last three refueling outages, the trend of actual repairs to DEI forecast repairs reflects that fewer repairs than forecast by DEI were actually needed.
SDG&E states that, based on degradation forecasts prepared by DEI for SCE, without SCE's adjustments, there is a 33% probability that Unit 2 will exceed the plugging limit by RFO 17 in 2011.21 Likewise, there is a 44% probability that Unit 3 will exceed the plugging limit by RFO 20 in 2017.22 Therefore, SDG&E states that the SGRP will not be needed according to the schedule SCE proposes. SDG&E further states that it may be reasonable for SCE to acquire the replacement steam generators and to store them against possible future need.
For Unit 2, referring to only the past DEI forecasts without any subjective adjustment by SCE, the RFO 11 and RFO 12 actual repairs were both 30% less than forecast, and the RFO 13 actual repairs were 6% less than forecast. For Unit 3, the RFO 12 actual repairs were 3% more than forecast, and the RFO 13 actual repairs were 55% less than forecast. Therefore, we agree with SDG&E that the actual repairs were generally less than forecast by DEI. Since forecast repairs are based on forecast tube degradation, this means that the actual tube degradation was generally less than forecast by DEI (without SCE's subjective adjustments).
In its application, SCE increased the forecast tube degradation, over the amount forecast by DEI, based on its subjective estimate of the effect of industry experience, and NRC guidance and requirements. This subjective adjustment was an increase of 6% for Unit 2, and an increase of 12% for Unit 3. Since the results of the most recent DEI forecast used by SCE for Unit 2, incorporating more recent data than used in SCE's application, indicate less degradation, one would expect SCE to reduce its degradation estimates accordingly.23 However, SCE did not do so because, according to its witness, no adjustment was necessary. This means that, since the more recent DEI forecast decreased and SCE's forecast was unchanged, SCE effectively increased its subjective adjustment from 6% to 12% for Unit 2. SCE has not explained why such an increase in its subjective adjustment is reasonable.24 There may be future effects of industry experience, and NRC guidance and requirements on forecast tube degradation. However, SCE has not shown that its subjective adjustments are reasonable as discussed above. Thus we will base our cost-effectiveness analysis on the most recent DEI degradation forecasts.
The forecast of tube degradation is relevant to when SONGS would shut down if the SGRP is not performed. The record demonstrates that there is considerable uncertainty as to when the steam generators will reach the plugging limit. For determining cost-effectiveness, it is reasonable to assume the steam generators will reach the plugging limit when the probability of doing so is 50%. This is the point at which there is an equal probability they will shut down at an earlier or later date. Therefore, the question we have to address is when this will occur.
The most recent DEI forecasts indicate a 32% probability of Unit 2 reaching the plugging limit by RFO 17 in July 2011, and a 70% probability of reaching the plugging limit by RFO 18 in April 2013. These forecasts also indicate a 46% probability of Unit 3 reaching the plugging limit by RFO 19 in January 2016. This means that without the SGRP, there is approximately a 50% probability that Unit 2 will operate until mid-2012, and that Unit 3 will operate until the beginning of 2016. We find these most recent DEI forecasts reasonable and will use them in our cost-effectiveness analysis without SCE's adjustments. We note that if SCE was to apply for and be granted a higher plugging limit by the NRC, the original steam generators would be allowed to run longer. We also note that SCE has not done so.
SCE states that it has reasonably and prudently maintained the original steam generators. Therefore, SCE believes it should recover all reasonably incurred capital costs for SONGS. In addition, it states that it should recover all reasonably incurred SGRP costs if the Commission denies this application.
TURN points out that an assumption underlying SCE's cost-effectiveness calculation is that, if SONGS shuts down at any time prior to the end of its license lives, the undepreciated plant balance will remain in ratebase and be fully recovered from ratepayers. TURN asserts that in D.85-08-046, the Commission concluded that the early shutdown of Humboldt Bay Unit 3 (Humboldt), a nuclear power plant, resulted in investment that was no longer used and useful and, therefore, excluded the undepreciated plant costs from ratebase. PG&E was allowed to recover plant costs, but was not allowed to earn a return on the unrecovered amount. TURN also points out that in D.92-08-036, the Commission adopted a settlement regarding the early shutdown of (SONGS) Unit 1 that allowed SCE to recover its remaining investment, but only allowed a return on the unrecovered amount equal to the embedded cost of debt. Based on the above, TURN recommends that SCE be required to run its cost-effectiveness model assuming the treatments adopted in D.85-08-046 and D.92-08-036.
Aglet believes that recovery of net plant costs in the event of an early shutdown is not assured. It states that the Commission has no firm policy on this matter, and that full recovery is unlikely.
In D.85-08-046, the Commission addressed the recovery of the remaining undepreciated plant investment in Humboldt, which was shut down before the end of its license life.25 The Commission allowed a four-year amortization of the remaining unrecovered plant investment without a return on the unamortized balance during the amortization period.
In D.92-08-036, the Commission addressed the recovery of remaining undepreciated plant investment for Unit 1, which was shut down before the end of its license life. The Commission adopted a settlement that allowed a four-year amortization of the remaining undepreciated plant investment. It also allowed a return equal to the embedded cost of debt on the unamortized balance during the amortization period. Since this decision adopted a settlement, it did not set a precedent.
It is possible that, in the event of an early shutdown, the undepreciated plant balance may be amortized over a four-year period with a reduced or no return on the unamortized balance. However, we normally base depreciation rates on the remaining life of the asset being depreciated. Therefore, it is also possible that depreciation rates for SONGS, in the absence of the SGRP, will be increased based on the shorter expected life. If that is done, the remaining undepreciated capital costs will be fully recovered over its remaining life with a return earned on the undepreciated balance. At this time, it is premature to make these determinations, and there is no fixed policy as to how any undepreciated plant balance would be recovered, if at all. Therefore, we will calculate the cost-effectiveness of the SGRP without assuming a limitation on capital recovery if the SGRP is not performed.
The discount rate is used in this proceeding to determine the present value, in 2004 dollars, of future expenditures. SCE uses a 10.5% discount rate in this proceeding, which is its estimate of its incremental cost of capital. SCE states that the discount rate is higher than its authorized cost of capital because it is a forward-looking long-term cost of capital. SCE argues that it would be inappropriate to use its authorized cost of capital because it is a short-term average cost of capital, and does not reflect the cost of new or incremental capital.
Aglet states that SCE's discount rate is based on speculation as to the incremental cost of debt and equity. Aglet also points out that the Commission has never endorsed the incremental cost of capital as a basis for cost-effectiveness analysis. Aglet states that it would be more reasonable to assume that customers, especially low-income customers, have higher discount rates.
The authorized cost of capital is often used as a discount rate to evaluate cost-effectiveness. Most of the costs in this cost-effectiveness evaluation occur in the early years of the SGRP, whereas most of the benefits occur later. Therefore, the use of a higher discount rate would tend to make the SGRP less cost-effective, and the cost-effectiveness analysis more conservative. In this case, SCE's recommended discount rate is higher than its authorized cost of capital, and no party has recommended a specific discount rate higher than 10.5%. The record is not sufficient to determine whether, in theory, an incremental cost of capital is more appropriate as a discount rate than the authorized cost of capital. Nevertheless, since SCE's recommended discount rate does not appear likely to overstate the cost-effectiveness of the SGRP, we find it reasonable and will use it in our cost-effectiveness analysis.
The current ownership shares of SONGS are:
· SCE 75.05%
· SDG&E 20.00%
· City of Anaheim (Anaheim) 3.16%
· City of Riverside (Riverside) 1.79%
SDG&E has indicated its intention, pursuant to the operating agreement, not to participate in the SGRP. However, its ownership share will be reduced accordingly, with a corresponding increase in SCE's ownership share. Although they do not agree on the amount of the reduction, SCE and SDG&E have agreed that SDG&E's likely remaining ownership share will be 0-14% if the SGRP goes forward.
SCE states that it performed its cost-effectiveness evaluation of the entire project, as opposed to only its ownership share. It says that this is appropriate because, if the overall benefits of the project exceed the overall costs, there are sufficient benefits to justify the project even if the distribution of benefits between the owners is not uniform.
TURN states that the SGRP should be evaluated assuming that SDG&E does not participate. TURN recommends that, since SCE and SDG&E submitted their dispute regarding the ownership share reduction to arbitration, the results of the arbitration should be used to evaluate the cost-effectiveness of the SGRP. TURN also states that, since Anaheim has decided not to participate in the SGRP, the results of Anaheim's non-participation should be considered in evaluating the cost-effectiveness of the SGRP.
Aglet states that the uncertainty about the economics of SDG&E's decision not to participate in the SGRP contributes to the uncertainty of the SGRP's cost-effectiveness for SCE's customers.
SDG&E plans to file a § 851 application for approval of its resulting ownership reduction pursuant to the SONGS ownership agreement. SDG&E states that it will only participate in the SGRP if the Commission finds in that proceeding that it should do so.
While SCE and SDG&E have submitted their dispute to arbitration, the result of the arbitration is not binding. It remains for the Commission to decide in a § 851 application whether SDG&E should participate in the SGRP and if not, what the ownership share reduction should be. Therefore, the arbitration results will not be considered herein. As a result, we will evaluate the cost-effectiveness of the SGRP assuming the 0-14% range of ownership by SDG&E.26
Anaheim has also decided not to participate in the SGRP. The record does not indicate what its likely ownership will be as a result of its non-participation in the SGRP. Therefore, we will use an ownership range for Anaheim that is proportionately similar to SDG&E's. This results in an ownership range of approximately 0-2.2%.27
As a result of the decisions by SDG&E and Anaheim not to participate in the SGRP, SCE's ownership share will increase in the same amount SDG&E and Anaheim's will decrease. As a result, SCE's ownership will range from 82.00% to 98.21% with a midpoint of 90.10%. We will consider this ownership range in our cost-effectiveness evaluation.
In its application, SCE assumed that if the SGRP is not performed, Units 2 and 3 will shut down at the same time. The Commission's Office of Ratepayer Advocates (ORA) recommends that a split shutdown scenario be considered. A split shutdown scenario is where the SGRP is not performed and each unit is shut down as it reaches the NRC imposed plugging limit. Unit 2 has more tube degradation than Unit 3. Therefore, under the split shutdown scenario, Unit 2 would shut down first and Unit 3 would remain in operation for a longer period.
SCE states that it must maintain systems and facilities that are shared by both units so long as either unit is in operation. SCE represents that this would substantially increase the base O&M required for the remaining operational unit. SCE also represents that the NRC would require mid-cycle outages once one of the units is shut down, which would increase the RFO O&M expenses.28 In other words, shutting one unit down reduces the O&M expenses by less than half. SCE states that if the SGRP is not performed, it would be more cost-effective to shut both units down when one of them reaches the plugging limit than to keep the remaining unit running.
The record demonstrates that Unit 2 will likely shut down before Unit 3, and that Unit 3 can be operated when Unit 2 is shut down. Therefore, we will include a split scenario shutdown of SONGS in our cost-effectiveness evaluation of the SGRP to determine whether it is more appropriate than shutting both units down at the same time.
SDG&E proposes to sell its entire interest in SONGS to SCE, and take back a purchase power contract (PPC). SDG&E states that this would be cost-effective for SCE's customers because the terms of the PPC would financially motivate SCE to manage SONGS costs.
Whether SDG&E should participate in the SGRP is not at issue in this proceeding. Likewise, the sale of all or part of SDG&E's ownership share to SCE is also not at issue. Therefore, we will not address SDG&E's proposal in this proceeding.
SDG&E proposes that SCE be required to form a partnership for SONGS for tax purposes. The purpose of the partnership would be to avoid possible income tax consequences to SDG&E and SCE of a transfer of some portion of SDG&E's ownership share in SONGS to SCE.
The sale of all or part of SDG&E's ownership share to SCE is not at issue in this proceeding. Therefore, the tax consequences of such a sale to SDG&E are not at issue. In addition, SDG&E has stated that the consequences of such a sale to SCE would not affect the cost-effectiveness of the SGRP for SCE. As a result, we will not address SDG&E's proposal in this proceeding.
Aglet notes that SCE includes in its cost-effectiveness analysis the air quality benefits of nuclear power through calculation of a carbon adder, but does not include unquantified costs resulting from risks associated with the additional spent fuel that will be generated by SONGS due to the SGRP. Aglet points out that public health risks are inherent in nuclear power plant operations.
In D.04-12-048 and D.05-04-024, we adopted a Green House Gas (GHG) adder for carbon dioxide emissions to be used when comparing fossil generation to non-fossil generation in utility resource plans, and energy efficiency programs. The purpose was to explicitly account for the financial risk associated with GHG emissions. However, those decisions did not address major repairs to nuclear power plants as an alternative to fossil fuel fired generation. Therefore, those decisions did not address whether and to what extent the GHG adder should be applied to this proceeding.
CCGTs will produce the emissions the GHG adder is intended to address. However, nuclear power plants have their own safety, public health, and environmental risks and effects. Inclusion of such risks and effects, if they could be quantified, would decrease the cost-effectiveness of the SGRP. However, nothing in the record places a dollar amount on such risks and effects. At the same time, inclusion of the GHG adder would increase the cost-effectiveness of the SGRP. Therefore, we will consider both the GHG adder and the safety, public health, and environmental risks and effects associated with SONGS in our cost-effectiveness evaluation of the SGRP.
Aglet points out that SCE claims the SGRP will avoid statewide natural gas price increases due to a greater demand for gas if the SGRP is not performed. Aglet agrees with the concept, but states that the SGRP will increase the demand for the goods and services necessary to perform the SGRP, which will raise the prices for such goods and services. Aglet states that there is no reason to believe that this effect does not counterbalance the effect on natural gas prices. As a result, Aglet recommends that no effect of the SGRP on natural gas prices be considered in the cost-effectiveness evaluation.
Aglet's argument that the SGRP will affect the prices for goods and services involved in the SGRP is correct in theory. The goods and services indicated by Aglet are the replacement steam generators, the materials and labor to install them and remove and dispose of the original steam generators, and future O&M costs. SCE's SGRP could affect the prices for these goods and services associated with other SGRPs occurring at about the same time. However, since SCE customers will not be paying for other SGRPs occurring at about the same time as the SONGS SGRP, they will not be affected. While it is possible that the SGRP could affect the costs for goods and services other than those associated with other SGRPs, the record does not indicate that there would be any significant effect on SCE's customers. Therefore, we will not adopt Aglet's recommendation.
The SGRP, if it is approved, will be paid for by the customers of SONGS owners, not the state as a whole. The cost-effectiveness evaluation in this proceeding is limited to SCE's customers. Therefore, we will not consider the effect on other customers.
10 100% level is the total for the project. Individual participating owners' shares will be a portion of this amount. These numbers exclude construction financing costs, and allowance for funds used during construction.
11 In this proceeding, SCE defined "contingencies" as a specific provision for unforeseeable costs within the defined project scope. SCE estimated $131 million for such contingencies. SCE also added an additional $10 million for growth in the scope of the SGRP, which it calls "additional adjustments." For simplicity, we use the term "contingencies" in this decision to include both the contingencies and additional adjustments amounts included by SCE.
12 SCE's 2006 GRC is Application 04-12-014.
13 $40 million is about 10% of the recorded total 2004 O&M costs.
14 The SONGS Board of Review consists of representatives of the owners of SONGS. It oversees the SONGS budget and other matters related to SONGS.
15 CEC does not say what the annual O&M costs would be after shutdown, but presumably they would be $16 million as in the second and third scenarios.
16 This scenario assumes that the enhanced security requirements include more stringent steam generator tube integrity requirements that lead to shutdown in three years.
17 Out of 102 operating nuclear plants in the United States, less than one percent experienced a shutdown of one year or longer in the last seven years. Less than 15% experienced a shutdown of one year or longer in the last 15 years.
18 The heat rate is the amount of heat in British thermal units (Btu), from burning natural gas, that is necessary to generate one kilowatt-hour (kWh) of electricity.
19 Voltage support equipment is installed to maintain transmission system voltage. The capacity of the equipment is stated in units of reactive power called volt-amperes reactive (VARs). MVARs means millions of VARs.
20 The forecast may also be expressed as the percentage probability that a unit will continue to operate until a specified date. For example, a 25% probability that a unit will exceed the plugging limit at a specified date is equivalent to a 75% probability that it will continue to operate until that date.
21 This means that there is a 67% probability that Unit 2 will operate until RFO 17 in 2011.
22 This means that there is a 56% probability that Unit 3 will continue to operate until RFO 20 in 2017.
23 A more recent DEI forecast for Unit 3 was not available until after reply briefs were filed.
24 The Commission has not previously adopted such subjective adjustments.
25 Humboldt shut down in 1976 with the expectation that it would eventually be returned to service. It remained in ratebase until December 1979 when it was removed from ratebase.
26 Pursuant to the SONGS ownership agreement, the reduction in the ownership share will take place after the SGRP is completed.
27 Since Anaheim is not a regulated utility, a § 851 application is not required.
28 A mid-cycle outage is a scheduled outage approximately mid-way between RFOs for the purpose of performing tube inspections.