6. URG - Hydroelectric (Hydro) Generation

During the Record Period, SCE operated and maintained 33 hydro generating plants including 33 dams, 43 stream diversions, and approximately 143 miles of tunnels, conduits, flumes and flow lines. These resources have an aggregate 1,176 MW of nameplate generating capacity. SCE has provided information on the characteristics of its hydro generation resources, organization of the Hydro Division, recorded hydro production, and operating results of its facilities. Approximately 86% of SCE's hydro generation is provided by SCE's Northern Hydro Division. Despite below-average precipitation levels in 2009, SCE's hydro plants generated over 90% of the long-term historic annual average.

In its testimony and brief, DRA found that two hydro forced outages were unreasonable. The first was an outage in Northern Hydro, also known as Big Creek, which took place on December 14, 2008 in Big Creek 3, Unit 1. The second was an outage at the Mammoth Pool Generator which lasted from June 11, 2008 until May 1, 2009.

DRA does not challenge any other SCE hydro operations as unreasonable. Based on the testimonies of SCE and DRA, we also conclude that all other SCE hydro facilities were operated reasonably during the Record Period.

6.1. Big Creek 3, Unit 1 Outage

The Big Creek 3 powerhouse relayed off-line on December 14, 2008 apparently due to a 15 KV transfer switch fault in a Station Service Transformer switch cubicle. The outage lasted approximately one year. The switch (which is several cubic feet in size) is housed in a metal enclosure, located outside in a cabinet on the transformer deck just outside the powerhouse. The switch fault may have been the cause of a fire in Unit 1 generator stator. The fire caused damage to the stator coils and iron core, requiring a generator stator rewind.¹⁸ DRA contends that SCE's failure to maintain a dry environment around high-voltage switchgear, failure to be mindful of equipment history, and imprudent design at the plant led to approximately $10.2 million in power replacement costs, which should be disallowed.

DRA claims that the record shows the following:

· Because the transfer switch was outside and exposed to inclement weather, decades of service can cause deterioration or corrosion to the equipment.

· The 15 kV transfer switch showed evidence of corrosion and moisture, which can accelerate a final insulation failure.

· The cabinet is surrounded by an oil containment barrier which the switchgear was never designed to be placed in.

· An inch of standing water in the cabinet containing the transfer switch was coincident with the ongoing rainstorm at the time of the failure.

· The fact that high voltage switchgear had standing water in it is not a good practice by design.

· The transfer switch gear was compromised, and corrosion and moisture were present.

Therefore, DRA concludes that SCE knew or should have known that this combination of factors could cause any number of problem situations in the electrical system, up to and including a fire in a generator. DRA lists several precautions SCE could have taken to protect the transfer switch and, in turn, the generator.

SCE states that its report on the incident indicates that the fault may have initiated in the switch, but that this has not been established conclusively. While SCE agrees with the basic facts laid out by DRA (which stemmed from SCE's exhibits), SCE also adds the following:

· Large portions of the power grid must be located outdoors because of economic necessity.

· The switch had been subjected to other rainstorms and fog episodes since it was installed in the 1970s and since the oil containment curbs were installed in the 1990s.

· It is unknown whether the switch's compromised insulation was present before the fault or appeared because of the fault itself. Absent obvious degradation, SCE personnel had no basis to conclude the insulation needed repair or replacement.

· SCE personnel had no record prior to this event that standing water had ever been present in the cabinet before and had no reason to believe that this particular switch would fail during this particular rainstorm.

· The windings in the generator stator that faulted were scheduled to be replaced in 2012. The generator was inspected, passed operability testing, and was expected to operate through the time for its planned rewind.

Based on these additional facts, SCE claims that there is no basis to conclude that SCE could have done anything other than maintain, test, and operate the generator consistent with its historic practices to prevent this outage.

6.2. Discussion

There is persuasive evidence that the failed switch caused the fire and outage in Unit 1, given that considerable damage to the switch was found around the switch concurrent with the damage to the generator. On the contrary, there is no persuasive evidence that the alternative suggested by SCE - an electrical fault initiated within the generator -- was the most likely cause of the fire and outage. We conclude the failed switch most likely was the cause of the outage.

The question is whether SCE, in the context of the reasonable manager standard, acted reasonably in the time period before the switch failed based on information available at the time. The main questions are whether SCE should have been aware of corrosion in the insulation, and should have known that a rainstorm resulting in an inch of standing water could cause a compromised switch to fail.

DRA in its brief suggests SCE should have taken more precautions to protect the transfer switch and, in turn, the generator. DRA contends that SCE could have covered the outdoor switch cabinet to protect it from the elements. Certainly, this is true, but it cannot be considered unreasonable for SCE to locate this equipment outdoors, since accepted industry practice has been to place many types of equipment (including this type of transfer switch) outdoors over many years. DRA contends SCE could have inspected the equipment more often. Again, this is true, but DRA offers no evidence that SCE failed to timely or properly inspect the equipment at issue. DRA suggests SCE could have replaced the transfer switch more often, but offers no rationale for why SCE should have done so in this case.

If SCE had known that the equipment was at risk, it would have been unreasonable to not take corrective action. SCE is incorrect that there was nothing it could have done to prevent the incident from occurring. But there is not sufficient evidence that SCE had information before the incident upon which a reasonable manager should have acted. Therefore we cannot conclude SCE's actions were unreasonable on this point.

More troubling is the presence of one inch of standing water in the cabinet at the time of the incident. SCE's investigators recommended forward-going redesigns so that contaminants and water cannot cause a phase-to-ground fault. SCE agrees that, as a matter of practice, standing water that accumulates in the containment area should be drained as soon as practicable following a rainstorm. SCE is correct to apply such redesigns now; the question is whether SCE should have proactively designed or redesigned its equipment before such an incident could occur.

If SCE had known that there was a reasonable likelihood that standing water could cause a fault in the switch, SCE should have taken proactive measures to prevent this situation from occurring. It is possible that SCE was aware of, or should have been aware of, this possibility, given that a similar event occurred in the 1988-1989 winter season in a different location. However, DRA has not shown that lessons from the incident 20 years before should have or could have been applied to this incident, or that applying such lessons would have prevented the incident from occurring. Because SCE is now aware that such a circumstance can occur, it is likely that application of the reasonable manager standard in the future would result in a disallowance in similar circumstances if proactive measures are not now taken.

We conclude that SCE's actions were reasonable with regard to the Big Creek 3, Unit 1 outage.

6.3. Mammoth Pool Outage

Mammoth Pool is a powerhouse that includes two hydro generating units, each with 95 MW capacity. Consistent with most large generators, the Mammoth Pool Generator has two sets of windings, one for the rotor and one for the stator. On June 11, 2008, after a Mammoth Pool Unit 2 turbine overhaul, SCE cleaned Unit 2's stator windings and initiated a three phase test known as a high potential test ("hi-pot") to examine the stator's operability. The Unit 2 stator windings failed the test and had to be replaced. The unit remained out of service until May 1, 2009.

DRA contends that SCE's own internal report investigated the stator winding failure to determine what caused the failure, which occurred approximately 17 years into the equipment's expected 30 year lifetime. DRA claims the report concluded that the stator windings were damaged because the generator was consistently operated at higher than normal temperatures, which escalated the thermal aging of the equipment.

DRA cites SCE's Stator Winding Failure Executive Summary (SCE Summary),¹⁹ which reached three conclusions about the stator winding: 1) The equipment failed prematurely; 2) the premature failure was due to thermal aging that was taking place at escalated levels; and 3) excessive thermal cycling of the equipment also contributed to the premature failure of the equipment. DRA contends that these three conclusions point to negligence and unreasonable use and operation of the generator by SCE absent any justification for the manner in which equipment was operated.

Specifically, DRA contends that SCE's unreasonable conduct in the failure of the stator windings was that SCE knew or should have known that operating the generator consistently at 90ºC and with frequent or excessive cycling would cause thermal aging which would prematurely damage the stator winding. Further, DRA contends that SCE failed to conduct more frequent planned maintenance of the stator winding given its heightened use and operation.

DRA cites the SCE Summary as stating: "A major contributor to the premature failure of this winding is that thermal aging is taking place at an escalated level. The unit is normally run at and limited by the 100ºC temperature limit. Thermal aging does take place at 90ºC and above. Thermal aging escalates exponentially with temperature."²⁰ The SCE Summary also states: "Thermal aging is also a contributor to the premature failure."²¹ DRA also cites from SCE's report from Voith Hydro (Voith Report),²² the manufacturer of the stator winding, which investigated the reasons for the outage, to support its recommendations. DRA's brief quotes the Voith Report as stating: "For a period of 4-5 years, the unit operated with a stator winding temperature of 110 -115ºC, but more recently, the unit operated around 100ºC, with an alarm at 115ºC."²³

DRA contends that SCE claims that it may have run the generator excessively to meet customer demand, but that there is no evidence to support this claim. Further, even if that was the case, DRA contends that SCE had a duty to maintain its equipment at a level consistent with the usage that the equipment is subject to. DRA also contends that SCE provided no evidence that it made efforts to balance longer life versus higher output.

SCE contends that all actions taken by plant personnel related to the outage were prudent, reasonable and consistent with the plant's historical operating and maintenance practices. SCE agrees that two after-the-fact reports of the outage make clear that thermal aging and thermal cycling were the major contributors to the stator's failure. However, SCE contends that the precise effect of these factors on stator life expectancy is unknown.

SCE agrees that the effects of thermal aging increase at a higher rate whenever temperatures exceed 90ºC, but claims that the winding manufacturer recommended an operating temperature of 100ºC on hot days to achieve the longest winding life, while accommodating production needs. SCE argues that running the generator at more than 100ºC on some seasonal peak days was necessary to increase generation production and for other operational factors. Because of these factors, SCE agrees that it ran the generator at 110 -115ºC in the early to mid 1990s on some hot summer afternoons.

SCE contends that it had no reason to conclude that the generator windings were approaching the end of their service life, because tests performed in 1993, 1996, and 2003 did not raise any concerns about remaining life expectancy. SCE contends that two prior winding failures in 1983 and 1990 were due to manufacturing defects and unrelated to the 2008 failure.

In its Reply Brief, SCE contends that it had no reason to believe that occasionally operating the generator above 90ºC would cause the windings to fail several years before a more typical winding service life. Thus, SCE claims it had no reason to conduct such an analysis because SCE had no reason to expect the windings to deliver a service life that would be significantly different than that of all SCE's other hydro generators.

6.4. Discussion

The Voith Report shows that the Mammoth Pool Unit 2 generator operated with a stator winding temperature of 110-115ºC for a period of 4-5 years. In more recent years, the unit operated at temperatures around 100ºC, with limited periods at higher temperatures. Two post-failure analyses concluded that the Mammoth Pool Unit 2 generator stator windings likely failed the hi-pot test due in part to degradation from time periods when operating temperature exceeded 90ºC. The question is whether SCE acted as a reasonable manager in the period before the failure of the hi-pot test. Specifically, we seek to determine whether SCE knew or should have known that operating the generator at high temperatures would cause thermal aging and premature failure.

SCE is correct that the post-failure analyses findings were not known to SCE at any other time leading up to the 2008 winding failure experienced during routine operability testing. However, the record shows that SCE did have information available at the time which it should have considered when setting the temperature at the generator.

SCE in its Rebuttal Testimony states: "The standard design temperature of 110ºC follows a generic rule-of-thumb for balancing generator winding life (that degrades whenever temperatures exceed 90ºC) with generator MW output on hot days."²⁴ SCE further states that "both longer life and higher MW output have economic value, and the operator must balance the two factors."²⁵

SCE was aware that it was running the generator at higher than recommended temperatures over several years. SCE claims that it ran the unit at these higher temperatures in order to achieve higher output, but SCE does not show that it performed any studies to determine if the higher output was worth a shorter plant life, either to shareholders or to ratepayers. We conclude that SCE, in the absence of such studies, unreasonably chose to operate the generator at higher than recommended temperatures for extended periods, thus knowingly diminishing the life of the plant. While the particular failure was unexpected, a reasonable manager should have known that, by running the generator at higher than recommended temperatures for extended periods, the generator was likely to fail far in advance of its expected useful life. SCE should have taken measures to mitigate this outcome, including adhering to temperature guidelines from the manufacturer and consideration of an earlier stator rewind.

DRA recommends a disallowance of approximately $7,691,411 for the outage related to the valve at Unit 2. SCE places the upper limit of replacement energy at $693,800. DRA's estimate is based on an estimated upper limit of loss of 231,840 MWh. SCE's estimate is based on an estimated upper limit of loss of 27,069 MWh. One difference between the calculations is related to the capacity factor used for the generator. DRA uses a capacity factor of 100% (i.e., DRA assumes the plant runs 100% of the time), while SCE witness Kurpakus testified that the annual capacity factor for Mammoth Pool is approximately 33%.²⁶ The other difference arises because SCE claims that when one hydro unit is out of service, SCE has the ability to utilize water from other sources in the overall Mammoth Pool Generator project, thereby continuing to produce electricity,²⁷ known as "outage bypass energy."

In its Reply Brief, DRA accepts SCE's annual average capacity factor of 33% for the Mammoth Pool generator. DRA also accepts a 50% reduction to take into account seasonal factors and outage bypassed energy. Thus, DRA calculates a lost energy value of 38,254 MWh, and a disallowance of $979,350. We will accept DRA's modified calculation for this disallowance.

¹⁸ Exhibit SCE-8, Appendix L.

¹⁹ Exhibit SCE-8, Appendix O.

²⁰ Id, at 3

²¹ Ibid.

²² Exhibit SCE-8, Appendix P.

²³ Id, at 2.

²⁴ Exhibit SCE-7, at 79:16-17.

²⁵ Exhibit SCE-7, at 79:16-17.

²⁶ RT Vol. 1, at 148, 9-24.

²⁷ Exhibit SCE-7, at 88, line 21 through at 89, line 2.