5. PG&E's December 2002 Storm Response

In December 2002, PG&E and its customers were affected by a series of severe storms that brought heavy rains and hurricane-force winds to central and northern California. Two severe storms, occurring on December 14 and December 16, were followed by two additional storms of lesser magnitude on December 19 and December 21; all four storms were characterized by high wind conditions and periods of heavy rain and thundershowers. During those storms, thousands of PG&E customers experienced power outages, with many customers experiencing outages for extended periods of time.

PG&E states that the storm on December 14, 2002, was one of the worst 10 storms since 1995 and the December 16, 2002, storm was the second most severe storm to hit PG&E's service territory in the last 20 years.¹¹ The December 14 and 16 storms were characterized by sustained winds reaching 49 miles per hour and lasting up to 11 hours. Peak wind speeds during these two storms reached 100 miles per hour and over 10 inches of rain fell. The storms of December 19 and 21 delivered wind gusts of 30 to 50 miles per hour throughout the service area, accompanied by up to two inches of additional rainfall.

The series of storms caused significant damage to PG&E's electric distribution facilities, resulting in 1.97 million customer interruptions. PG&E reports that 1.3 million customers experienced service interruptions, with approximately 900,000 customers experiencing one outage, approximately 250,000 customers experiencing two outages, and approximately 150,000 customers experiencing three or more outages over the course of the storms. Some customers experienced outages that extended for several days.

PG&E's distribution system suffered damage to 2,056 distribution poles and pole hardware, 817 distribution transformers, and 3,884 spans of primary, secondary, and service conductors. PG&E's transmission system suffered damage to approximately 237 wood poles, one transmission tower, and 267 spans of transmission conductor. Over 3,600 field workers responded to restore service in harsh wind and rain conditions. The majority of the storm damage was tree-related.¹²

During the nine-day period beginning December 13, 2002 and ending December 21, 2002, PG&E's call centers and representatives received 1.8 million calls, compared to the 300,000 that would normally occur over an eight-day period. Of the 1.8 million customer calls received, call center representatives handled approximately 290,000 customer calls and IVR units handled approximately 1.5 million calls. Many customers contacted PG&E and the Commission to express concern about multiple, extended outages and their difficulties in contacting PG&E or receiving information from PG&E.

PG&E claims that its response to the December 2002 storms was reasonable. PG&E states that the December 2002 storms were unusual in that the two weather systems on December 14 and 16 were severe, and close together in time. PG&E states that the multiple storm fronts resulted in many circuits suffering repetitive damage, forcing PG&E to continually reprioritize restoration efforts and resource deployment plans in order to make conditions safe and assess the new damage and repair requirements. The resulting effect was that many customers were interrupted more than once and many other customers suffered extended outages due to the resource deployment needs (e.g., crews planned to be deployed to smaller outages from an earlier storm were redirected to larger outages from subsequent storms).

PG&E maintains that the outages that occurred during the storms were not caused by a lack of maintenance or a lack of tree-trimming. PG&E states that it currently trims or removes approximately two million trees per year, yet 60 percent of the outages were tree-related. PG&E also reports that it has increased the number of pole replacements since 1999, yet over 2,000 poles or associated pole hardware were damaged during the storms. PG&E believes that this indicates that the magnitude of the damage that occurred was unavoidable and cannot be attributed to a lack of maintenance, lack of tree trimming, or lack of pole replacements.

PG&E demonstrated that it met the ASA standard from 1996 through December 2002, with small variances in January 2001 and February 2001 during the energy crisis. During the storm period, from December 13-21, 2002, the call centers maintained a 21-second ASA, and only 236 customers received a busy signal. PG&E believes that this result demonstrates that its call center performance was reasonable. However, PG&E admits that following the conversion to the new CIS in December 2002, the call centers experienced a substantial increase in the time required to handle customers' calls and PG&E has not been able to meet the ASA standard since.

PG&E also argues that its OIS performed adequately during the December 2002 storm event. PG&E bases this claim in part on the fact that the average restoration time for all outages was less than 8 hours, more than 1.5 hours less than the performance standard set in GO 166. PG&E acknowledges that the difference between the first ETOR and the actual restoration time during the December 2002 storms was considerably larger than the differences observed during other major events but PG&E believes that this difference is largely due to the multiple severe storms causing ETORs to be revised several times by the time power was actually restored.

PG&E also admits that it identified several problems associated with its field restoration, outage communications, and call center systems during the December 2002 storms.

The field restoration and outage communications problems are listed below.

· PG&E learned from various police and fire departments that they were dissatisfied with PG&E's handling of their emergency calls regarding hazards associated with PG&E's facilities during the storms. PG&E admits that calls from emergency personnel were not handled consistently among the various Operations Emergency Centers, resulting in some police and fire personnel standing by hazardous conditions for excessive periods of time during the storms.

· Some customers who suffered service interruptions during the first storm remained out of power when the second storm hit, and some were placed in the back of the queue for outage restoration.

· The degree of detail provided by the field assessment personal varied considerably during the December 2002 storm event. Since the written assessments were not always detailed enough to accurately determine the resource and material needs, the ETOR was negatively affected.

· PG&E has three separate computer systems that require data to provide assistance in outage management and communication. These three systems are not currently connected, and require three separate entries to store complete outage information. Outage restoration is delayed by the time it takes to input data into these three systems, rather than into one integrated system.

· The circuit maps currently in the OIS are not always detailed enough to identify the specific transformer affected by the outage. In addition, due to the level of system detail currently in OIS, the OIS may indicate a greater number of customers out of service than are actually experiencing an outage. The overall effect is providing potentially inaccurate information to customers.

· During the December 2002 storms, some customers were informed that PG&E could not find a record of their earlier call. OIS software that addresses single customer outage information was originally configured so that a single customer outage would be "aged off" after 30 minutes. The "aged off" outages would not be prioritized along with multiple customer outages, but would be managed within the FAS. During the December 2002 storms, when the system "aged off" the single customer outage, it inadvertently deleted the recorded history of that outage. In an effort to temporarily resolve this problem, the aged off feature was disabled. The resulting number of single customer outages on the OIS made the system slower and the dispatcher's job more difficult. PG&E applied a software change to eliminate the record from being deleted and allowed PG&E to turn the "aged off" script back on.

PG&E's call centers also experienced some technical complications during the December 2002 storms as follows:

· A small number of customers received a busy signal on December 14, 2002 (236 callers of 414,903 calls received a busy signal).

· Some customers expressed concern that after making an initial call to report their outage in the Outage VRO system, they were presented with a message stating "We are not aware of a power outage at your location" when making a subsequent call to obtain the ETOR update.

· Some callers experienced a period of silence and a call ending message when waiting for an outage status message to play in the VRU.

· Some customers experiencing outages said they did not have an option to talk with a service representative when they called.

· Some outage restoration callbacks by the VRU system were delayed because of the high volume of callback requests.

· Some customers were unaware of the significance and advantage of having their phone number listed on their account record in PG&E's CIS.

To address these problems, PG&E identified and implemented several process and technology improvements intended to improve its outage communication and field restoration process as well as its call center performance during future storms. The identified changes to the outage communication and field restoration process include:

· PG&E is implementing a process by which all calls from emergency agencies will be routed to service dispatch operators located throughout PG&E's service territory. These personnel will be trained to handle these calls during both routine and emergency operations, which will ensure consistent prioritization of these calls.

· PG&E has modified its restoration prioritization to balance the length of time small numbers of customers are out of power with the need to restore the largest numbers of customers. Specifically, PG&E intends to identify emergency response personnel in each of its OECs who are responsible during major storms to focus attention on small primary level outages and single customer outages.

· Utilizing mobile data terminal units currently in Troubleman and Gas Service Representative trucks to accelerate the input of outage cause and damage assessment information to the Operations Emergency Centers.

· Integrating three of the company's internal information control systems to reduce the number of entries required by a system operator. The three systems that would be integrated include the Supervisory Control and Data Acquisition System (SCADA), the Distribution Operators Logging Information Program (DOLIP), and the OIS.

· Improving the mapping associations within OIS to better provide accurate numbers of customers affected by outages and more accurate outage information on a real time basis.

· Eliminating the "aged off" feature of the OIS and developing software to more efficiently manage single customer outages.

PG&E has also taken steps to eliminate or reduce the occurrence of call center problems. For example, PG&E states that it has re-programmed the VRU system to preclude customers from experiencing a period of silence and call ending message when waiting for an outage status message. PG&E has also added a special toll-free number for customers who are without power for 48 hours or more to improve the ability of those customers to speak with service representatives. PG&E has undertaken a communication campaign urging customers to ensure that the correct phone number is on record for their service address to allow customers to better receive timely and accurate outage information. PG&E also states that it is taking action to reduce its call center ASA, including 1) hiring and training an additional 100 service representatives, 2) improving service representatives' call handling efficiency, and 3) improving the customer information system's performance.

PG&E states that its overall response to the December 2002 storms is consistent with the "adequate service" standard discussed in the last PG&E GRC decision and in the context of past Commission decisions addressing the reasonableness of storm response activities, including D.95-09-073, the decision on PG&E's response to the January and March 1995 storms, D.99-06-020, the decision on PG&E's response to the December 1995 storm, and the decisions adopting standards and requirements for storm response, inspection, maintenance and reliability. (D.96-09-045 and D.98-07-097)

PG&E maintains that throughout this phase of the GRC, no party presented evidence demonstrating that PG&E's response to the December 2002 storm event was unreasonable or that PG&E's maintenance activities or work practices contributed to the amount of damage incurred.¹³ PG&E argues that the severe damage suffered by its system cannot be attributed to a lack of maintenance, lack of tree trimming or lack of pole replacements. PG&E argues that it effectively utilized the resources at its disposal to restore service caused by the series of storms that began on December 12, 2002. PG&E points out that although ORA and its consultant Stone and Webster asserted that various factors associated with PG&E's performance negatively affected service restoration, neither ORA nor Stone and Webster claimed that these factors resulted in unreasonable performance. PG&E also points out that the PG&E/ORA joint testimony identifies improvement opportunities and plans to better document and measure performance, but does not include any findings of lack of compliance with any existing standard or code.

Although PG&E maintains that its response to the December 2002 storms was reasonable, it recognizes that there has been public dissatisfaction with its storm performance. PG&E acknowledges that the damage cause by these successive storms resulted in many customers experiencing multiple interruptions of service. Because of the severity of the damage, the high number of outages, and the number of repeat outages, approximately 50,000 customers were without power for more than two days.

PG&E explains that because the storms did not meet the definition of a "Major Outage" under GO 166, the service restoration standards and call center standards applicable to measured events do not apply. Nevertheless, PG&E notes, its performance during the storm met the requirements of GO 166. PG&E also notes that its reliability performance in 2002 (excluding major events) as measured using the SAIDI and SAIFI was better than any year going back at least 10 years. The outages associated with the December 2002 storms are not included in the calculation of SAIDI and SAIFI for 2002, however, because they met the Commission's definition of "excludable major event" since more than 10 percent of PG&E's customers were without power during each of these events.

ORA agrees with PG&E that the December 2002 storms were severe, and that the current definition of "Major Outage" within GO 166 prevents the restoration benchmark from being applied to the December 2002 storm event. ORA finds, however, that certain of PG&E's actions may have contributed negatively to the outage durations experienced by customers and offers several recommendations designed to improve PG&E's performance, as well as the Commission's ability to effectively evaluate PG&E's performance. In particular, ORA believes that PG&E's restoration performance would have been more effective if PG&E had mobilized and deployed additional staff earlier in the storms, provided better training to trouble assessors, and developed a clear policy to ensure that no customers are left without service for an inordinate amount of time. ORA highlights the significant differences in performance among PG&E's divisions and recommends that reliability measurement standards be reported at the division and operating area level as well as at the system level. ORA also recommends that the definitions used for Excludable Major Event and Major Outages be made consistent and that outage performance be measured against realistic benchmarks.

TURN maintains that the Commission cannot find PG&E's response to the December 2002 storm reasonable. In particular, TURN finds that PG&E's failure to adequately manage the outage information flows in its system, resulting in reported single outages being "aged off" without a record of the outage, demonstrates performance that was not reasonable. TURN points to the fact that PG&E has discontinued its use of the FAS during major events and has decided to modify its OIS as indications that PG&E acknowledges that its performance was unreasonable. TURN points out that not only is there no finding of reasonableness in ORA's testimony, or the PG&E/ORA joint testimony, but PG&E specifically acknowledges that customers' complaints regarding outage communications during storms led to various PG&E proposals to modify and improve such outage communications in the future.

CUE argues that PG&E's service is far less reliable than Southern California Edison's (SCE's) or San Diego Gas and Electric Company's (SDG&E's). According to CUE, outages on the PG&E system are more frequent and of longer duration than outages of the other California utilities. Of particular concern to CUE is that PG&E's average restoration time, as measured by CAIDI, has increased as PG&E field staffing levels have decreased. CUE believes that the Commission should adopt a reliability performance incentive mechanism to improve PG&E's performance.