The parties participating in the workshop generally agree with the UDC definition that load profiling is the process of taking the cumulative kilowatt-hour (kWh) usage of a customer over a billing cycle and assigning it to individual hours in the cycle, based on the aggregate characteristics of the customer segment in which the customer resides. The Workshop Report notes that several parties would clarify this definition by stating that statistically valid methods would be used in the assignment process. The Environmental Defense Fund would base the assignment of usage on an estimate of the time pattern of the customer's proportionate energy usage, rather than the aggregate characteristics of the customer segment.
The UDCs' load profile design involves the use of daily load shapes by hour, which the UDCs will make available to all applicable parties. The daily load shapes will consist of hourly kW loads for each applicable load profiling category, or equivalent normalized values. The UDCs propose to use various load profiling methodologies to develop daily load shapes from load research data that is currently available or can be readily developed. The specific methods that they plan to use will vary by utility and by rate category within each utility.
The methods that the UDCs plan to use are dynamic load profiles, static load profiles, and deemed load profiles. Dynamic load profiles are created by reading the load research meters on a daily basis, and producing daily load shapes which reflect the actual usage for that customer segment for the day. Edison proposes to use dynamic load profiles for its Domestic and GS-1 (small commercial and industrial) classes.
Static load profiles are created by averaging historical data from load research samples by class, and creating load shapes that approximate customer segment usage for the given day. Static load shapes are fixed in advance. They can be differentiated by season, month, and day. PG&E and SDG&E propose to use static load profiles for the majority of their rate categories. PG&E's load profiles use three years' worth of data to create average load profiles that smooth out unusual operating characteristics for any particular year. SDG&E and Edison propose to use similar methods to develop their static load profiles.
Deemed load profiles are created by using engineering estimates to create daily load shapes. This method is used for those rate schedules with predictable loads, such as street lights and traffic control devices, where metering may not be available.
The UDCs agreed to provide the load profile information for all applicable load profile categories to the appropriate market participants. This information was made available by PG&E, SDG&E, and Edison on August1, 1997 by electronic distribution to interested parties, and posted on ORA's web site. PG&E, SDG&E, and Edison request that the Commission approve the load profiles for interim use starting on January1, 1998.
The following are the listings of the rate schedules PG&E, SDG&E, and Edison propose be eligible for load profiling, and the type of load profile method that they propose to use.
PG&E4
Class |
Eligibility |
Estimation Method |
Residential |
E-1, E-7, E-8
|
Static |
Small Commercial |
A-1, A-6
|
Static |
Medium Commercial |
A-10, E-19V
|
Static |
Agriculture |
<35 hp connected load
|
Static |
Traffic Control Devices |
TC-1 |
Deemed |
PG&E-Owned Streetlights |
LS-1 |
Deemed |
SDG&E5
Class |
Eligibility |
Estimation Method |
Domestic |
All residential rate schedules |
Static |
Small Commercial: |
Schedules A, A-TC |
Static |
Large Com/Ind. |
Schedules A-TOU, AL-TOU, AO-TOU, and AY-TOU |
Static |
Agricultural |
Schedules PA, PA-TOU |
Static |
Lighting |
Schedules LS-1, LS-2, LS-3, OL-1 |
Deemed |
Edison6
Class |
Eligibility |
Estimation Method |
Domestic Domestic (master metered) |
D (includes all schedules except DMS) DMS |
Dynamic Static |
Small Commercial
|
GS-1 |
Dynamic |
Small Agriculture
|
PA-17 |
Static |
Traffic Control Devices |
TC-1 |
Deemed |
Edison-Owned Streetlights |
LS-1 |
Deemed |
The UDCs propose to periodically redraw their load research samples and redeploy their sample meters in response to customer turnover and customer migration to hourly meters.
The UDCs' proposal also calls for the continuation of stratification based on current practices for the interim period. Stratification is a process in which characteristics of a population, such as average usage or territory, are used to enhance the information contained within the sample for the entire class so as to increase the sample's accuracy.
The Workshop Report notes that one of the concerns with using static load profiles is that static profiles do not follow actual weather patterns. As a result, static profiles are subject to greater daily errors than other load profiling methods. Some of the participants have suggested that approved econometric techniques be used to estimate load profiles and to verify the fit of load profiles.
The UDCs agree that econometric techniques reflecting weather and other relevant variables would be an improvement, but suggest that dynamic load profiling is preferable in the long term. The UDCs also point out that the process of gaining approval of econometrically derived load profiles may be slow, and may not permit implementation on January1, 1998.
The UDCs are not opposed to moving towards dynamic load profiles, although the Workshop Report states that for PG&E and SDG&E, the use of dynamic load profiles is not feasible by January1, 1998. The UDCs contend that there are additional costs associated with dynamic load profiles, but some of the parties oppose allowing the UDCs to recover these costs through the use of Public Utilities Code Section 376.8
The Workshop Report states that some of the parties believe that any retailer should be able to design its own load profiles so that the profiles reflect the characteristics of the retailer's customers. This could mean that customers in the same market segment could choose between retailers offering different load profiles.
The UDCs contend that there are three problems with allowing retailer-developed load profiles. The first problem is the "churning" effect that is created when customers switch between the UDC and the retailer, or between retailers. That is, if retailers are permitted to develop aggregated load profiles that are specific to their own customer mix, customer migration between retailers will result in a systematic loss of accuracy in the load profiles, even with a significant increase in the frequency in which load research samples are redrawn. The second problem is that there are no existing procedures to verify the accuracy of a nonstandard load profile. This could result in gaming of the load profile and fraud, which could lead to cost shifting between the market participants. For example, ESPs might have the incentive to create flatter load profiles than may actually exist in order to offer lower prices to consumers. The third problem is that standards would need to be developed for retailer-developed load profiles. These standards, such as the sample selection process and the frequency of redrawing samples, would need to be prescribed by an independent regulatory entity.
The comments to the Workshop Report raise several issues with regard to the methodologies used for load profiles. These issues are discussed below.
Applied Econometrics, Inc. (AEI) points out that dynamic load profiles have the advantage of measuring current conditions that affect load, rather than employing profiles from previous years. AEI does not believe that static load profiles should be used for cost settlements because they do not account for current weather patterns, changes in consumption, or any changes in loads resulting from competition. However, Edison's dynamic load profile approach still has drawbacks because it does not provide for any segmentation of load profiles within the traditional rate classes.
Calpine Corporation, Cinergy Services, Inc., Mock Energy Services, NorAm Energy Management, Inc., and Stone & Webster Management Consultants, Inc. (jointly referred to as "Calpine et al.") agree with Edison's use of dynamic load profiles, but question why PG&E and SDG&E cannot do the same. Calpine et al. contend that dynamic load profiles are inherently more accurate than static profiles because the dynamic profile represents actual daily usage within the customer class. Calpine et al. believe that the UDCs should be required to implement dynamic profiling methods for all residential and small commercial and industrial customers by January1, 1998. Calpine et al. contend that it is inequitable for commercial or industrial customers to see markedly different electricity prices due to different load profiling practices of the various utilities.
Enron contends that the release of historic use information and load profile information for all applicable load profile categories is important. Although Edison plans to use dynamic load profiles, it should not be exempted from releasing the historical use data for current rate categories.
The California Energy Commission (CEC) believes that static load profiles should only be used as an interim measure pending the development and implementation of a dynamic load profiling methodology. The drawback with static load profiles is that the information may not capture the weather and operating conditions that are unique to the settlement period. The CEC recommends that PG&E and SDG&E consider how static load profiling can be refined to make it more accurate for use in the interim, while working to develop and implement dynamic load profiling. The CEC suggests that this can be accomplished by weather sensitive regression adjustments to the static load profiles; matching the load profiles to the overall system shape; and collection of load research data on a monthly basis, and producing time period specific profiles prior to final settlements.
The CEC recommends that a schedule be adopted to develop a dynamic load profiling methodology for all load profile customers by January1, 1999. The CEC points out that dynamic load profiles may be developed when there are a statistically meaningful number of interval-metered sites equipped with telemetry, which allows the data to be uploaded and processed daily. The load profiles can then be used as a source of information on which to base load forecasts, make bids into the PX, schedule power with the ISO, and determine settlements for energy imbalances. Since large financial commitments are riding on the PX load bids and the ISO load schedules, the CEC believes that accurate load forecasts are essential.
Itron, Inc. (Itron) agrees that refining load profile methodologies will benefit load forecasting and settlement applications. However, refining load profiles to use them to bill customers is inappropriate. Instead of refining load profiles for billing, Itron contends that the Commission should focus on the establishment of direct measurement of meters on a statewide basis using radio-based automated meter reading systems.
ORA asserts that a load profile should develop the average hourly pattern of consumption for a customer or group of customers, with the intent of providing the best possible representation of hourly loads in the absence of hourly metering. ORA endorses Edison's use of dynamic load profiles. However, ORA believes that Edison must still make available static load shape data for all customer classes in order for the ESPs to formulate workable business plans.
ORA believes that the dynamic load profiling process should consist of the following steps: (1) identify groups of customers having relatively similar loads by using monthly energy consumption data, location, climate zone, housing stock information, demographics, etc.; (2) for each customer in the sample, install an hourly interval meter and communication systems; and (3) use a sample weighting system to convert the actual load profiles of the sampled customers of a group into an estimated load profile for the average member of the group.
PG&E proposes to use the existing rate categories as the base for establishing load profiles for residential, small and medium commercial, agricultural, and street lighting customers. PG&E uses existing sample meters to derive the load shapes for these rate categories.
Edison states that dynamic load profiles are superior to static load profiles or to an econometrically adjusted static load profile, and that dynamic load profiles should be used where reasonably feasible. Where there is no dynamic metering capability available, Edison advocates using static load profiles. Edison states that it is able to offer dynamic load profiles for domestic and small commercial customers because Edison's load research group has been able to deploy additional hourly meters in recent years. This enables Edison to provide all the required information it needs to create accurate daily load shapes.
To maintain the accuracy of the load profiles, Edison plans to routinely update the samples. This update would be based on existing procedures, which typically involve the redrawing of samples about every four years. The samples are redrawn to reflect the changes in customer composition, migration between rate schedules, and other factors. Edison points out that in the future, more frequent sample updates may be needed to take into account those customers who install hourly meters.
The Commission recognized in D.97-05-040 that there were differences of opinion regarding how load profiles should be designed. The Commission stated:
"Statistical load profiles are estimates of the loads of a group of customers. Those estimates can be fairly accurate if they are based on an appropriate statistical sampling of customers, actual interval metering of some representative portion of the members of the group, and updated frequently. They can be less accurate if they are based solely on literature studies and are not adjusted to reflect actual usage from members of the group in question."
The UDCs propose that the Commission adopt the interim load profiling approach described earlier, and explore possible revisions to such an approach in 1998. The issue before us today is whether the UDCs' interim approach should be adopted, or should modifications be made to the approach prior to the use of such profiles. We first address the issue of whether static or dynamic load profiles should be used.
At present, only Edison plans to offer dynamic load profiles for its domestic and small commercial and industrial customers. According to the Workshop Report, PG&E and SDG&E are unable to offer dynamic load profiles prior to January1, 1998. The advantage of dynamic load profiling is that it reflects fairly current conditions which affect customer load. Conditions which affect customer load are such things as the weather and changes in the consumption patterns of the end-use customers. Load profiles which are based on a static approach that uses past history as a reflection of future consumption may not accurately reflect current consumption because of recent changes in the weather or operating conditions.
Some of the parties recommend improving the static load profiles by using econometric techniques to reflect changes in weather and other variables. They propose that these adjustments be made to the static load profiles to make them more accurate during the interim period. We recognize that the interim use of static profiles may not be the best solution for reflecting an accurate approximation of use by a particular customer class. The regression techniques, and other methods to improve the static profiles, will take some time to implement. Instead of adopting a requirement that the UDCs fine-tune their static load profiles, we expect PG&E, SDG&E, and Edison to move toward the use of dynamic load profiles for all eligible customer classes no later than July1, 1998, as discussed below. Thus, instead of working toward having more refined load profiles in place from January1, 1998 to June30, 1998, a period of six months, we expect PG&E and SDG&E to have in place dynamic load profiles no later than July1, 1998. Having dynamic load profiles in place for the three major electric utilities in this state makes sense because it will result in uniform load profiling methodologies throughout most of the state, and will result in improved accuracy of the load profiles.
We will therefore adopt the UDCs' proposal to use static load profiles on an interim basis for the majority of the customer classes, and Edison's use of dynamic load profiles for its residential and small commercial and industrial customers.9 Edison plans to use a static load profile for its master metered customers who serve residential customers. We indicated in D.97-05-040 that we would address the issue of master meters and direct access in an upcoming decision. Until the Commission resolves the issue of how direct access affects master metered customers and sub-metered customers, such a load profile should not be offered. Accordingly, we will defer the issue of load profiles for master metered customers to another decision.
The next issue is what customer classes should have dynamic load profiles, and when should dynamic load profiles be put in place for PG&E and SDG&E, and for Edison's other rate classes. The use of dynamic load profiling appears appropriate for most customer classes. One notable exception is for street lights and traffic lights. As discussed below, the loads of street lights and traffic lights are fairly consistent from day to day and year to year. Another possible exception is agricultural customers. Due to their varying sizes, their locations, and changing usage, the use of dynamic load profiles for these customers may not be appropriate or feasible.
With respect to the use of dynamic load profiles for agricultural customers, we will direct the Energy Division to convene a workshop with the UDCs and other interested parties, particularly agricultural interests, to discuss this issue. Such a workshop shall be held no later than February 15, 1998. The Energy Division shall then prepare and file a workshop report by March13, 1998 with its recommendations. Interested parties may file comments to the workshop report no later than March 31, 1998. The Commission will then determine whether dynamic load profiles for agricultural customers should be mandated.
The Workshop Report and the filed comments did not shed any light on why it is not feasible for PG&E and SDG&E to institute dynamic load profiling before January 1, 1998. We assume that part of the problem is that they may not have enough load research meters in place to create dynamic load profiles. PG&E and SDG&E should take action immediately, as should Edison, to ensure that dynamic load profiling is in place for all the customer classes described above no later than July1, 1998. We will direct PG&E and SDG&E to file with the Commission's Docket Office, within 30 days a status report as to the reasons dynamic load profiling cannot be instituted for residential and small commercial and industrial customers by January1, 1998. The status report shall also state what steps they will take to ensure that dynamic load profiling can be in place beginning on July1, 1998. Edison shall also be required to file a status report on what steps it needs to take to ensure that dynamic load profiling can be instituted for the other customer classes described above no later than July1, 1998. The status report shall also contain an estimate of the costs that the UDCs anticipate spending to institute load profiles. Interested persons will be provided with an opportunity to comment on these status reports.
The UDCs are free to implement dynamic load profiling before July1, 1998. Should that occur, to effect a smooth transition from the use of static to dynamic profiles, the UDCs shall provide the ESPs with 45 days advance notice before such profiles are used.
The four smaller investor-owned electrical corporations, Kirkwood Gas and Electric (Kirkwood), PacifiCorp, Sierra Pacific Power Company (Sierra Pacific), and Southern California Water Company (SCWC), have not submitted any load profiles or commented on the Workshop Report or the supplements. The four smaller utilities have filed applications addressing the requirements of Assembly Bill (AB) 1890 (Stats. 1996, ch. 854.). Unless those utilities are exempted from having to provide direct access to its customers, they remain obligated under Ordering Paragraph 5 of D.97-05-040 to provide load profiles to its customers with loads of less than 20 kW. These four utilities shall be required to file status reports with the Commission's Docket Office within 30 days from today. The status report shall explain what their intentions are with respect to offering load profiles to their customers who have a maximum demand of less than 20 kW, their reasoning for their actions, and what their timetable is for implementing load profiles. Interested persons will be provided with an opportunity to comment on the status reports of the four smaller utilities.
Some of the commenting parties have suggested requiring Edison to make available historical use data for current rate categories with a maximum demand of below 20kW, even though Edison plans to use dynamic load profiles for its residential and small commercial and industrial customers. These parties assert such information is important for the ESPs' marketing efforts. We see some value in making this kind of information available. However, given the number of days remaining in 1997, we believe that it is a more productive use of everyone's time to focus on making dynamic load profiles available for the remaining customer classes. Once Edison begins to offer dynamic load profiles, the ESPs will be apprised of the electricity usage by residential and small commercial and industrial customers. In order to provide the ESPs with some idea of the usage data for Edison's residential and small commercial and industrial customers, we will require Edison to make available its dynamic load profiles for these rate schedules starting November1, 1997.
Itron contends that for billing purposes, the Commission should adopt direct measurement of meters using radio-based automated meter reading systems. We decline to adopt Itron's suggestion. Although metering standards and interoperability standards need to be in place, the technology to carry out the metering functions is better left to the marketplace, rather than to this Commission, to decide. (See D.97-05-040, p. 35, fn. 14.)
Entities other than UDCs should not be allowed to create and design load profiles at this time. First of all, if entities other than the UDCs are permitted to design their own load profiles, a process to evaluate the accuracy of the load profile would be needed. Given the short time remaining before direct access is to begin, we believe that the Commission's resources and time can be spent more effectively addressing other direct access implementation issues. A second reason for not permitting other entities to design their own load profiles is that there will be an opportunity to revisit this issue as discussed in the segmentation section below. A workshop will be held to discuss, among other things, whether other entities should be permitted to develop segmented load profiles.
The California City-County Street Light Association (CAL-SLA) contend that all street lights should be load profiled using the deemed estimation method. CAL-SLA points out that the hours of operation for street lighting is consistent, i.e., street lights come on at dusk, stay on during the night, and go off at dawn. A photocell acts as the switching unit which turns the lights on and off.
CAL-SLA states that according to the Workshop Report, PG&E and Edison propose that the deemed estimation method be used for street lights that the UDCs own, but not for customer-owned street lights. The rationale of PG&E and Edison for not permitting customer-owned street lights to be on load profiles is that the municipal owner is not required to meet any maintenance requirements for the photocell which controls the lighting. 10 SDG&E, on the other hand, plans to permit customer-owned street lights (LS-2) to be on a load profile.
CAL-SLA point out that for customer-owned street lights on Edison's LS-2 schedule, Edison furnishes and operates the switch to operate the street light. Under PG&E's LS-2B or LS-2C schedules, PG&E maintains the photocells. If a load profile is not permitted, these customer-owned street lights would not be able to use direct access because it would require the installation of a costly interval meter that would exceed the cost of the street light. CAL-SLA also contend that there is no evidence that indicates that there are maintenance problems with customer-owned street lights. ORA supports CAL-SLA's interpretation, and argues that load profiling for LS-2 customers should be permitted using the same profile as the LS-1 customers.
The California Department of General Services, the University of California, and the California State University (DGS/UC/CSU) state that there was not agreement among the stakeholders that load profiles should apply only to utility-owned street lights. Most stakeholders appeared to be supportive of the concern that customer-owned street lights should be treated no differently than utility-owned street lights for purposes of eligibility for load profiles.
With respect to the traffic control signals, CAL-SLA points out an inconsistency in the approach adopted by PG&E and SDG&E. PG&E and Edison propose to use the deemed estimation method for traffic control devices on the TC-1 rate schedule. SDG&E characterizes traffic control signals as small commercial. CAL-SLA recommends that SDG&E use a deemed load profile for SDG&E's A-TC tariff.
Edison proposes to use deemed load profiles for lighting and traffic control services. The deemed load profiles are based on engineering estimates of hourly usage for customers on those schedules. However, if the customer owns the photocells or other control devices for these kinds of services, Edison may require auditing of customers' maintenance practices, and may restrict the use of load profiles if maintenance does not meet appropriate standards.
Whether deemed load profiles should be used for all street light rate schedules revolves around the issue of who controls the maintenance of the photocell. Edison, and formerly PG&E, asserts that unless there is regular scheduled maintenance of the photocell switching device, there is no assurance that the load profiles of street lights will apply because of possible malfunctioning photocells. That is, if the photocell is not operating properly, the usage of the street light may exceed the usage estimated by the load profile.
SDG&E proposes to apply load profiles to all of its street light accounts. SDG&E appears to be unconcerned about who is responsible for the maintenance of the photocell. In its comments to the Eligibility Supplement, PG&E revised its position to allow all street light and traffic control accounts, regardless of lamp and photocell ownership, to go on deemed load profiles. (PG&E Comments to Eligibility Supplement, p. 4.) Edison proposes to restrict the use of load profiles if the customer-owned or -maintained street light does not meet appropriate maintenance standards.
We have carefully weighed the arguments regarding the use of deemed load profiles for all street lighting accounts. Edison's position has merit in that one needs to ensure that the photocell switching device is working properly. However, customer-owned street lights should not be barred from load profiling entirely simply because of the ownership of the photocell or who maintains it. We will require that the UDCs make all street lights, regardless of ownership, eligible for direct access based on deemed load profiles. We will also allow the UDCs to impose in such tariffs and profiles that the customer must meet or exceed the photocell maintenance requirements of the UDC in order to be eligible for direct access using that load profile. Such a requirement helps to ensure that the street light usage is consistent with the estimate of usage in the deemed load profiles.
CAL-SLA points out that SDG&E's load profile for traffic control lights, A-TC, uses a static load profile instead of a deemed load profile. The traffic control rate schedules of PG&E and Edison both use deemed load profiles. In deciding whether we should require SDG&E to change its load profile method for its A-TC schedule, we need to consider whether the static profile would result in a more accurate usage profile than a profile based on a deemed profile. As with street lights, one would expect the electricity usage of traffic lights to be consistent. Usage estimates could either come from historical data or engineering estimates of usage. Since electricity usage for traffic lights is fairly consistent, SDG&E's use of a static load profile is unlikely to vary much from a deemed load profile. Therefore, we will permit SDG&E to use a static load profile for its traffic lights rate schedule.
4 For PG&E customers who have a demand metered account, PG&E proposes to screen annually for accounts with maximum demands of less than 20 kW in nine of the last twelve billing cycles.
5 SDG&E proposes to screen annually for accounts with maximum demand of less than 20 kW, or if no demand meter is present, with usage of less than 12,000 kWh in each of the previous twelve billing cycles.
6 Since dynamic load shapes will be used for Domestic and GS-1 customers, advance load shapes will not be available.
7 Edison points out that its PA-1 load profile does not conform with the 20 kW load profiling threshold established in D.97-05-040. About 20% of the 5000 accounts have connected loads above 20 kW. Due to the small number of customers, and because Edison's load research sample reflects the usage of all PA-1 customers, Edison proposes that these 5000 accounts be eligible for load profiling.
8 Unless otherwise noted, all section references are to the Public Utilities Code.
9 The use of deemed load profiles is discussed below.
10 PG&E subsequently changed its position as discussed below.
This Web page was produced using a Beta version of HTML Transit 3.0