Divestiture Modeling Methodologies and Assumptions

To support the economic and operational characterization of the operational and emission impacts of the proposed divestiture, the proprietary computer models SERASYMÔ and the Surplus Energy Resource Assessment Model (SERAM IIÔ ) (1) were used to simulate the future operations of the California electric system and its interactions with the rest of the Western Systems Coordinating Council (2) after restructuring with and without the occurrence of divestiture of the four power plants as proposed by PG&E.

Two "no-project" cases postulating that the proposed divestiture project does not occur were modeled. The no-project case run for 1999 is the CEQA baseline case. Both the 1999 and 2005 no-project modeling assume the persistence of current utility generation ownership beyond the first phase of divestiture by PG&E and Southern California Edison (Edison). For consideration of divestiture effects and those of the numerous alternatives identified, these cases were used as starting points to produce further scenarios which incorporated additional assumptions and modeling stratagems to characterize the various alternatives and their operational and emissions impacts.

2.1 Baseline

The 1999 baseline assumes that neither further divestiture nor market power manifestations occur. It provides a CEQA "no project" baseline which is to be used as the basis of comparison with all other cases and scenarios.

A myriad of assumptions and modeling methodologies go into every model forecast of electric system operations. For the baseline case, the same set of methods and assumptions were used, except as described below, as those employed in the baseline forecast reported in the Initial Study/Mitigated Negative Declaration for PG&E’s first power plant auction in 1997 (Phase 1). (3) The assumptions from the first sale of particular importance to the proposed sale (Phase 2) include the following:

As the baseline case is intended to simulate present or near-present operations, the year 1999 was chosen for simulation. New and updated assumptions for the baseline case and the entire set of modeled scenarios where appropriate include the following:

New requirements and questions became important in the process of performing the modeling for this phase of the divestiture. This necessitated applications of different and/or extended methodologies:

The tabulated results for the baseline are presented in Table G-1, which exemplifies the standard presentation form used in this Attachment to present modeling results. Unit and plant specific results are presented for each of the fossil and geothermal units proposed for divestiture. Results are also presented for the remainder of the generation employed to serve California ISO load. The table shows, for instance, that the unit with the highest expected capacity factor (given normal hydro conditions) of the fossil units proposed for divestiture is Potrero 3. (9) At a capacity factor of 41 percent, Potrero 3 is operated mostly in support of the SFOC as established in a separate analysis (see Section 3.1 below).

The row second from the bottom ("Non-BAAQMD California Load-Related") describes the operations of the remaining plants. For comparison, it indicates that all the other sources of generation used to satisfy ISO load in California emitted over 200 thousand tons of NOx at a rate of about 1.78 lb per MWh. (11) The bottom row incorporates the results for the plants proposed for divestiture, presenting overall results for the state.

2.2 The 2005 No-Project Case

The 2005 no-project case depicts post-1999 cumulative impacts upon PG&E-built Bay Area electric generation of electric industry restructuring, changing air quality regulatory requirements, load growth in San Francisco and the remainder of California, resource additions to meet growth requirements, and resource retirements according to current and pending regulatory agreements. The year 2005 was chosen to ensure capture of all effects of the existing air quality regulations yet be sufficiently within the restructured period. The BAAQMD air quality rules currently call for increasingly stringent emission limits in the region over the next few years, reaching a steady state before 2005.

Tabulated results for the 2005 no-project case are presented in Table G-2. Probably the most interesting results from this case are the very high capacity factor seen for the new 480-MW station in San Francisco, as befits such an efficient plant, and the reduction in generation from Potrero Unit 3 due to the presence of the new 480-MW station and its individual capability, except under very high San Francisco loads, to meet the SFOC. It is also obvious from reviewing the results to see that Pittsburg Unit 1, absent major emission controls, is the least "clean" of any of the fossil-fueled boiler units analyzed.

2.3 Analytical Maximum Generation

The analytical maximum generation cases are intended to study the potential impacts of selling the plants to private parties with incentives greater than those of PG&E to maximize generation sales from the plants they own. (16) Reasons a new owner might have incentive to run its units more than PG&E are comprehensively discussed in Attachment C and include, among others, having a smaller generation portfolio and thus relying more on this unit’s income; reduced operating costs from such actions as staff cuts, new equipment, refurbishments, or acquiring cheaper fuel supplies; and signing direct access contracts.

In order to investigate these potential effects, both the baseline and cumulative impact cases were used as starting points for a series of sensitivity scenarios to test the environmental impacts of running these plants at their maximum credible levels of generation consistent with the market into which they sell. This market features some unavoidable limits that absolutely preclude all of these plants from simultaneously running full out. These constraints include the availability of the individual units; transmission limits; limited hourly demand net of ISO obligated must-run and must-take generation; existing take-or-pay contracts with out-of-state utilities; and competition from negligible or very low variable cost of generation sources including California hydro, lower cost coal generation and imports. In addition, there is sufficient generation being offered under this phase of divestiture that postulated generation from one unit can be displaced during some hours of lower load by postulated generation from another of the units being offered for sale.

To implement these scenarios we chose to reduce their fuel costs in order to preserve relative economic dispatch order between the units. In the "analytical maximum" cases for the fossil plants, the natural gas prices seen by the natural gas-fired boiler units were replaced with the cheapest natural gas supply utilized in the baseline and cumulative impact case modeling, that of the Coolwater plant in southern California, reduced by a further twenty-five percent and provided in unlimited quantities. (17)

The Geysers geothermal plants supplied by Calpine wells (Nos. 13 and 16) are already running at their steam-limited maximum levels; the remainder, supplied by UNT, are not. Geothermal units have fewer options than natural gas units for reducing fuel costs due to the fuel’s very nature as a product of local geothermal heating of the water table, but UNT could nevertheless run the units baseloaded to employ all available steam if so inclined. Consequently, for the geothermal case we reduced the price of the geothermal steam for the UNT plants to the level provided by Calpine for Geysers Plants 13 and 16.

This "analytical maximum" concept is not applicable to peaking units such as the combustion turbines to be divested, although it does apply and was employed with the new 480 MW combined-cycle (CC) resource addition in the 2005 no-project case. A CC can be a highly cost effective baseload or intermediate resource, while CTs are unavoidably high cost, normally reserved for transmission disturbances or generator outages, and cannot operate more than ten percent of the hours in a year, in any event, due to BAAQMD rules.

To take account of the mutual displacement effect among the plants being operated at their analytical maximum levels, we chose to run each of the three natural gas plant packages (18) and the geothermal plants in separate simulations at their maximum levels. A case featuring simultaneous analytical maximum operations of all the natural gas plants was also run to investigate their joint maximum in case a single owner were to purchase all of them, as happened in the first phase of PG&E divestiture.

Table G-3 represents the analytical maximum of the baseline for the Contra Costa and Pittsburg units, which are assumed to be sold together to a single buyer. The 1999 analytical maximum change in generation from these plants is quite large. Capacity factors at both plants doubled in the 1999 case and total generation rose by 9,026 GWh, equivalent to approximately 3.5 percent of the entire California load. Pollutants emitted by the plants increased apace, though the resulting average NOx emission rate actually declined due to these units being cleaner than the average for the Bubble. There was some reduction in San Francisco boiler generation but it was very modest (only 29 GWh).

No separate analytical maximum case was run for the Delta units for 2005; see Section 2.3.2.3 and Table G-6.

Table G-4 is the analytical maximum version of the baseline case for Potrero 3, the only steam boiler unit at the Potrero Power Plant. In 1999 much of Potrero’s analytical maximum increase in generation displaces generation from the Hunters Point Plant due to Potrero satisfying the SFOC in lieu of the Hunters Point units, whereas the increase in Potrero 3 generation has essentially no impact on CT generation at Potrero or at Hunters Point 1. This is to be expected since the CTs are used almost exclusively to satisfy reliability needs which do not vary with level of operation of boiler units, all of which are found relative early in the commitment order.

Table G-5 shows the 1999 baseline with all the natural gas-fired steam boilers proposed for divestiture and at the Hunters Point Plant run at analytical maximum. This run inspects the degree of joint backout attributable to the simultaneous running of all the area natural gas units at these maximal levels. Each natural gas boiler increases its capacity factor significantly compared to the 1999 baseline (Table G-1); however, the level of operation is reduced slightly at each plant in the joint maximum case from that found in the individual maximum cases (Tables G-3 and G-4). Interestingly, in spite of the reduced generation at each plant, individual units may operate more in the joint maximum case. For example, Hunters Point 2 and 3 increase their total generation while generation from Hunters Point 4 drops by a greater total amount, and Potrero 3, also in the City, drops the most of any unit. Overall, these differences are fairly minor.

Table G-6 shows the analytical maximum version of the 2005 no-project case for all area natural gas boilers. No individual cases were run for 2005 for the Potrero or Delta analytical maximum because the reductions in generation, minor in 1999, were expected to be even less significant in 2005 since the overall greater capacity factors were anticipated in 2005 to meet increased load requirements. Additionally, the change in overall system resources and particularly the retirement of Pittsburg 3 and 4 change the overall commitment and reduces the total amount of generation from the Delta units. Note that the analytical maximum change in generation from the new 480-MW station was minimal in any event, since it was already sufficiently cheap and would reduce generation at the other fossil-fueled plants being divested.

Table G-7 shows the analytical maximum baseline for operations of the Geysers at maximum steam usage. Here, generation increased for all the UNT plants, while it remained unchanged at maximum output for Geysers 13 and 16, the two Calpine-supplied units with the very low priced steam.

Table G-8 shows the 2005 no-project case with the Geysers at analytical maximum. Interestingly, the increase in generation in 1999 is much larger than in 2005 because, by then, the UNT steam price has escalated less than the price of natural gas so that most of the steam available is used in the no-project case.

2.4 Sensitivity Scenarios

Numerous alternative possibilities for the future, of varying likelihood, may be identified, characterized, and investigated as to potential impacts. Besides the range of reasonably conceivable impacts of divestiture, these potentials include sales of varying facilities to parties with varying operating incentives and sales and operating policies, resource changes in the local areas of the plants, revisions to operating criteria, revisions to fuels used by units, and levels of air quality regulation.

For analyzing the sensitivity of the system to these alternatives, one or both of the cases above were modified to reflect each of the sets of possibilities and the differential impacts analyzed.

This scenario studied a variant to the 2005 no-project arrangement case with less reliance on new in-city generation while still meeting emission limits and the SFOC, through a combination of increasing transmission import capability into the City and further reducing the level of in-city load covered by in-city generation for the SFOC. No consideration is given to costs or to determining cost effective actions. All that are considered are the technical changes necessary to support the scenario and the time to implement the change. Further, these actions will tend to reduce overall reliability in the City to the degree additional elements of the network load are not sustainable from indigenous generation.

To eliminate some of this new generation while maintaining the SFOC, even the relaxed SFOC following the transmission improvements prior to 1999, at 2005 load levels does not appear feasible; additional transmission into the city, with concurrent further relaxation of the SFOC, is required. The 2005 no-project case was run with all Bay Area natural gas-fired generation at analytical maximum, and two additional changes. First, further transmission enhancements to the San Francisco import corridor were assumed to add an additional 250 MW of import capability over the 2005 no-project case. These additional changes are necessary to account for future load growth in the City, replace the capacity from the other new 480-MW station (two new 240-MW units) no longer assumed, and keep the capacity factors of each of the CTs well below ten percent in spite of increased use for the modified SFOC and to help support the overall ISO system. Transmission upgrades might involve reconductoring the 1000 foot segment of the "overhead" 115 kV lines between San Mateo and Martin substations that is actually a subterranean cable and increasing the amount of transformation at the Martin Substation, and/or building a second, buried 230 kV cable from San Mateo to Martin. Other improvements north of Martin would also be required including possibly a 115 kV cable from Martin to Hunters Point and a buried cable from Hunters Point to Potrero. Second, this additional transmission allowed the SFOC’s native on-line generation requirements to be relaxed enough to allow one of the two new 240-MW units in the City previously assumed in the no-project case to be eliminated, leaving one new 240-MW unit equipped with SCR added in San Francisco.

Table G-9 presents the results of this scenario. The new 240-MW station is seen to run at the same high capacity factor as the full 480-MW station because it is assumed to have the same economics on a per MW basis. Potrero generation rises only slightly at the CTs. Generation also rises noticeably at the Delta Plants and even increases slightly for the geothermal plants, even under analytical maximum conditions. Given the assumptions regarding load growth and new and remaining resources available to serve system load, no problem is detected in satisfying the remaining lost generation from the postulated new 240-MW station though the variable cost of serving the load is significantly increased.

2.4.2 Suspended Natural Gas Fired Emission Control Technology

Over the next few years, BAAQMD air quality rules are set to require gradually lower emission rates for utility electric steam boiler generation. Also, load growth in San Francisco will cause PG&E to run both its steam and combustion units progressively more to meet load and reliability requirements, causing these units to approach current BAAQMD emission and operation limits they had previously remained comfortably below.

PG&E has contemplated installing numerous emission control technology measures between now and 2005 at its natural gas-fired generating units in order to maintain compliance with its BAAQMD regulations. (19) These regulations (specifically Regulation 9, Rule 11) would not, in their current form, apply to new non-utility owners. BAAQMD has stated its intent to modify Regulation 9, Rule 11 to ensure its continued applicability to all of the steam boilers at the four Bay Area power plants, regardless to whether they are utility owned.

To study the maximum possible emissions impacts of the sales from a regulatory perspective, these scenarios postulate for the no-project cases that the new owners (of the PG&E electric generating plants being divested) install none of the PG&E-planned NOx emission controls between now and 2005, consistent with BAAQMD not modifying Regulation 9, Rule 11 to apply to the new ownership.

The tabulated results for the cases with emission controls frozen at 1998 levels are presented in Tables  G-10 and G-11. In 1999, total NOx emissions from Pittsburg, for example, increase from 3,000 (see Table G-1) to 3,685 tons and the emission rate increases by 22 lb/BBtu. In 2005, Pittsburg emissions increase from (see Table G-2) 661 to 3,034 tons and the rate goes up by 74 lb/BBtu.

Tabulated results for the analytical maximum generation cases for 1999 and 2005 with emission controls frozen at 1998 levels are presented in Tables  G-12 and G-13. Here, in 1999 total Pittsburg NOx emissions increase from 3,000 tons in the 1999 baseline case (see Table G-1) to 7,444 tons, and the emission rate increases by 22 lb/BBtu. In 2005, at analytical maximum dispatch Pittsburg NOx emissions increase by 4,261 tons (see Table G-2) due to the freeze in emission controls ceteris paribus, and the emission rate goes up by 79 lb/BBtu.

2.4.3 Proposed ENRON Power Plant

In April 1998, ENRON Corporation announced that it plans to partner with the City of Pittsburg and a steel producer to build a natural gas-fired combined-cycle plant of about 500 MW in or near the city. The partners expect to file a siting application with the CEC and hope to begin operations by 2001. Part of the plant’s capacity would be reserved for an "over-the-fence" steel producer with an existing 50 MW boiler that would be retired.

A 2005 case was added with a new 450-MW combined-cycle (CC) natural gas-fired plant in Pittsburg, based on the new 480-MW station in San Francisco (described earlier) and with the same natural gas supply prices as adjacent PG&E plants. The new plant was incorporated into, and made subject to, the BARR, and the existing Pittsburg Plant Units 1 and 2 were retired as PG&E had previously planned, under the assumption that the local voltage support function they would otherwise have provided could be handled by ENRON. Lastly, all Bay Area steam units were dispatched at their analytical maximums.

The tabulated 2005 results for the case with the ENRON plant are presented in Table  G-14. The major impact of the presence of the ENRON facility is that generation from remaining boilers declines by 1,830 GWh and total NOx emissions from the boilers and CCs decline by 484 tons, while total Bay Area generation rises by 1,696 GWh. Along with increased total generation associated with the proposed new 450-MW CC comes increases in other pollutants: e.g., annual PM-10 emissions increase by 21 tons.

2.4.4 Extended Delta Plant Fish Entrainment Limits

The National Pollution Discharge Elimination System (NPDES) permits for the Contra Costa and Pittsburg plants, among other operational restrictions, require limits to cooling water intake to protect striped bass spawning in the Sacramento and San Joaquin River delta. (20) These rules, which restrict generation and capacities for the units in question, take effect about May 1 of each year and end on July 15 or when the California Department of Fish and Game advises that striped bass levels in the river warrant ending the restrictions.

The U.S. Fish and Wildlife Service and state authorities have been in negotiations with PG&E about applying similar but more extensive restrictions than these in order to protect endangered native species. (21) The new restrictions, which would protect endangered and threatened species from entrainment as well as temperature increases from cooling water, are tentatively planned to shift the beginning of the special operations season from May to February with the ending remaining in July.

The baseline case for 1999 was modified to extend the Delta plant fish entrainment limitation periods and associated operational restrictions to be in effect from the beginning of February to mid-July. These conditions, without predicted retirements or necessary concomitant transmission upgrades, were carried into 2005 for purposes of comparison. The tabulated results for the runs with extended Delta protect periods are presented in Tables  G-15 and G-16. The impact of this change appears very modest; the generation at Contra Costa declines slightly while the generation at Pittsburg increases more than enough to make up for the decline at Contra Costa. However, in all instances the increase in generation occurs primarily at Pittsburg 7, which has a cooling pond, and declines in generation occur at units with once-through cooling where the detriment to the fish population occurs.

2.4.5 Gas Conversion Of San Francisco Combustion Turbines

Natural gas-fueled combustion turbines are, for certain pollutants, cleaner burning than those burning residual or distillate petroleum products. The Potrero and Hunters Point CTs, which burn distillate, are the only commonly used utility electrical generation units being controlled by the ISO whose primary fuel is not natural gas. (22) Conversion of these units to natural gas burning would reduce NOx and other emissions in the local area, all other things being equal. However, by converting to natural gas all other things would not remain equal. This fuel conversion would have the effect of reducing each unit’s marginal operating costs due to expected continuing significant differences in fuel prices, and thus increase their frequency of usage.

This case studies the net operational benefits that might arise from fuel conversion of the CTs. No estimate was made of the cost effectiveness of such a procedure though PG&E has studied the possibility and it appears to be feasible from an engineering perspective.

The analytical maximum case for all Bay Area natural gas generation in 2005 (1999 not providing sufficient lead time for the conversion) was modified to convert Potrero Units 4 through 6 to burn natural gas, available at the typical northern California price. The tabulated results are presented in Table  G-17. The impact of this change was as expected in terms of operations: the CTs were called on much more often when converted to natural gas. In fact, their usage exceeded slightly the ten percent operational ceiling now imposed upon these units while they burn distillate. The key question is whether or not the pollutant emission declined with the conversion in spite of increased operation. The answer appears to be mixed, depending upon the pollutant. SOx emissions decline drastically while PM-10 remain basically unchanged. NOx, on the other hand, increases but is insignificant as compared to the overall emissions of the boiler units.

Several analyses were made of special possible circumstances that were not based on the no-project case because they either involved no actual modeling or a significant modification to the system modeling to employ a modeling "trick" or stratagem to isolate an effect.

3.1 San Francisco Stand-Alone with Alternate In-City Generation and Displacement Generation Requirements

The electricity generated by the Hunters Point and Potrero plants in San Francisco is normally used only to supply local loads and satisfy the SFOC. Since even at full output these plants cannot regularly support San Francisco’s entire load without resorting to several of the CTs as well as all the boilers, the city is usually a net importer of electricity over its transmission link with the rest of the PG&E system. However, occasional conditions exist under which the City units operate at greater than the level dictated by the SFOC, thereby reducing electricity imports into the City and helping to support overall, and not just in-city, PG&E loads. For example, at times of high summer peak loads, as available capacity outside San Francisco becomes scarcer and dearer, the generation in San Francisco, which is strategically located in the heart of the Bay Area Reliability Area, will be called upon to increase its level of generation to reduce imports into the City, freeing up other generation to serve overall system loads. (23)

To analyze the extent to which the generation in San Francisco operates to support only in-city load and system requirements versus out-of-city needs, the San Francisco system forming part of the total system was severed from the rest of the system and run by itself. Local loads were unchanged; the transmission link into and out of the City was connected to a firm source of generation which served only to supply San Francisco loads at inframarginal cost; i.e., a cost level sufficiently below those of any of the City’s generators so that it was always economic to completely fill the transmission line. The total generation from the San Francisco units under these conditions constitutes the minimum needed beyond that required to satisfy the SFOC and/or supplement the full import capability available. This amount of total in-city generation was then subtracted from the corresponding in-city generation levels found in the cases, and the difference taken to represent the amount of in-city generation that ran to serve purposes outside the City.

Because the City is such a small region and the individual generators are such large percentages of this load, we employed the Monte Carlo solution technique in SERASYMÔ to permit accurate results.

The tabulated results for 1999 for the San Francisco stand-alone cases are presented in Table  G-18, which differs in form from the tables presented heretofore. This table only reports generation from the in-city plants and reports both the stand-alone and full-system cases and the differences between them. Examination of these results show that very little San Francisco generation is used to displace non-city generation. It totals 122 GWh for an 8.0 percent increase over the minimum amount of generation needed to observe the SFOC, given the assumed level of import capability. The majority of the increase occurs in Hunters Point Unit 4 although the largest increase arises in the CTs, which are presumably needed to help serve system peak load conditions. Annual emission increases are also quite modest averaging just about ten per cent for each of the pollutants.

3.2 Pittsburg Power Plant Peak daily PM-10 Emissions

In each year of electrical system operations each plant has a peak day when the specific plant sees and must meet its highest level of generation. That day is unlikely to be the peak load day of the year due to the impacts of availability of other, cheaper sources of generation. Also, that day may or may not be the same day as when selected emissions also peak for that plant. Further, emissions of NOx which are driven by the control technology on a given set of units may not peak the same day that other pollutants such as ROG, CO or PM-10 peak, since those pollutants all peak on the day of maximum fuel burn which is not necessarily the day of maximum generation or NOx outputs.

Special analyses were undertaken to identify the hourly emissions from the Pittsburg Plant on the day of its maximum-modeled PM-10 emissions. For Pittsburg, the day of maximum-modeled PM-10 emissions is the second to the last Monday in September. The hourly tabulated results for the Pittsburg Power Plant for the no-project case in 1999 are presented in Table G-19. The hourly tabulated results for the analytical maximum case in 2005 are presented in Table G-20. Hourly generation and emissions are shown on the tables by unit. The results show that all of the units are committed at all times of the day and that the peak operations and peak PM-10 emissions occur from hours 14 and 17; i.e., 1:00 p.m. and 5:00 p.m. in 1999. Due to the high demand, all units are operating at their peak outputs during that afternoon except for Pittsburg 7, which has a very expensive incremental cost of operations at peak output due to its closed cycle cooling so it was not fully dispatched even though the remainders of the units were. For 2005, with a higher demand level throughout the system and Pittsburg 3 and 4 retired, the peak emissions occur in hours 8 and 11 through 22: that is, through most of the afternoon and evening.

From comparison of the results of the cases, the various scenarios, and the special analyses of alternatives, observations can be made that provide guidance in the divestiture EIR process. Observations evident from the results to date include: