V. Cost-Effectiveness Conclusion
In D.05-02-052, we adopted the following changes to PG&E's modeling assumptions to be used in our cost-effectiveness calculations:
· SGRP cost of $706 million (base case), and $815 million cap.
· Base capital additions of $87 million (2003 dollars) for 2016 and after.8
· 4.5% operations and maintenance expense escalation rate after 2011.
· September 5, 2003 and April 19, 2004, NYMEX closing prices for gas.9
· 30-year facility life for combined cycle generation.
The following table from D.05-02-052 shows the mean net present value of revenue requirements (NPV), in 2003 dollars, of five scenarios illustrating the results of our cost-effectiveness analysis.10 A negative NPV indicates that the costs of the SGRP exceed the benefits, which means that the SGRP is not cost-effective. Likewise, a positive or zero NPV indicates that the SGRP is cost-effective. The term "High Gas" refers to replacement electricity costs based on the September 5, 2003 NYMEX closing prices for gas. The term "Low Gas" refers to replacement electricity costs based on the April 19, 2004 NYMEX closing prices for gas.11 The base case (Scenario 1) used the above modeling assumptions and a $706 million SGRP cost. Subsequent scenarios incorporated additional assumptions. Each scenario is shown using the 90.6% capacity factor used by PG&E in its application, as well as an 85% and an 80% capacity factor to illustrate the effect of a lower capacity factor on the cost-effectiveness of the SGRP.
Table of NPV Results From D.05-02-052
Scenario |
Assumptions |
Capacity factor12 |
Low Gas |
High Gas |
1 |
Base |
90.6% |
522 |
804 |
2 |
Base +1 unit refueling outage13 |
90.6% |
429 |
687 |
3 |
Base + 1 unit refueling outage |
90.6% |
333 |
591 |
4 |
Base + 1 unit refueling outage |
90.6% |
194 |
439 |
5 |
Base +2 unit refueling outage |
90.6% |
217 |
450 |
PG&E performed steam generator tube inspections during the October-November 2004 refueling outage of Unit 2. When D.05-02-052 was written, the tube degradation results of the tube inspections for Unit 2 were not available, and the tube degradation results of the tube inspections during the Unit 1 refueling outage in early 2004 were not in the record. Scenarios 3, 4, and 5 were based on possible results of these inspections. We determined that Scenario 3 was the most probable outcome of those inspections, and found the SGRP cost-effective.
On March 17, 2005, the assigned Administrative Law Judge (ALJ) issued a ruling requiring PG&E to provide cost-effectiveness calculations using the tube degradation results of the tube inspections performed during the last refueling outages for both units. These calculations were to be identical to the calculations performed in connection with D.05-02-052, the results of which are shown in the above table, except that the tube degradation results of the last refueling outages for each unit would be substituted for the possible results used in Scenarios 3, 4, and 5. The ruling stated the ALJ's intention to make PG&E's response to the ruling an exhibit. On April 3, 2005, PG&E filed its response to the ruling. Since no party filed an objection to PG&E's response, it was marked for identification and received into evidence. The tube degradation results indicate that Unit 1 will run for one additional refueling cycle, and Unit 2 will run for one half of an additional refueling cycle (relative to the Scenario 1 assumptions) if the SGRP is not performed. The cost-effectiveness calculations incorporating these results are shown in the following table in 2003 dollars:
Table of NPV Results
Scenario |
Assumptions |
Capacity factor |
Low Gas |
High Gas |
6 |
Base (updated) |
90.6% |
422 |
678 |
7 |
Base (updated) |
90.6% |
226 |
458 |
Scenarios 6 and 7 replace Scenarios 3, 4, and 5 because they are based on actual rather than possible tube degradation results of the last inspections for each unit.16
As indicated in D.05-02-052, we have no reason to believe that a one-year outage of one unit is likely. Therefore, we believe Scenario 6 is the most probable. Under this scenario, the SGRP will be cost-effective, even at the low gas price, the $815 million SGRP cost, and an 80% capacity factor.
Scenario 7 shows that, although we do not believe it likely, if we add a one-year outage in 2015 to Scenario 6, the SGRP remains cost-effective at the low gas price and the $815 million SGRP cost as long as the capacity factor remains at or above approximately 85%.
The NPVs for Scenario 6 show the SGRP to be more cost-effective than Scenario 3, the scenario determined to be most likely in D.05-02-052. Scenario 7 is Scenario 6 plus a one-year outage, and corresponds to Scenario 4 in D.05-02-052. The NPVs for Scenario 7 show the SGRP to be more cost-effective than Scenario 4, except under the High Gas assumption at 85% and 80% capacity factors where it is only slightly less cost-effective ($2 million and $7 million less, respectively). Therefore, for the reasons discussed herein and in D.05-02-052, we find the SGRP cost-effective.
8 Base capital additions exclude the SGRP costs and specified major capital projects included in the application. After 2015, only base capital additions were included in PG&E's forecast.
9 NYMEX is the New York Mercantile Exchange.
10 The NPV refers to the net present value to ratepayers of the revenue requirements resulting from the estimated costs and benefits of the SGRP. It is calculated using PG&E's Monte Carlo simulation model.
11 The "Low Gas" estimate is lower than the "High Gas" estimate.
12 Reducing the capacity factor reduces the replacement energy costs because Diablo is generating less energy that needs to be replaced.
13 At the time D.05-02-052 was written, the tube degradation results of the inspections of the steam generator tubes during the November 2004 outage of Unit 2, and the early 2004 outage of Unit 1 were not known. Therefore, for the purpose of this scenario, it was assumed that the results of these inspections would indicate that Unit 2 would go out of service without the SGRP one refueling cycle later.
14 A one-year outage was assumed to occur in 2015 if the SGRP is performed. A one-year outage occurring after 2015 would have a lesser effect on cost-effectiveness because of the time value of money.
15 At the time D.05-02-052 was written, the tube degradation results of the inspections of the steam generator tubes during the November 2004 outage of Unit 2, and the early 2004 outage of Unit 1 were not known. Therefore, for the purpose of this scenario, it was assumed that the results of these inspections would indicate that both units would go out of service without the SGRP one refueling cycle later.
16 Scenario 3 assumed one unit will run for an additional refueling cycle without the SGRP. Scenario 5 assumed both units will run for an additional refueling cycle. Since Scenario 6 assumes that Unit 1 will run for one additional refueling cycle, and Unit 2 will run for one half of an additional refueling cycle, one would expect the results for Scenario 6 to be between the results for Scenarios 3 and 5. However, that is not the case. The calculations for Scenarios 3 and 5 were developed by adding the estimated effects of delaying the expected shutdown dates for one or both units by one refueling cycle respectively. No changes to the operations of Diablo were included except in the additional refueling cycles. This tended to understate the cost-effectiveness of the SGRP. For Scenario 6, a more comprehensive calculation was performed. It included changes to plant operations other than in the additional refueling cycles due to the updated degradation results. The revisions included such things as changes to the expected shutdown dates, increased probability of mid-cycle outages and extended outages, and increased inspection and repair costs. Scenario 7 also reflects these changes.
