> 685; bjbj -8]npppppp$_Snnvnn kFѴ4nCriticism of VELCO's Critical Load Study and Probabilistic Analysis, in further response to VELCO's specific question in deposition and in response to my testimony.
1. Mistake duration for frequency. Probability of failure is much less than the probability of unavailability. The longer a "contingency" lasts, the more the operator can take remedial action that changes the way he operates the system and in turn reduces the likelihood and therefore the statistical independence of the other contingencies. This is the problem of mixing the apples of frequency with the oranges of duration even in a deterministic study. The Critical Load study pretends that, because of its likely long duration, a Highgate outage is therefore the most likely besides being the single largest, an assumption that is not necessarily correct because a slightly smaller much more frequent "contingency" would be more "critical" or significant than a given contingency. Therefore, if a long-lasting "contingency" like a Highgate outage is to be used as a base to examine second contingencies (CL p. 8), it should itself be probability weighted to reflect its effect of reducing the likelihood of the other contingencies through planned remedial action during its duration and thereby of placing less stress on the line than were the remedial action not performed. Accordingly the Critical Load Study understates the critical load levels for not probability weighting the Highgate-outage starting point.
2. Arbitrarily apply probability. The documents fail to count the 100-120 MW of extra thermal generation because of "low probability" when contingencies of equally low probability are treated as certain in the deterministic study (CL p. 9, PA p. 13). This is hypocritical if the generation is still included in rate base (as it likely is, even if largely paid off) by the embedded utilities because it could not honestly be included in ratebase if it truly had a near-zero probability of deployability. The critical load study selectively applies "probability" where convenient: here (and questionably), but not to the Highgate outage in 1 above (where it would be very reasonable). Inclusion of the 100-120 MW of thermal generation significantly raises the critical load above the level it is in the Critical Load Study.
3. Base performance strictly on the past. Because the critical points have never actually been reached, the analysis fails to take into account price incentives to run the 100-120 MW of extra thermal generation (PA p. 13). Operating at the critical point would prompt enormously high spot prices that would surely incent deployment of more of the 100-120 MW of thermal generation, throughout the Highgate outage if not instantaneously. Proper pricing of reserve by ISO-NE for its deliverability and its quickness of response and not just for its capacity to provide energy would incent the operation of thermal generation in a mode to provide instantaneous support for a Highgate outage.
4. Fail to explain the 27 system states that explode the list of 178 contingencies into 4806 load flow simulations (CL, p. 11).
5. Fail to adequately explain the source of the data on transmission outages other than by vague reference to Canadian Electrical Association data (PA, p. 12). There is no systematic industry record keeping of transmission outages, only generation outages (NERC's GADS database).
6. Don't justify tight low/high voltage criteria (CL p. 10). The voltage criteria used in the industry are system specific, set locally. It would not be difficult to bias a critical load study by arbitrarily selecting the voltage limits. In this case it is difficult for an outsider to make a judgement whether the 1.05/0.95 per unit times the normal voltage at 230 kV and 345 kV used in the VELCO study is justifiably high relative to the 1.08/0.92 (used by VELCO at 115 kV), 1.1/0.9, 1.12/0.88 sometimes used in other systems
7. Do not address calibration of VELCO's planning model against ISO-NE's state estimator.
8. Mistake "conditional probability" for "joint probability". The Probabilistic Analysis should be based on the "joint probability" of two contingencies but is instead based only on the probability of a contingency conditional on the non, joint or separate occurrence of a Highgate outage and a McNeil outage (PA p. 13).
9. Omit a probability distribution to measure the probability of a load loss of at least a certain size.
10. Don't define one-event-in-ten-years in terms of how small a load loss to consider as an event.
11. Bundle both the size and the frequency/probability of the load loss into "expected" load loss (PA p. 15). But one-event in 10 years cares about separating this information because a small enough more frequent "contingency" would fall within the one event in ten years threshold, while a big enough less frequent one would not, while both have the same "expected value".
12. The Probabilistic Analysis doesn't determine whether once-in-ten-years standard is violated or not (PA, p. 11). Only expected load loss conditional on occurrence or not of Highgate or McNeil outage is indicated and compared. No expected load loss based on joint probability is indicated.
13. Makes a circular time-line forecast by assuming no reduction in demand by congestion pricing. The DPS load forecast basis for the load time-line apparently assumes no impact on load of increased congestion price, in other words assumes the presence in the future of the very transmission capacity that it is being used to justify! It must certainly ignore the potential impact on demand of locational pricing in a NW VT demand-price zone, or the impact on demand of a higher demand price from the triggering of NW VT into a Designated Congestion Area.
9:RTefnpKMqstB*B*OJQJB*OJQJOJQJ
:;TUfgpqMNstqMNst +0P/ =!"#$%
[$@$NormalmH <A@<Default Paragraph Font8
\c:BA
G
JP<
B
ACd=Robert Blohm*C:\My Documents\VELCoCriticalLoadStudy.doc@g`@G:Times New Roman5Symbol3&:Arial"qhN&S&F '!20Criticism of VELCO's Critical Load Study and Probabilistic Analysis, in further response to VELCO's specific question in deposition and in response to my testimonyRobert BlohmRobert Blohm
Oh+'0DPht
Criticism of VELCO's Critical Load Study and Probabilistic Analysis, in further response to VELCO's specific question in deposition and in response to my testimonyrit
Robert BlohmVELobeobeNormal.dotm
Robert BlohmVEL1beMicrosoft Word 8.0s@Ik@FѴ@z{ѴF
՜.+,0hp
x
Robert Blohmf' *Criticism of VELCO's Critical Load Study and Probabilistic Analysis, in further response to VELCO's specific question in deposition and in response to my testimonyTitle
!"#$&'()*+,./012347Root Entry F~ѴgѴ91TableWordDocument-8SummaryInformation(%DocumentSummaryInformation8-CompObjjObjectPoolgѴgѴ
FMicrosoft Word Document
MSWordDocWord.Document.89q