System Analysis Advisory Committee Meeting Notes
February 27, 2003 - 8:30 a.m. - 3:30 p.m.
NORTHWEST POWER PLANNING COUNCIL OFFICES
PORTLAND, OREGON
I. Greetings, Introductions and Review of the Agenda.
The
February 27, 2003 System Analysis Advisory Committee meeting, held at the
Northwest Power Planning Council's offices in Portland, Oregon, was
chaired by Michael Schilmoeller of the Council staff.
The
following is a distillation (not a verbatim transcript) of items discussed
during the call, together with actions taken on those items. Please
note that some enclosures referenced in the body of the text may be too
lengthy to attach; all enclosures referenced are available upon request
from Schilmoeller at 503/820-2314.
Schilmoeller welcomed everyone to today's meeting, led a round of
introductions, then reviewed today's agenda. Schilmoeller noted that
copies of his presentation are available via the NWPPC website; please
refer to this document for full details, including graphs and
charts.
2. Notes from the February 7 SAAC Meeting.
The
notes were approved as written.
3. Incentives for New Generation.
Schilmoeller began by noting that incentives for new generation is
specifically referenced in the fifth Power Plan; he said the events
of recent years have led many to ask whether industry should return to
planning criteria that guarantee a level of reliability. Some believe we
may return to a period of volatility, high prices and short supply,
similar to what we saw in 2000-2001, he said. Using the overhead
projector, Schilmoeller went through some of the options available to
address resource inadequacy, then touched on FERC and California
incentives for new generation. Schilmoeller described some of the problems
associated with a centralized administration to address such incentives,
then moved on to wealth transfer issues.
Next, Schilmoeller addressed difficulties with a reserve margin planning
criterion; mainly, that it is ?yesterday's? solution and does not
address the general issue of risk. The group devoted a few minutes of
discussion to this topic, debating Schilmoeller's statement that the
2000-2001 energy crisis was a problem primarily of over-reliance on
wholesale power markets and unexpected prices in those markets, and had
its roots in poor resource adequacy.
Schilmoeller then provided an example illustrating his proposed approach
to this issue in the Portfolio model, noting that the least-cost hourly
deterministic solution, in this instance, proved to be reliance on the
market. One participant noted that this is not a long-term equilibrium
solution; Schilmoeller replied that it is not intended to be. It reflects
the fact that the future rarely unfolds as we would predict.
We're in agreement that, looking ahead, there may not be enough
incentive to build a plant, in this case, said Ken Corum. The solution is
that people have to recover their costs, said Sher ? they can't just
ignore the sunk costs. Greg Nothstein said the thing that concerns
potential resource builders isn't the dry year ? it's the wet year.
Corum observed that Schilmoeller's statement that this is the least-cost
deterministic solution is only true if there is no reserve margin
requirement; Schilmoeller agreed.
So
what we're doing here is trying to analyze a reserve margin requirement?
Corum asked. Correct, Schilmoeller replied. He then moved on to another
example, intended to illustrate the least-cost probabilistic solution ?
in 2,000 trials with Crystal Ball, assuming 57% uncertainty in gas and
power market prices, the total cost of relying on the market was $38
million. The optimum solution, under this scenario, was 383 MW of
combined-cycle combustion turbine (total cost: $36 million).
The
last thing we looked at was a least-cost probabilistic solution with a
risk constraint, Schilmoeller said. The optimum solution, in this case,
turned out to be 500 MW of combined-cycle combustion turbines and 417 MW
of wind generation, with a 22% reserve margin (total cost ? $48
million). The group offered a variety of clarifying questions and
comments, to which Schilmoeller replied.
Next, Schilmoeller put up a graph illustrating the concept of the
efficient frontier ? the relationship between risk constraint and cost.
So as the cost of meeting the load goes up, the risk goes down? one
participant asked. Right, said Schilmoeller ? that's what happens when
you buy insurance.
Moving on, he touched on the implications of incentives for generation:
risk constraints provide more general protection (the choice of wind
provides a hedge against the cost of gas generation); the issue of wealth
transfer is moot; drivers and homeowners are required to carry insurance
to protect others or society -- who is protected by ?insurance?
carried by a load-serving entity? What are the externalities? These
issues are at the core of whether load-serving entities (and ratepayers)
should be required, effectively, to pay for insurance.
The
thing that bothers me is the distinction between the load-serving entity,
who is managing the risk, and the ultimate consumer, who is the one who is
bearing the risk, said Dick Watson. In my mind, it is the PUCs which are
effectively writing the check in the name of the consumer, Schilmoeller
replied ? they're in the best position to make the call about how much
risk and how much volatility their consumers are willing to tolerate. The
game of what is the diversity needed in the market is a tough game to
play, Sher observed ? that's a tough question whether you?re a
self-regulated or a publically-regulated utility.
Schilmoeller then offered a series of conclusions:
?
Those responsible for rate stability should use risk-constrained
least-cost planning to protect their constituents
? Such
analysis is feasible
?
Discussion should turn to the extent to which there are externalities
associated with high retail rates, such as social harm
?
To the extent that there are social externalities, some kind of
enforcement would probably be necessary.
The group discussed the issue of how this portion of the Portfolio model
would address curtailment; Kurt Granat observed that there are some
instances in which everyone must and should pay ?taxes.? I'm
concerned about FERC saying the reserve requirement is 15%, without
taking into account diversity, Sher said ? it's a very complicated
equation. What I'm hearing is that you?re advocating the idea that
it makes the most sense for individual utilities to analyze and address
their own reserve requirements, Schilmoeller said. Correct, Sher
replied.
One topic for a more long-term discussion, said Watson ? talking
about CVARs among the tekkies is one thing, but how do we turn this part
of the model into something that makes sense for the actual
decision-makers? it's a question of balancing portfolios to manage
risk, Schilmoeller replied ? Hopefully, we could show them portfolios
and risks that resemble circumstances they are already familiar with --
such as prior states of the system.
4. Detailed Assumptions Around Renewables and Distributed
Generation.
Schilmoeller began with a slide showing some specific renewables and
distributed generation technologies (wind, solar, microturbines, diesel
engines, fuel cells); he noted that there will be an internal discussion
tomorrow among Council staff to stimulate thinking about what
information is needed about these technologies to feed into risk
assessment. Will you add conservation and demand-side resources to this
list? one participant asked. We'll be handling those separately,
Schilmoeller replied.
Are there other technologies that should be on this list? he asked. Sher
suggested landfill gas recovery. Jeff King noted that the focus of this
list is really ?green? resources and resources with lead times of
less than one year. It seems to me we're trying to gain understanding
about risk reduction potential, he said; we should probably begin by
discussing which risks, such as CO2 emissions and capacity shortfalls,
we're trying to avoid. The commercial availability of some of these
resources is another factor to consider, Schilmoeller said.
Schilmoeller then went through the cost, service life, development
period, mothball periods, capacity factor, shaping penalties,
dispatchability and risk attributes of each of these resources classes.
The group devoted a few minutes of discussion to this information.
Schilmoeller then provided a series of conclusions. The group
agreed that -- for the purpose of analysis -- we could aggregate
resources into groups based on their operating and risk characteristics.
5. Statistics (Continued).
Marty Howard led a discussion of statistics associated with power and
gas; he noted that most of the statistics he has uncovered have to do
the gas. He said his objectives were to be analytically descriptive of
what the data looks like, to show how the results may be useful for
Portfolio modeling, to provide plausible answers to specific questions
-- for example, how gas market prices are related to power market prices
-- and to provide a final revelation. Essentially, we need to understand
know how these factors behave and interact with one another, said
Howard.
Howard touched on some basic concepts, including dependence and
independence; he then provided some examples illustrating these
concepts. Howard noted that he has data from nine hubs, which he has
combined into six; he then moved on to a series of bivariate gas price
graphs, pointing out the interesting characteristics of and
relationships between each graph, then providing a few observations on
this data. Howard then turned his focus to Sumas, running through a
variety of statistical analyses focused on this firm's gas prices over
time, including results from a Sumas seasonal and trend model. He noted
that his purpose is not to create a predictive model to forecast future
prices; rather, the model is simply intended to duplicate the shape of
the historic price trend. The reason for its development, of course, is
to inform Michael's Portfolio model.
In
terms of next steps, Howard said he is planning to do a similar trending
and seasonality treatment for the Mid-C; he said he will also be
applying the same ARMA and Del 1 transformation to the Mid-C raw
results. Further in the future, he intends to examine other series and
discrete events.
Next SAAC Meeting Date.
The next meeting of the System Analysis Advisory Committee was set for
Thursday, March 20 at 9:30. Subjects to be discussed include more
discussion of statistics and results with Olivia. Meeting summary
prepared by Jeff Kuechle, NWPPC contractor.
|
