Energy Power Supply

Adequacy/Reliability Study Proposed Analytical (Phase 1) Work Plan

December 4, 1998

What questions are we addressing in the overall study?

• What is the level of reliability, in probabilistic terms, that we would expect in the future assuming no new generation supply, no new transmission and no new demand-side response?   Ability to meet load on an hourly basis, i.e., both energy and capacity, is at issue.

• What is the level of reliability, in probabilistic terms, that we would expect to get from market-driven generation suppliers?

• What are the current institutional constraints on the generation supply market, if any (lack of market information, lack of purchasers with load responsibility, etc.)?  Are they significant enough to be concerned about?

• What are the current institutional constraints on demand response to market conditions?   (In this case we can start by assuming they are currently significant.)

•  What actions could be taken to make the market more efficient, both for supply entrants and demand-side entrants?

• What actions are likely to lead to a dead end and should be avoided?

The technical analysis described in this workplan is intended to answer the first two questions and set the background for the policy analysis of the remaining three.

Note that there will be locational advantages to generation and/or demand reductions in some areas (generation deficient areas with impending transmission constraints like the west side, south of the north-of-John Day cut plane to replace potential fish loss of generation supporting intertie capability).  Bonneville has raised at NRTA an transmission issue that is parallel to ours on the generation side: what transmission improvements are needed to maintain future reliability and who will pay for them.  In certain areas, generation and transmission are substitutes for each other and the two studies will need to be considered together.  The modeling for the Council's study will represent some transmission constraints within the Northwest, and their effects, in a simplified manner.

Analytical Process

The key analytical problem is doing a frequency analysis on a hourly basis for unserved load under water, load, and resource availability uncertainty.  The ideal would be to end with a duration curve of unserved load with existing resources:  The peak would be load not served under the adopted reliability criterion and the remaining layer would be load we would expect to be met by new resources, however they would be developed (Part 2 of the project ? institutional and policy analysis).  The reliability criterion is a potential variable in this.

The random variables are water, extreme temperature events and thermal forced outages.

The following sections describe pieces of the analytical process.  The expanded model described below is in need of a new name.  ?ISAAC? is used below, recognizing its lineage.  Suggestions are welcome and may be rewarded.

Weather-driven demand input into ISAAC 

• Monthly and hourly load shapes from hourly load model, using regional temperature data as input.
• We are interested in random variation in hourly load shapes within ISAAC to account for weather effects, which would replace secular trend uncertainty which we take as fixed for this study.
• Issue: degree of correlation of temperatures (and loads) with water conditions
• Proposed mechanism:  Pursue availability of regional hourly temperature data, weight according to ELCAP site weights (this structure built into ISAAC and the load model already), attach to historical hydro data and make random draws from the combined set.  Each draw of a year would give both a hydro record and an hourly temperature record that occurred that year.
• To investigate: Whether cold spells, the real area of concern, are correlated with winter average temperatures.  If they are not, an additional random variable would be needed; if they are, the above procedure should capture it properly.
• For scenario analysis, specific load shapes and levels can be used as well.

Import capability

Import capability on the southern intertie is a function of net west side load and east-to-west flows on the Midpoint-Summer Lake line.  The winter 1998-99 operating limits based on analysis of 1996-98 winter data suggest that a reasonable reliability limit for this kind of planning study would be about 5,000 MW.

• Under less severe conditions, the operating limit would be substantially higher.  A frequency analysis can be done on this data as on the other data.
• As a check, it would be useful to compare west side loads in January 1998 (the maximum in the data set used to derive the operating limits) with the west side loads in February 1989 or December 1990 (with some adjustment for load growth) or alternatively, the west side temperatures for the two periods, to see if the data are representative enough of the conditions we are seeking.
• Proposed mechanism:  Make import capability from the south a rough function of net west side load in ISAAC, using the regional temperature data.  Assume fixed import capability on the eastern and northern interties.

Western market prices

Use HOSS output to define sustained peak, minimum generation and instantaneous peak for the various hydro regulations for different fish scenarios to use as input to Aurora.  Aurora calculates western prices and resources for input to ISAAC.

• Aurora is our only tool that allows an estimate of both 1) the generation supply response as a function of market prices and 2) the reliability level (calculated as reserve margin) as a function of demand bids (captured in different costs for blocks of outage).

Basic analytical tool

Use ISAAC for basic analysis.  The one dam model is probably sufficient for the future value of water calculation (which requires multiple passes through the future) rather than trying to modify ISAAC to incorporate the full regulator for this function.  ISAAC needs to be modified to incorporate the full regulator for dispatch.

• ISAAC is being modified to do chronological dispatch for capacity (replacing the current load duration curve method using Booth-Baleriaux probabilistic capacity analysis.  It is set up to do multiple sub-daily periods (currently using four) chronologically over a typical week.
• This approach would still use the trapezoid (described below), or some other shortcut so as not to require a within-week regulation, as HOSS does.
• ISAAC is being modified to dispatch with multi-area capability, recognizing limited transmission constraints, captured as fixed megawatt limits.  Currently we envision two areas within the Northwest: east and west of the Cascades.
• Are there others that would have a significant effect on the results of this analysis?

Capacity modeling validation

For selected weeks:
1) Use the chronological multi-area ISAAC dispatch to estimate potential hourly problems within the week.  This will be based on an internal redispatch of the output of the trapezoidal model, which calculates the ability of the system to meet a sustained peak approximation of the hourly load.  The trapezoid is currently a stand-alone model but is to be incorporated into ISAAC.
2) Hydro regulator output to HOSS.  HOSS calculates ability of system to meet actual hourly load for the week.
Compare HOSS and ISAAC output.  Does ISAAC give a significantly different answer about ability to meet the hourly load characterized in the two different ways?  A significant difference could come from some sequential drafting problem through the week that forces misses of a non-uniform daily pattern.  Instantaneous lack of machine capacity is another possible significant difference.  Define adjustments that might need to be made in interpreting ISAAC output.

• We may need to do this test for several different fish scenario hydro regulations to ensure confidence in consistent similarities or differences between HOSS and ISAAC.

An alternative to using the trapezoid is the function that relates hydro sustained peak capability to monthly energy that is currently contained in ISAAC.

The effect of extreme cold spells on side flows needs to be established.

Analytical Flow Chart [graphic not available]

Advisory Committee