Fortnightly
Using Hourly System Lambda to Gauge Bulk-power Prices
May 01, 1995By William C. Booth and Judah L. Rose
The electric power industry lies in the midst of major change, including a shift to market-based wholesale prices. Market players and regulators will recognize that competition requires a shift in thinking on key issues such as resource planning before the market is developed enough to provide adequate price information. It is during this interim period that newly available hourly system lambda information could prove useful.
The outlines of the future wholesale power market are becoming increasingly clear. The market will probably resemble other commodity markets. Suppliers and buyers will engage in price discovery and a mix of short- and long-term transactions. As the markets develop, market-price hedging mechanisms such as futures contracts may emerge and even become commonplace.
This final outcome will likely occur only after years of debate and experimentation. Currently, market participants, regulators, and others are debating key structural issues. These debates cover system operations, reliability, market making, and transmission issues. For example, in the California restructuring debate, "PoolCo" and "NetCor" conceptual frameworks are being debated. In many other states, regulatory agencies are initiating review of retail wheeling, stranded assets, and other aspects of deregulation.
And these debates won't occur in a vacuum. As they continue, utilities will downsize, merge, write off assets, and search for ways to cut operating costs. Wholesale and retail customers will continue to look for cheaper suppliers for their existing loads. Power marketers will work to shape commercial arrangements to their vision of a full-fledged electric commodity market.
Nevertheless, as it awaits the future, the electric industry in some key respects will appear as it always has. Utilities will maintain reserves, operate their systems, and remain subject to regulation. But regulators, producers, wholesale customers, marketers, and others will increasingly seek price information \(em information not yet fully available. We believe that during this period the industry can and will use hourly system lambda data as price proxies.
Hourly system lambda data has been available since June of 1994 from the Federal Energy Regulatory Commission (FERC). But this data now offers new uses, such as:
s Tracking incremental cost savings
s Evaluating wholesale offers
s Resource planning and plant valuation.
For many participants, these commodity market concepts will represent a "gestalt shift." We believe that the changes occurring in the electric market will recast planning and lead to greater public scrutiny of the industry. Although planning approaches will be different, we do not expect the substance of decisions to change. The most economic options will not change. They will just be found in a different way.
Hourly System Lambda
Electric competition received a major boost with the passage of the Energy Policy Act of 1992 (EPAct). This legislation created "electric wholesale generators" and gave wholesale customers the right to request transmission service. The EPAct also requires the FERC to collect and make public transmission planning information. In creating the new FERC Form 715, Annual Transmission Planning and Evaluation Report, the Commission also changed its existing FERC Form 714 to require reporting of hourly system lambda and demand data. This information was first submitted in June 1994 and is now available to the public via electronic bulletin board.
System lambda is closely associated with the marginal cost of producing electricity and offers a good indicator of the competitive price for electrical energy. System lambda is a product of control area economic dispatch. Certain plant owners are accustomed to tracking the changing level of customer demand. Each of these dispatchable plants exhibits its own production cost curve, which identifies variable costs (mostly incremental fuel cost) at each level of output. The system lambda represents the variable cost of the last kilowatt produced over a particular hour. The plants most closely following load tend to be utility units with higher variable operating costs (see Figure 1). That's intuitive; you want to use the power plants with the lowest variable costs as much as possible.
Electricity demand can fluctuate widely, between hours, days, and seasons. In general, system lambda correlates with load levels (see Figure 2).1 Utilities with high cost generators can be expected generally to carry higher system lambdas. Thus, for example, utilities with mostly oil and gas power plants show higher average system lambdas than utilities that operate low-cost coal plants.
A Look at the FERC Data
Some of the newly available FERC system lambda data are shown for contiguous regions in the eastern United States (see Figure 3). This data shows the 1993 annual average (that is, an average over 8,760 hours of data). For 1993 the average lambda for the PJM power pool was nearly 22 mills (2.2 cents) per kilowatt-hour (Kwh).2 In contrast, the average system lambda is only 15 mills (1.5 cents) per Kwh for two leading utilities (American Electric Power and Allegheny Power System) located in the North American Electric Reliability Council Region (NERC) known as the East Central Area Agreement (ECAR). These companies own many power plants located close to the Appalachian coal fields to the west of PJM. The ECAR region is directly interconnected with PJM. Utility average system lambda data are also presented for some utilities to the north and south of PJM.
Several factors (among many) explain differences between PJM and ECAR utilities: 1) the generating fuel mix, 2) economy interchanges, and 3) transmission bottlenecks.
In 1993, PJM oil and gas generation was 5 percent of total mix. In contrast, in ECAR, oil and gas accounted for less than one percent of utility generation. A higher PJM cost seems logical. Also, sales of economy power from ECAR to PJM would yield cost savings. Both sides would benefit. The availability of low-cost ECAR power means that PJM's incremental sources of power come from ECAR utilities. In this respect, the large differential in average annual lambda is at first glance surprising. However, consider that high-voltage transmission interties between PJM and ECAR are heavily loaded. In actuality, this link is one of the most used inter-NERC regional interties in the country. Trades that would decrease the lambda difference may not be possible due to lack of transmission capacity.
Of course, differences in computational methods do place certain limitations on the FERC data. The FERC requires submission of system lambda data, but does not specify calculation procedures. Thus, while utilities may calculate their system lambda differently, the FERC requires utilities to explain how they calculate their reported system lambdas. This information facilitates analytic adjustments necessary for inter-utility comparisons.
Key calculation differences may come from the utility's fuel-cost method or whether it treats non-fuel O&M production costs as fixed or variable. For instance, the utility may calculate fuel costs on a marginal (e.g., spot coal) or average basis (e.g., contract). In other cases, variable costs may include production cost items in addition to fuel that other utilities might treat as fixed.
Analytical adjustments can prove difficult. System lambda may diverge from the marginal cost calculation made using all plants in a given area. In other words, the calculation of lambda can be more complex than it appears in Figure 1. This effect occurs for several reasons. First, some plants are not dispatched. A plant may operate to maintain transmission system reliability. Others may supply power at the decision of the plant owners, not the central dispatcher (e.g., some cogenerators). Second, some plants are limited in their ability to respond to load fluctuations by their designs (e.g., plant turndown constraints).
Finally, lambda should approximate marginal costs and competitive prices for small increments in demand, but may diverge for larger increments. The marginal cost of the next megawatt-hour (Mwh) can be different than the incremental cost of producing larger amounts of power (e.g., 1,000 Mwh). Variable costs might change for the marginal unit as the loading increases. If not, additional plants with higher costs must be used.
Despite these limitations, one thing is certain: Wholesale prices will gain in importance as competition grows in the power industry. But since the industry has functioned so long under regulation, the importance of system lambda data may not be apparent. Regulators and market players may be surprised at how the data will change their world. Consider these three examples: 1) incremental cost tracking, 2) wholesale competitiveness, and 3) resource planning and plant valuation.
Cost Tracking
Under traditional cost-of-service regulation, regulators ensure that electric utilities minimize costs consistent with prudent management. System lambda lends itself well to this task in two areas.
First, differences in system lambda may indicate added cost savings available from additional economy power interchanges. (Utilities with low system lambdas sell power to utilities with higher lambdas.) Such sales can make dispatch more efficient.
Second, differences in system lambda may also highlight needs for more transmission capacity to remove bottlenecks that block economic power transfers. Granted, utility operations may be close to achieving most available economies in these areas. After all, state and federal regulators periodically review bulk-power trading and transmission investment. But this key data should facilitate review.
Evaluating Wholesale Offers
Whether firm or nonfirm, wholesale price offers can be evaluated with the help of system lambda information.
Firm power transactions are backed up with capacity available to meet demand. Firm power sales are considered adequate to satisfy the regional power pool's requirement that all member utilities have capacity to meet expected peak demand with a reserve margin. In contrast, economy or nonfirm power transactions do not count toward the utility's reserve requirement and are provided on an "as available basis."
In the economy markets, a high system lambda should encourage a power plant owner to buy bulk power from utilities with lower marginal costs. In the firm market, system lambda also plays a large role. The typical sale is by an investor-owned utility to a municipal or cooperative on a full-requirements basis, with the seller bundling all services together: energy, backup capacity, and transmission services. However, by unbundling the price component by component, it can be seen that the firm wholesale supply price actually reflects three factors \(em average system lambda, capacity, and transmission.
Thus, the seller could choose to obtain the required electrical energy from the grid at a competitive price equal on average to the average annual system lambda. As for capacity, the seller could obtain firm backup by purchasing combustion turbine capacity (combustion turbines set a ceiling on capacity prices since they represent the newest plants with the lowest capacity cost). This capacity would hardly ever be used, but would provide necessary reliability. The costs of this capacity are shown to be spread over the kilowatt-hours of energy. Lastly, the seller would need to reserve firm transmission capacity.
In many cases, current firm prices based on regulated rates that reflect average historical ("embedded") costs are higher than competitive market prices. For example, 1993 firm wholesale contracts in PJM averaged 46 mills per Kwh (source: FERC Form 1). In contrast, the competitive price was about 33 to 35 mills/Kwh. This price reflects: 1) energy prices (lambda) in 1993 averaged 22 mills/Kwh; 2) capacity obtained at the annualized cost of a combustion turbine plant at 9 mills/ Kwh;3 and 3) 3-5 mills/Kwh for transmission.
And what's good for sellers is good for buyers. Retail customers (such as large industrials, public intervenors) will use the same information to ensure that their utilities are aggressively pursuing market offers. Power marketers will also use this information to identify possible sales opportunities based on purchases currently above market.
Resource Planning,
Plant Valuation
In the past, when precious little wholesale bulk-power price information was available, plants were valued by their ability to decrease system cost. This theoretical approach led to correct assessments. Comparing costs or revenue requirement offered a screening method for evaluating new capacity (see Figure 4).
The process will change in the future, however, with the framework recast not only in terms of prices (system lambda), but also energy value and capacity value.
Electrical energy value is measured as the difference between variable operating cost and system lambda (the marginal decrease in electrical energy costs). In Figure 5, the cost of a new generating unit (its variable operating cost) is compared to the system lambda over the course of a year. When the lambdas are high, savings are high (shaded area), and the plant or resource has high energy value. This energy value can be estimated over the useful life of the resource option on a present-value basis. Thus, system lambda will offer crucial information for evaluating resource options. While future rather than historical data is what is needed, forecasts will be calibrated to the new Form 714 and other data.
Capacity value reflects plant reliability. In regions with a reserve margin constraint, and no excess capacity, the value of this capacity is equal to the cost of the least expensive type of capacity to build \(em the combustion turbine. Thus, regulators and competing power producers who operate in regions that need capacity will ask two questions. First, "Is more capacity required?" If yes, then, "Is any resource effective other than a simple-cycle combustion turbine?" Plants other than combustion turbines will compete only if their energy savings offset their higher initial capital costs, relative to the costs of a combustion turbine. Energy savings in turn depend on system lambda.
The importance of lambda is highlighted by comparing the value of alternative plants for systems with annual average lambdas of 25 and 35. Table 1 shows
the annual costs and revenues
of new power plants, assuming average annual lambda equals 25 mills/Kwh. The net value is the sum of the operating revenues \(em which also equals marginal energy savings plus capacity value less the costs of the power plants. Costs include operations and maintenance and the carrying costs of the plant's capital investment, with these assumptions:
s Combustion Turbines. Capacity value only (zero operating hours).
s Combined Cycle Plants. Revenues exceed costs (the most economic).
s Coal Plants. High energy value (many operating hours); negative net value (high investment costs).
The results change at lambdas of 35. This effect occurs because the net value of new combined cycles increases, but not as fast as the value of coal plants (see Table 2). The energy value of the coal plant increases faster than that of the combined cycle since it operates more frequently (that is, peak and off-peak). Energy savings for coal plants become large enough eventually to offset their higher fixed costs.
Conversely, at lambdas well below 25 only combustion turbines make sense. Their net profits are zero (though they provide benefits to customers in terms of reduced blackouts and the like), but other options have negative values.
In essence, whether a power plant should be built will become largely a question of what the system lambda is and what assumptions (or in the case of turnkey bids, what "all-in" costs) underlie the analysis (such as power plant capital and other fixed costs, and variable operating costs).
Note also that a similar analysis can be performed for existing power plants, except that capital construction costs are likely to be zero. For example, in a competitive market, the value of existing power plants selling on a merchant basis may fall below book value or vice versa. t
William C. Booth is the special assistant to the director of the Office of Electric Power Regulation at the Federal Energy Regulatory Commission. The analyses and conclusions presented here are his own and do not necessarily reflect the views of the FERC, any individual Commissioner, or any other staff member. Judah L. Rose is a senior project manager at ICF Resources, Fairfax, VA. He thanks Charles Clark of the Electric Power Research Institute for introducing him to some of these background concepts. Nevertheless, the views presented here do not represent those of Mr. Clark or EPRI.
31
Articles found on this page are available to Internet subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.
Public
Utilities Fortnightly & Spark
