tinue, and commercial square footage continues to grow more slowly than GNP, commercial electricity use will only increase as fast or faster than GNP as long as electricity use per square foot continues to increase. Electricity use per square foot in commercial buildings may continue to increase for several reasons. Only 24 percent of existing commercial building square footage but almost half (48 percent) of the new building square footage is electrically heated (33). If these trends continue the share of buildings that are electrically heated could double. Air-conditioning in commercial buildings is probably saturated. About 80 percent of all buildings have some air-conditioning. Small increases in electricity use per square foot can come about by air-conditioning more of the building and by replacing window air-conditioners with central or package air-conditioners. Window air-conditioners cool 20 percent of the existing building stock, but only 9 percent of the newest buildings (33). Greater use of office machines and automation might increase electricity use both to power the machines and to cool them in office buildings, stores, hospitals, and schools. Machines, however, are less likely in churches, hotels, and other categories of commercial buildings. The potential for increased efficiency of electricity use in commercial buildings is less well known than for residential buildings because commercial buildings are very diverse and the potential for increased efficiency depends partly on success in balancing and integrating the various energy loads: lighting, cooling, heating, refrigeration, and machines. OTA analyzed the theoretical potential for reductions in electricity and fuel use in commercial buildings in a recent report, Energy Efficiency of Buildings in Cities (71), and found that electricity use for lighting and airconditioning in commercial buildings can be reduced by a third to a half. Heating requirements also can be reduced substantially by recycling heat generated by lighting, people, and office machines from the building core to the periphery. There is still very little verified documentation of energy savings in commercial buildings, however, and therefore, considerable uncertainty remains about the potential. The results of the bottom-up analysis for the commercial sector indicate a demand for electricity that will increase for a few years at about the rate of increase of GNP but with wide ranges of uncertainty. It could be higher as a result of faster penetration of electric space heating and more air-conditioning and big loads from office machines, or lower as a result of big improvements in the efficiency of commercial building electricity use. By the 1990's, however, when the demand for electric heat in commercial buildings will be largely saturated, the trend rate of growth in electricity demand is likely to settle to the growth in commercial square footage, somewhat less than the growth rate of GNP. Conclusion Utility executives contemplating the construction of long leadtime powerplants must contend with considerable uncertainty about the probable future growth rates in electricty demand. The range of possible growth rates encompasses low average growth rates of 1 or 2 percent per year, which would justify very few new large powerplants, up to fairly high growth rates of 3.5 to 4 percent per year, for which the pressure to build several hundred gigawatts of new large powerplants is great, as will be clear from the discussion in the next section. The sources of uncertainty are many. Future trends in electricity prices are viewed differently by different forecasters, because there is disagreement about future capital costs of generating capacity, future rates of return to capital and future prices of coal and natural gas for electricity. There also is uncertainty about how consumers and industry will respond to higher prices, given many technical opportunities for improved efficiency in appliance use and industrial electricity use and increasing numbers of promising new electrotechnologies that could substitute for the use of oil and natural gas for industrial process heat. Several of the industries, such as iron and steel, however, where the new electrotechnologies could have the greatest impact, face an uncertain future. At present, utility strategy for supplying adequate electricity is influenced heavily both by the recognition of uncertainty about future growth in electricity demand and by recent financial dif ficulties and regulatory disincentives for large capital projects. The influence of utility rate regulation and current utility strategies are discussed later in the chapter. RESERVE MARGINS, RETIREMENTS, AND THE Powerplants are planned and ordered many years in advance of when they are needed. In order to produce power for sale by 2000, nuclear powerplants on a typical schedule would have to be ordered by 1988, or 1990 at the latest, and coal plants or nuclear plants on an accelerated schedule, must be ordered by 1992 or 1993. Sources of generating capacity with shorter leadtimes, such as gas turbines or coal conversion, need not be ordered before 1996 or 1997. There is considerable disagreement about how many new powerplants will be needed by 2000. Those who believe that large numbers of new powerplants will be needed (several hundred GWs) anticipate rapid growth of electricity demand (3 or 4 percent per year), expect that large numbers of existing plants will be replaced because of deterioration in performance or retirement due to age and economic obsolescence, and expect only modest contributions from small power production (20,83). On the other hand, some believe that no new powerplants or very few (a few dozen GW) will be needed before 2000 because they anticipate only slowly growing electricity demand (1 or 2 percent per year), expect little or no need to replace existing generating capacity, and expect substantial contributions to generating capacity from small sources of power such as cogeneration, geothermal, and small-scale hydropower (83). This section lays out the range of possibilities from no new powerplants to several hundred gigawatts of new powerplants arising out of different combinations of growth in electric demand and varying utility decisions about powerplant retirements and use of small power sources. The electric utility industry currently projects average growth in peak summer demand of 3.0 percent per year between 1982 and 1991 (68).* The industry has also planned for an increase in electric-generating resources of 158 GW by 1991 bringing the total generating capability up to 740 GW (68). Only 13 GW of scheduled retirements have been included in the 1991 estimate (69). At the same time the current reserve margin** of 33 percent is forecast to fall to 20 percent. As a rule, utilities like to maintain a reserve margin of 20 percent to allow for scheduled maintenance and repair and unscheduled outages. Individual regions may require higher reserve margins if they are poorly connected to other regions, if they are dependent on a small number of very large plants or if they are dependent for a large share of generating capacity on older plants or plants that burn expensive oil and natural gas. As shown in figure 7, the planned resources of 740 GW scheduled for 1991 would allow a reserve margin of 20 percent to be maintained until 1996 if electricity peak demand grows only at 2 percent per year, or until 2000 if electricity demand grows only at 1 percent per year. On the other hand, the average reserve margin will fall below 20 percent by 1987 if electricity demand increases at 4 percent per year. However, the number of new powerplants that must be built to maintain a given reserve margin does not depend only on the rate of increase in *This had dropped to 2.9 percent per year for 1983 to 1992 in the 1983 North American Electric Reliability Council Forecast of Electric Power Supply and Demand (70). *"'Reserve margin" is defined as the percent excess of "planned resources" over "peak demand" where "planned resources" includes: 1) installed generating capacity, plus 2) scheduled power purchases less sales. Capacity/peak demand in GW Figure 7.-Projected Generating Capacity and Alternative Projections of Peak Demand NOTE: "Planned resources" is defined by NERC to include: 1) installed generating capacity, existing, under construction or in various stages of planning; 2) plus scheduled capacity purchases less capacity sales; 3) less total generating capacity out of service in deactivated shutdown status. "Reserve margin" given here is the percent excess of "planned resources" over "projected peak demand." SOURCE: North American Electric Reliability Council, Electric Power Supply and Demand 1982-1991, August 1982 and Office of Technology Assessment. electric peak demand. It also depends on how much existing generating capacity must be replaced because powerplants are retired, due to age or to economic obsolescence, or because existing powerplants must be derated to lower electricity outputs. Retirements Due to Age.-The "book lifetime" of a powerplant, used for accounting purposes, is usually 30 to 40 years. Over this period the plant is gradually depreciated and reduced as a recorded asset on the utility's books until, at the end of the period, it has no more book value and produces no return on capital. However, in practice powerplants may continue to operate for 50 years or more. As of 1982 there were about 10 GW of generating capacity that were more than 40 years old, more than a quarter of the total generating capacity that was in service 40 years ago. In fact, the bulk of the current generating capacity of the United States is comparatively new. Over half has been built since 1970, as shown in figure 8. The number of plants that would be retired by 2000 varies greatly with the assumed plant life. In the unlikely event that a 30-year life would be used, over 200 GW would be retired by 2000 (see table 6). A 50-year schedule would retire only 20 GW. Economic Obsolescence.-From 1965 to 1979 a large number of steam-generating plants using oil or natural gas were built (see fig. 8). They were natural gas to utilities is expected to increase steadily through the 1980's as the long-term contracts for natural gas sold at relatively low prices expire and are replaced by contracts for more expensive gas. As of 1981, there were 152 GW of oil and natural gas steam-generating capacity. Together they totaled 27 percent of all generating capacity but produced only 22 percent of all electricity. As shown in table 7, oil-fired steam plants produced only half as much electricity relative to their share of generating capacity. Natural gas-fired steam plants, on the other hand, produced a greater share. Even though oil and natural gas will be expensive, plants burning these fuels can be used as part of the reserve margin. Oil and gas are, in fact, just about 20 percent of two regions, the Southeast (SERC)* and the West (WSCC). The fraction of oil and gas-generating capacity, however, is much larger than 20 percent in three regions: Texas (ERCOT) about 72 percent, Southwest Power Pool (SPP) about 56 percent, and the Northeast Power Coordinating Council (NPCC) about 51 percent (68). If oil and gas steam plants were retired continuously in these regions until they formed no more than 20 percent of total generating capacity, the total retired would be about 55 GW. Loss of Availability of Generating Capacity.The percent of time that nuclear and fossil baseload plants were available to generate electricity averaged around 70 percent** over the decade of the 1970's (67). If there were a reduction from 70 to 65 percent in the average availability of nuclear and coal powerplants this would be the equivalent of a loss of 21 GW out of a total current coal and nuclear-generating capacity of 294 GW (see table 6). A recent study for DOE assesses the prospects for changes in average availability (82). Statistical *These are the regions of the Northeast Electric Reliability Council (NERC). **The availability figure used here is equivalent availability and includes service hours plus reserve hours less equivalent hours of partial outages. (66) From 1971-80, nuclear plants and coal plants over 575 MW averaged 67.8 percent in equivalent availability and coal plants from 200-574 MW averaged 74.3 percent in equivalent availability. Table 7.-Installed Capacity and Net Electricity Generation by Type of Generating SOURCE: North American Electric Reliability Council, Electric Power Supply and Demand 1982-1991, August 1982. evidence from the past two decades would support an estimate of a loss of 3 to 5 percentage points of average availability for every 5-year increase in average age of coal plants. Looking ahead, there could also be losses in availability of several percentage points due to emission controls and requirements for low sulfur coal that is at the same time of lower combustion quality. Offsetting these tendencies to reduced availability, however, there are also forces that might increase average availability. The utility industry has completed a period of construction of coal plants with poor availability, and the newest plants (from the late 1970's and early 1980's) should have substantially higher average availability. If this were to continue, overall availability could increase. If utilities invest in higher availabilities (e.g., by converting forced draft boilers to balanced draft), this will also increase availability (82). It is clear that attention to fuel quality and good management also can raise availabilities. Some Public Service Commissions (e.g., Michigan) are including incentives to improve availabilities in utilities' rate of return formulas. On balance, it is unlikely that availability will increase or decrease dramatically. If a change in availability should occur, however, it would have a noticeable impact on the need for new capacity. A 10-percentage-point change could imply an increased (or reduced) need for powerplants of more than 40 GW by 2000. Summary-The Need for New Powerplants.the need for new powerplants depends on both the growth rates in electricity demand and on the need for replacement of existing generating capacity. Table 8 summarizes most of the range of disagreement and its implications for new powerplants. Estimates of growth in electricity demand range from 1 to 4 percent per year. (The table shows the implications of electricity demand growth rates of 1.5 to 3.5 percent.) Judgments about replacement of existing plants can, somewhat arbitrarily, be divided into high, medium, and low replacement. A high-level replacement of about 200 GW by 2000 would be necessary to: offset a slippage of about 5 percentage points in availability, meet a schedule of 40-year life expectancy for all powerplants, and retire about half the oil and gas capacity in this country (see table 6). A low-level replacement of 50 GW would meet a 50-year schedule, retire a little oil and gas capacity and would assume no slippage or an actual increase in average availability. If these alternative replacement assumptions are combined with alternative growth rate assumptions (table 8), they lead to a wide range of needs for new plants. About 454 GW of new capacity would be needed, for example, by 2000 (beyond NERC's planned resources for 1991) to meet a 3.5 percent per year increase in peak demand for electricity and the high replacement re |