Ratio of later to earlier quantities

0

5

6

3

2

70

Concrete (yd3 per MW)

Piping (ft. per MW)

Cable

(ft. per MW)

Raceway (ft. per MW)

with much tighter tolerances to the pipes they support.

Quality-control procedures and paperwork have added to the cost of materials and components. Although there has been no comprehensive study, there are individual examples and anecdotes to illustrate the claim that quality control represents a bigger and bigger share of nuclear materials cost. In one such example (86) structural steel supports now required for nuclear plants cost between two and three times the cost of the same quality steel supports that are still used for general construction projects and that were permitted on nuclear projects until 1975. Of this amount, the quality control procedures account for virtually all the increased cost.

Finally, there has been a steady increase in the amount of labor required per kW, both manual (craft) and nonmanual. For a series of typical plants costed out over 15 years in a study for DOE, craft labor requirements increased from 3.5 workhours/kW for a plant starting construction in 1967 to 21.6 workhours/kW for the average of 16 plants under construction for completion in 1982-85 (17). Nonmanual field and engineering services also have increased dramatically. For a slightly different series of typical plants, estimates of field and engineering services increased from 1.3 workhours/kW in 1967 to 9.2 workhours/kW in 1980 (16).

The increase in labor per kilowatt of capacity is the result of complex interactions resulting from

increasingly demanding regulations, quality-assurance requirements and the subsequent utility management response to these. These are described in more detail in chapters 4 and 5 and in several case studies (92).

There is large variation in material and labor requirements from plant to plant, just as there is large variation in overall capital cost. For a group of 16 plants scheduled to be completed in 1982-86, craft labor varied from a low of 15 workhours/kW to a high of 33 workhours/kW. Similarly, for a group of eight plants, linear feet of cable varied from a low of about 3,300 ft/MW to a high of about 7,300 ft/MW (see fig. 17).

The Increase in Nuclear
Construction Leadtimes

Nuclear construction leadtimes also increased over the decade, making it increasingly difficult to match nuclear plants to demand, adding to interest and escalation costs, and exacerbating problems with cash flow. At the same time, leadtimes for coal plants increased very little (from an average of 58 to about 60 months) (1).

Documenting the increase in leadtimes for nuclear plants is made difficult by the fact that some plants have been delayed deliberately by their

Photo credit: OTA Staff

Quality-control inspectors must approve the position and tolerance of each pipe hanger before a section of a nuclear plant under construction can be considered complete. These ribbons indicate preliminary approval

utilities because of slow growth in electricity demand and financing difficulties. There also appears to be important differences in the regulatory environments for different generations of plants that must also be taken into account.

A recent study of leadtimes for EPRI took both deliberate delays and regulatory stage into account* (1). The study identified from published sources those plants that had been delayed deliberately more than a year by their utilities and analyzed their leadtimes in a separate group. In a more detailed case study of 26 of these plants EPRI found that 8 had been delayed significantly (averaging 27 months) while 22 had only been delayed an average of 2.5 months.

Grouped by date of permit, it is plausible to identify three generations of nuclear plants. For the first generation, for which construction permits were issued from 1966 to 1971, leadtimes** increased steadily from about 60 to 80 months. This appears to reflect an increase in the designed complexity of nuclear powerplants and possibly the strains of rapid growth as well.

A second group of plants had their construction permits issued from 1971 to 1974. Leadtimes for that group were much higher than the first, averaging 120 months and ranging from 100 to 160 months. Leadtimes for plants without significant deliberate delays averaged about 10 months less than those with significant deliberate delays. It appears that this group of plants suffered a major increase in regulatory complexity (including the 1974 Calvert Cliffs decision, the regulations following the 1976 Browns Ferry fire and the 1979 accident at Three Mile Island) without the opportunity to develop construction and regulatory planning techniques to handle the increased complexity. They may have suffered as well from some of the effects of rapid growth in the industry, such as incomplete designs and inexperienced supervisors.

Although the data are sketchy, it is possible that a third generation of nuclear plants is now emerg

*The study grouped the plants by date of construction permit to avoid the obvious problem that later completion dates, by definition, include a larger proportion of long-leadtime plants.

**Leadtimes for this analysis are defined as time from date of construction permit to commercial operation.

ing, with construction permits issued later, in 1975-77. After adjusting for deliberate delays and excess optimism in time estimates, EPRI found that this latest group of plants appears to have somewhat shorter leadtimes than the 1971-74 group. Leadtimes for all plants in the group average about 100 months and range from 65 to 120 months. Plants without significant announced deliberate delays average 10 to 15 months less than the average. The plants with shortest leadtimes in this group are already in operation and were completed faster than the shortest leadtime plants in the earlier group (see fig. 18). The numbers are so small, however, that it is too early to tell if these plants are anything but anomalies. Those plants with longer leadtimes in this latest group still may experience significant delays beyond the adjusted estimates calculated by EPRI. At the same time there is some case study evidence that the plants that were started later were able to compensate for increased regulatory complexity in the plant design and construction management and were also able to plan systematically

their dealings with the NRC. (Case Study 2 in ch. 6.)

The Impact of Delay on Cost

In a period of substantial general inflation characteristic of the last 5 years, a delay in nuclear plant construction can cause an alarming increase in the current dollar cost of the plant. Increases in the current dollar cost, however, must be distinguished carefully from increases in the real or constant dollar cost of the plant (after the impact of general inflation has been eliminated). These in turn must be distinguished from increases due to changes in regulations or other external influence during the period of delay.

For a hypothetical plant that has been expected to be completed in 8 years but instead has been delayed to 12 years, with no increase in complexity or scope, there are two sources of increases in total capital cost in constant dollars. One is that nuclear components, materials and labor may

have increased 1 or 2 percent faster than general inflation (escalation). The second is real interest during construction* that is capitalized and added to general total plant cost as AFUDC (AIlowance for Funds Used During Construction). (See box B above.)

Table 9 shows the increases in constant dollars and in current dollars of several different cases: 5, 7, and 9 percent general inflation with no real escalation in nuclear components and with 2 percent real escalation and real interest rates of 3 percent and 5 percent above general inflation (13). Several of the examples in table 9 can serve as illustrations of the difference between increases in current and constant dollars. For example, in case 3, if a plant takes 12 years to build during a period of general inflation of 7 percent, escalation of nuclear components of 9 percent (2 percentage points faster than general inflation) and an interest rate of 12 percent (a rather high real interest rate of 5 percent), the "mixed current dollar cost" of the plant will be 233 percent higher than its overnight construction cost. Two-thirds of the increase, however, is general inflation. The real constant dollar increase in the plant cost is only 48 percent. Construction of the same plant in 8 years time would cause a current dollar increase of only 123 percent and a constant dollar

*Real interest is the nominal rate of interest less the rate of general inflation, e.g., real interest is 5 percent for nominal interest rates of 12 percent when general inflation is 7 percent.

increase of 30 percent. For this case, shortening the plant's leadtime would save about a third of its current dollar cost but only about 12 percent of its constant dollar cost.

The Cost of Electricity From Coal and Nuclear Plants

The steadily increasing capital costs of nuclear power (including the increasing costs brought about by increasing leadtimes) leads to a crucial question: at what point does the increasing capital cost of nuclear plants make nuclear power a more expensive source of electricity compared to alternative generating sources, especially coal? As long as it is likely that utilities will avoid the use of oil and gas for baseload electricity generation, the chief competitor to nuclear is coal.

Initially (for most nuclear plants completed by the early or mid-1970's) there was no doubt that electricity generated from these plants was substantially cheaper than coal-generated electricity. Because of the way capital charges are recovered in the rate base (see box C above), the cost of electricity from these plants has become steadily cheaper relative to electricity from coal plants built at the same time. As the capital cost of nuclear plants has risen, however, the relative ad

*In this case 2.23 (8 years) is about 67 percent of 3.33 (12 years) and 1.30 (8 years) is about 88 percent of 1.48 (12 years).

vantage of nuclear power has diminished. For plants presently under construction, the average capital cost is now so high that the typical nuclear plant probably would produce more expensive electricity over its life time than the typical coal plant. Only electricity from the least expensive nuclear plants still may be competitive with average cost coal-generated electricity. Average cost nuclear plants, however, still can compete with more expensive coal plants.

Comparing the costs of nuclear and coal-fired electricity is made difficult by the different impact of fuel and capital cost components for each type of plant. Capital cost is a far more important component of nuclear-generated electricity than for coal plants. While the levelized cost of fuel (uranium ore, enrichment, storage shipment and disposal) and operations and maintenance each run about $0.0075/kWh, the capital charge (levelized charge* over the life of the plant) per kWh for new plants may range from as little as $0.01/kWh for older reactors to $0.10/kWh or even more for the most expensive of today's reactors. The capital charge per kWh increases with higher total construction cost (including the impact of longer leadtimes), with higher interest rates, and with a shorter capital recovery period. The capital charge (as well as operations and maintenance) per kWh also increases as the plant capacity factor** is reduced because there is less output among which to apportion the annual capital cost. Since the earliest nuclear plants were built, there have been significant increases in all the categories that increase annual capital charges. The capacity factors of nuclear plants also have been less than expected. (See ch. 5 for more discussion of nuclear capacity factors.)

For coal-fired electricity, on the other hand, the cost of the fuel is at least as important as the capital cost in determining the price of electricity over the life of the plant. Operations and maintenance of coal plants cost somewhat less than

*Various techniques are used to "levelize" costs over a plant's lifetime. One simple method, used in the EIA study of coal and nuclear costs, is to take the present discounted value of the stream of costs and divide it by the number of years to get an annual levelized cost (36).

**Capacity factor equals the number of hours of actual operation divided by the hours in the year.

for nuclear plants. Fuel cost, however, may range from less than $0.01/kWh in regions where plants can be built near the coal mine to almost four times that in regions located far from coal fields (assuming rapid increases in coal prices). The rate at which coal prices are likely to escalate over several decades has a significant influence on the forecast average cost of electricity from the plant over its lifetime. Levelized electricity prices will be about $0.015/kWh (in constant dollars), higher, on average, if coal prices escalate at a real annual rate of 4 percent than if they don't escalate at all in real terms (36).

Unfortunately, there is no study of recent plants using actual reported capital cost of coal plants. In a DOE study (36) using coal and nuclear plant model data on capital cost, the cost of electricity is about equal in five of the ten DOE regions, slightly lower for nuclear in two of the regions and considerably lower for coal in two of the regions. There is reason to believe, however that nuclear capital costs are higher and coal capital costs may be lower than the study results. Capital costs of the typical nuclear plant reported in the study are about 15 percent lower than the average (in constant dollars) of the plants now under construction. On the other hand, the capital cost of the typical coal plant reported in the DOE study is more than 40 percent higher than the capital cost of the typical 1978 coal plant (including a flue-gas desulphurization scrubber) in the 1981 Komanoff study updated to constant 1982 dollars. While it is possible that the capital cost of coal plants may have increased 40 percent since 1978, several factors make it unlikely. Coal plant construction leadtimes (unlike nuclear plant leadtimes) have not increased since 1978. Since the cost of scrubbers already is included in the 1978 typical coal plant capital cost, it is unlikely that further pollution control improvements and design improvements would add more than 20 percent. If, indeed, actual nuclear construction costs are higher and actual coal plant construction costs are lower than the plant model results, the typical nuclear plant would be expected to produce more expensive electricity in all regions.

Low-cost nuclear plants, however, still would be competitive with the average coal plant. Com