The nuclear power plants in service today were conceptually designed and developed during the 1960s. At that time, it was deemed necessary to achieve maximum efficiency and minimum cost in order to compete successfully with coal- or oil-burning plants. The latter were priced at 15% of their present cost and used fuel that was very cheap by current standards. In order to maximize efficiencies in the nuclear plants, temperatures, pressures, and power densities were pushed up to their highest practical limits. Safety features were exemplary for that era, and even for current safety practices in other industries. But they were not up to present-day demands for super-super safety in the nuclear industry.
As the public became more concerned with nuclear safety, the Nuclear Regulatory Commission required that new safety equipment and procedures be added on, in the process discussed in Chapter 9 as "regulatory ratcheting." The amount of labor and materials for these add-ons exceeded that for the plant as originally conceived. With this added complexity, the plants became difficult and expensive to construct, operate, and maintain. Moreover, the level of safety was still limited by the original conceptual design.
The New Design Philosophy1,2,3
By the early 1980s it became apparent that a new conceptual design of nuclear reactors was called for. The cost of electricity from coal- and oil-burning plants had escalated to the point where their competition did not require maximum efficiency from nuclear plants. Furthermore, the added efficiency achieved by pushing temperatures, pressures, and power densities to their limits was overshadowed by the efficiency lost due to shutdowns when these limits were exceeded. But above all it would be much easier to satisfy the public's demand for super-super safety by starting over with a new conceptual design than by using myriads of add-ons to a design originally targeted on rather different goals.
In the mid-1980s, several reactor vendors undertook these new designs. Let us consider some of the thinking that served as their basis.
In attempting to obtain maximum performance per unit of cost, designers nearly always find it advantageous to build plants with higher power output. This is the widely applicable principle of "economies of scale." For example, it is the reason why, before fuel costs became an issue, American automobiles were large in stamping out an auto body, machining an engine cylinder, or in any other such operation, it doesn't cost much more to make it large than to make it small. It thus costs substantially less to build and operate a 1,200,000-kW plant than two 600,000-kW plants of the same basic design. In line with the old goals, plants were therefore built in the former size range, and constantly grew larger. This trend still dominates in foreign countries.
However, if safety is the primary goal, as it is in the United States today, it is much easier to assure that adequate cooling will be available if there is only half as much heat to dissipate. In fact, in a 1,200,000-kW reactor, cooling requires elaborate pumps, while in a 600,000-kW reactor it can be handled by simple gravity flow with natural convection cool water enters the bottom of the reactor, which heats it, causing it to rise because warm water is lighter. This process sets up a natural circulation driven only by gravity. Unlike pumps which can fail and are driven by electric power which may not always be available, gravity never stops working. That makes the 600,000-kW reactor inherently safer. This is called "passive stability," since no active measures by operators or by mechanical or electrical control systems are required. The operator could shut off the electric power and go home without any harm coming to the reactor.
Natural circulation can also be used to protect the containment from breaking open due to excess pressure. In present-day power plants, active cooling using water pumps is necessary to control the pressure. But with the smaller reactor, there is less energy to dissipate, making natural circulation a viable alternative.
Of course the safety of the larger reactor system can always be improved by adding extra pumps and extra diesel-driven generators to provide power for them in the event of failures, but this drives up the cost. Moreover, this extra equipment must be periodically tested, maintained, and repaired. The control systems that start the equipment when it is needed, and open and close the necessary valves, must be maintained and frequently tested. As a result of problems like these, a super-super safe 1,200,000-kW reactor might cost almost as much as two super-super safe 600,000-kW reactors. Moreover, the 1,200,000-kW reactor is still not as safe because human failures can enter the picture for instance, a person can push the wrong button in starting a pump or opening a valve, or the person who installed an automatic system for performing these functions can make a mistake in wiring. In addition, with more pipes entering the reactor there are more opportunities for a pipe to break, causing a loss-of-coolant accident (LOCA).
We have seen that the time required for construction can be a very important ingredient in determining costs, and it obviously takes longer to construct a larger plant. This was not an important factor in the early 1970s, when large plants could be built in 4 or 5 years, as they still are today in France and Japan. But with the very elaborate and time-consuming practices now required in the United States, construction times have become an important factor in determining the relative costs of large and small reactor plants. For example, with elaborate tests and documentation for these tests required for every weld, reducing the number of required welds results in substantial time saving. Since they have fewer pumps and valves and less piping, smaller reactor plants require fewer welds, which reduces the time needed for constructing them.
Still another consideration is that many more parts of a smaller reactor than of a larger one can be produced and assembled in a factory. Operations are much more efficient in a factory than at a construction site because permanently installed equipment and long-term employees can be used to produce many copies of the same item for many different power plants. Use of transient construction workers and portable equipment to do a wide variety of jobs just once introduces all sorts of inefficiencies.
When all of these matters are taken into account, it turns out that if super-super safety is a primary concern, two 600,000-kW reactors may even be cheaper and are certainly considerably safer than a single 1,200,000-kW reactor.
In France and England, one government-owned utility provides electricity for the whole country. But in the United States there are hundreds of utilities, and they are generally rather small. Consider a typical American situation: a particular utility's service load is growing by 100,000-kW per year, and signs indicate that it will continue to grow at that rate for the next few years. A 600,000-kW plant might be an attractive investment to take care of this growth, but a 1,200,000-kW plant would be much more expensive and would provide a lot more power than needed for a long time. Moreover, it is hard to predict how long the growth will continue; if it should fall off, the utility may never need the full 1,200,000-kW. Thus a 600,000-kW size is better suited to the needs of an American utility. This is not the case for the huge French and British national utilities, which generally have a load growth of several million kilowatts per year.
Since the 600,000-kW size range is optimum for U.S. utilities, reactor vendors are competing to provide them. In fact, not only are U.S. reactor vendors developing them, but foreign vendors who build larger reactors for their own countries are developing 600,000-kW reactors to sell to American utilities. Coal- and oil-burning power plants of this size have always been available and are the most widely used plants in the United States.
Another aspect of the new design philosophy is to favor a larger tolerance to variation in operating conditions over optimizing efficiency. This represents an abandonment of the controlling principle of reactor design in the 1960s. For example, optimum plant efficiency requires heating the water to the highest possible temperature, but increased temperature accelerates corrosion of steam generator tubes. In the new design, water temperatures are reduced from 615°F to 600°F, and, in addition, a new stainless steel alloy that is more durable under exposure to high-temperature water is used for steam generator tubes. As another example, the reactor must be shut down for safety reasons if the power density the energy per second produced by each foot of fuel assembly exceeds a predetermined limit. Present reactors operate at 5.5 kW per foot, but in the new reactors this is reduced to 4.0 KW per foot. The change reduces efficiency, but it means fewer shutdowns due to excessive power density.
Another change in the new design is to favor simplicity over complexity. Adding complex equipment and operations may improve the efficiency of a reactor, but it also introduces more possibilities for equipment failures and operator mistakes. In earlier reactors, living with these problems was deemed to be a worthwhile sacrifice, but the new design philosophy is that it is not.
In addition to changes in design philosophy, there have been many lessons learned from experience that can be incorporated into the new reactors. These include systems for adding small quantities of various chemicals to the water to reduce corrosion problems, employing different materials for certain applications, and using new methods for construction. The principal lesson learned from the Three Mile Island accident is the significance of making important information easily available to the operator. The revolution in computer and telecommunications technology over the last two decades offers bountiful opportunities for improvements here in revamping the control room and instrumentation systems, and providing graphic displays, diagnostic aids, and expert systems to handle a variety of situations. Because of the passive stability features, there are few situations which require the operator to take emergency action. This also makes the reactor easier to operate.
Licensing of reactors has become a major problem for U.S. utilities. They must obtain a construction permit to start building the plant, and when the plant is nearing completion they must apply separately for an operating license. These are long, tedious, expensive processes that sometimes cause disastrous delays. In Chapter 9 we mentioned that the Shoreham and Seabrook plants encountered delays of several years, costing over a billion dollars for each, while the utilities struggled against political opposition to obtain an operating license. To avoid this sort of fiasco in future plants, the NRC is developing new rules under which both a construction permit and an operating license are obtained in a single procedure before construction is started. When the plant is completed, the utility need only show that it was constructed in accordance with the plans originally approved. It remains to be seen whether Congress will go along with these new rules. The interaction between the NRC and Congress is a matter of considerable complexity, and utilities are understandably wary of being caught in the middle.
Even more important in the licensing problem is the matter of standardization. Almost every U.S. reactor is a custom design and custom built. In the late 1960s and early 1970s, this seemed appropriate because lots of new ideas were available for trial, and designs were changing rapidly on the basis of earlier experiences. By the mid-1970s, things were settling down, and American reactor vendors began developing standard designs, but that was just when the orders for new reactors stopped. In France, however, reactor construction was just beginning to burgeon at that time, and the French government-owned vendor adopted the standardized design approach. This proved to be very efficient and successful, resulting in high-quality plants built quickly and cheaply. Construction crews, after having built several identical plants, knew exactly what to do, where to look for trouble spots, and how to correct them. Once a new reactor went into service, the operating crews were experienced from the start, since most of their members had done the same job before in identical plants.
For the United States, with its current tumultuous licensing procedures, standardization would appear to be a panacea. A reactor vendor would go through the laborious and expensive process of getting its standard design certified by the NRC. But once this was accomplished, a utility would need only prove that its site is suitable, a very much simpler procedure. Once this was done, they would have clear sailing with no delay before going into commercial operation. Since all equipment and procedures would be standardized, there would be much less that could go wrong.
The Next Generation4
In the mid-1980s, Electric Power Research Institute (EPRI), the private research arm of the electric utility industry, undertook to set design goals for new reactors that make use of currently available technology and incorporate the philosophy expounded above. That project developed the following specifications:1,5
A mammoth effort both in the United States and abroad is now underway to develop reactors to these specifications. The principal designs that have emerged include a pressurized water reactor called AP-600 (for advanced passive 600,000 kW), developed by Westinghouse and partner companies6; a boiling water reactor called SBWR (where S stands for simpler, smaller, safer), developed by General Electric, Bechtel, and the Massachusetts Institute of Technology; a liquid-metal-cooled reactor called PRISM (for power reactor inherently safe module), developed by General Electric in collaboration with Argonne National Laboratory; and a gas-cooled reactor called MHTGR (modular high-temperature gas-cooled reactor), developed by GA Technologies plus a consortium of other companies. Utilities from France, Italy, the Netherlands, Japan, Korea, and Taiwan have also participated with these companies in developing the designs.
In addition, large programs for developing advanced reactor designs have emerged in France, Britain, Germany, and Japan, although these have been on larger reactors without the passive stability features. A Swedish effort7,8 is developing the PIUS (process inherent ultimate safety) reactor, a 640,000-kW PWR with essentially no dependence on active equipment.
In the United States, the AP-600 and the SBWR are expected to pass the licensing requirements and be ready for commercial orders by 1995, but the PRISM and MHTGR will not reach that stage until after the turn of the century.
For the interim of the next few years, American reactor vendors have developed improved versions of the reactors now in service, with several changes targeted on reducing complexity and improving safety. They also expect to have a long-term market for these reactors, called the APWR and ABWR (A for advanced), in foreign countries. One of them has already been ordered for construction in Japan. The modifications reduce the probability of a severe accident tenfold, have increased margins for error in key areas, and give operators more time to respond to emergency situations. These reactors should be certified for licensing in the United States by 1992.
Let us examine the salient features of some of the next-generation reactors.
The AP-600 obtains its emergency cooling from huge water tanks mounted above the reactor. Some of these are pressurized with nitrogen gas, allowing them to inject water even if the reactor remains at high pressure, as it may in some accident scenarios. In most cases, neither electric power nor operator action are needed to start injection. For example, if the pressure in the reactor falls due to a break in the system, the valves connecting some of the tanks to the reactor are automatically pushed open by the fact that the pressure on the tank side is being higher than on the reactor side. Actually, present reactors have similar systems. The "accumulators" mentioned in Chapter 6 operate on the same principle, but with enough water for only about 15 minutes, as compared with several hours in the AP-600.
One of the large tanks above the reactor serves as a place to deposit heat. It is connected to the reactor by two pipes, one leading to the bottom of the reactor vessel and one to the top. As water in the vessel is heated, it automatically rises in the upper pipe and is replaced by cool water from the lower pipe, establishing a natural circulation which transfers heat from the reactor into the water tank. This is analogous to air heated by a radiator in a house rising (because it is lighter) and spreading through the room to transfer the heat from the radiator to all the air in the room.
Water tanks pressurized with nitrogen gas also provide sprays to cool the atmosphere inside the containment and remove some of the volatile radioactive materials from the air in the event of an accident. Again, no pumps are needed.
The steel containment shell is cooled by air circulating between it and the concrete walls, again by gravity-induced convection. In addition there is water draining by gravity onto the containment shell, though the air circulation alone is sufficient to provide the necessary cooling. Air circulation keeps the pressure inside below 40 pounds per square inch in the worst accident scenarios.
Probabilistic risk analyses yield estimates that a core damage accident can be expected only once in 800,000 years of reactor operation, and that there is less than a 1% chance that this will be followed by failure of the containment. This makes the AP-600 a thousand times safer than the current generation of reactors. It is also much simpler, reducing the number of valves by 60%, large pumps by 50%, piping by 60%, heat exchangers by 50%, ducting by 35%, and control cables by 80%. The volume of buildings required to have a very high degree of earthquake resistance is thereby reduced by 60%. It is estimated that the plant can be constructed in 3 to 4 years. All of these factors contribute to reducing the cost.
Other Next-Generation Reactors4,9
In a pressurized water reactor (PWR) the water is only heated, and it can therefore flow directly from the inlet to the outlet, passing through the core only once. But in a boiling water reactor, (BWR), all the water cannot suddenly turn into steam as it passes through the reactor core; boiling occurs, releasing bubbles of steam which rise to the top of the reactor vessel, but most of the water remains in liquid form. After passing through the core it must therefore be recirculated again and again with a fraction converted to steam on each pass. This recirculation is now done with pumps outside the reactor, but in the SBWR it will be done by natural circulation entirely inside the reactor.
The modification of the emergency core cooling system in the SBWR to achieve passive operation by natural convection is similar in principle to what is described above for the AP-600 PWR. In addition, new stainless steel alloys, advanced welding techniques, and improved water chemistry are used to eliminate stress corrosion cracking experienced in present BWRs. Reductions in the amount of piping, valves, and the like, comparable to those in the AP-600 are achieved to reduce complexity and cost. The SBWR should also be ready for orders by utilities in the mid-1990s.
The PRISM uses liquid sodium metal rather than water to transfer the heat. This has the advantage that the system does not operate under high pressure and therefore cannot burst open, which is a plus for safety. The disadvantage is that sodium burns in air or in water, releasing large quantities of radioactive materials if it gets out of the reactor. But there is now a great deal of technology and experience with safe handling of sodium.
The PRISM would be built in small modules producing 155,000 kW each. If the electric power should fail, the operators could leave without harm to the reactor. Its heat would be radiated to the containment shell, which would be cooled by natural air circulation. The reactor and even the containment vessel can be fabricated in a factory and shipped by railroad, making installation at the site very quick and easy.
Part of the reason for development of the PRISM is that it provides advantages in fuel fabrication, reprocessing of spent fuel, waste management, and protection against theft of plutonium. This design concept has a great deal of potential from many points of view, but a great deal of development and testing will be required before it can be evaluated in its totality.
The MHTGR is also modular and even smaller at 135,000 kW than the PRISM. The heat is transferred out of the reactor with helium gas, which allows use of much higher temperatures and thereby gives higher efficiency. The fuel consists of small particles of uranium plus thorium (which is converted to fissile uranium in the reactor), sealed inside very high melting point materials and encased in huge quantities of graphite, which can absorb tremendous amounts of heat. At any temperature above 1000°F, it radiates this heat to the surroundings much faster than it can be generated, leaving no possibility for the fuel to even approach the 4000°F temperature where it might be damaged. It is therefore believed that this reactor will not even require a containment. The modular character allows factory fabrication and quick, cheap construction.
Since they incorporate so many innovative features, prototypes will have to be constructed and tested before either the PRISM or the MHTGR can become candidates for purchase by utilities. This is not true of the AP-600 or the SBWR, because they represent relatively minor departures from well-established technologies. We can expect orders for these plants in the mid-1990s.
Cost per Kilowatt Hour
Quantities of electrical energy are expressed in kilowatt-hours. For example, our homes are billed for electricity use at a certain cost per kilowatt-hour, currently ranging from 8 cents to 12 cents. However, that billing includes distribution costs bringing the electricity from the power plant into our homes and these costs are substantial. Here we consider only the cost per kilowatt-hour as it leaves the power plant.
The history of declining electricity costs through the first three-quarters of this century is one of the great success stories of technology. In the 1930s, electricity was an expensive commodity. You never left a room without turning off the light, and keeping a refrigerator door open longer than necessary was a serious transgression. But as time passed, the cost per kW-h dropped steadily, even while the price of everything else was doubling and redoubling. By the early 1970s, it had dropped to 1.5 cents per kilowatt-hour, and electricity was being used profligately in all sorts of applications. With the energy crisis of 1974 raising the price of fuel, and the general inflation following it, the cost of electric power began to rise. By 1981, the average costs per kilowatt-hour were 2.7 cents for nuclear, 3.2 cents for coal, and 6.9 cents for oil.10 The low average price for nuclear power was due to the low-cost plants completed before 1975, for which the average cost was 2.2 cents versus 3.5 cents for those completed after 1975. As more new nuclear plants were completed, the average cost of nuclear power increased. A survey by the Atomic Industrial Forum found the costs in 1985 to be 4.3 cents for nuclear, 3.4 cents for coal, and 7.3 cents for oil.
The U.S. Energy Information Agency does not try to calculate averages but rather reports on the range of prices for all large plants. In 1984 it gave 1.52 to 8.17 cents for nuclear versus 1.86 to 6.41 cents for coal, and in 1987 it was 1.68 to 8.50 cents for nuclear versus 1.82 to 7.98 cents for coal.11These very wide ranges are indicative of how little meaning there is to averages over plants now in operation. For example, they depend very much on the mix of old and new plants in use, since there is no inflation factor applied to the original cost of the plant.
More germane to our interests here are the estimated costs of electricity from plants to be constructed in the future, since this is what is important in a utility's choice of a new plant. In Table 1, we break down the cost per kilowatt-hour for various candidates.12 The cost of operating and maintaining a facility, and of providing it with fuel are easy to understand. The contribution of the capital used in constructing the plant is calculated as the amount of money per kilowatt-hour needed to purchase an annuity which will pay off all capital costs and interest before the end of the facility's useful life. The decommissioning charge is the amount of money per kilowatt-hour that must be set aside so that, including the interest it has earned, there will be enough money to pay for decommissioning the plant when it is retired. There are many uncertainties in these calculations, but both the U.S. Government and the utilities employ groups of experts who have developed standard procedures that are normally applied and widely accepted. A utility gambles billions of dollars on decisions based on these analyses, so I would think they do everything possible to make them as reliable as possible.
The first two entries in Table 1 are data for present-generation reactors based on median experience and best experience as defined and used in Chapter 9. The APWR (advanced pressurized water reactor) refers to present-type reactors with various upgrades to improve safety, a standardized design, and the streamlined licensing procedures discussed above. The AP-600 represents the new generation of PWRs and BWRs that fully utilize passive safety features as we have described. Both the AP-600 and the coal-burning plant are of half the capacity of the present generation reactors and the APWR 600,000 kW versus 1,200,000 kW and Table 1 therefore considers two of these plants at the same site in drawing comparisons. Note that having two plants rather than one increases the operating, maintenance, and decommissioning costs and makes electricity from the AP-600 slightly more expensive than from the APWR. That is part of the price we pay for super-super safety, but it is hardly significant.
COST PER KILOWATT-HOUR FOR VARIOUS TYPES OF POWER PLANTS.
*Figures are in mills (0.1 cents) of 1987 dollars, for plants going into operation in the year 2000.
The lower capital cost of the APWR than of the "best experience" present-generation reactors is based on applying lessons learned from the latter, benefits from standardization and easier licensing, and freedom from regulatory turbulence. Actually the only large difference in Table 1 is the gap between the median experience and best experience present-generation reactors. While the costs of the others are similar, the APWR is about 10 times safer, and the AP-600 is 1,000 times safer than the present-generation reactors.
The most important point in Table 1 is that electricity from the new-generation reactors is about 20% cheaper than from a coal-burning plant. The coal-burning plant is cheaper to build, operate, maintain, and decommission, but its fuel cost is more than 3 times higher.
The fact that future nuclear power plants will produce electricity at a substantially lower price than coal-burning plants is confirmed by a more elaborate study by industry analysts.13 Their study also concludes that the cost of electricity from future nuclear plants will be about 20% cheaper than from coal-burning plants. It explores the sensitivity of this conclusion to a wide range of uncertain parameters used in the calculations, but always finds nuclear power to be less expensive.
Of course, this conclusion depends heavily on the assumptions of standardized design, streamlined licensing procedures, and no regulatory turbulence. These would reduce construction times to 5-6 years, according to the Department of Energy analysts who developed Table 1, and that fact alone would be responsible for a large fraction of the cost saving. Industry planners are hoping for even shorter construction times, especially after a few of these standardized plants have been built.
In the past, many other analyses have been reported on cost comparisons between nuclear and coal burning, always based on the assumption of reasonable (i.e., non-turbulent) construction practices. This is clearly a very practical question for a utility deciding to build a new power plant. Many utilities seek cost analyses from economics consulting firms, some utilities have their own in-house economists to make estimates, and banking organizations maintain expertise to aid in decisions on investments. From the early 1970s until the early 1980s, all of their reports found that nuclear power was the cheaper of the two. For example, the Tennessee Valley Authority (TVA) is the largest electric utility in the United States. Its profits, if any, are turned back to the U.S. Treasury. It maintained a large and active effort for many years in analyzing the relative cost advantages of nuclear versus coal-burning power plants, consistently finding that nuclear power was cheaper. The 1982 analysis by the Energy Information Administration, a branch of the U.S. Department of Energy, was the first to find that coal and nuclear were equal in cost; their previous analyses found nuclear to be cheaper. By 1982, these analyses were mostly discontinued, since it seemed unrealistic for a utility to consider building a nuclear power station, or even to hope that it could be done without regulatory turbulence. Only with the new optimism about the future of nuclear power have these analyses been resumed.
There have been claims that utilities are biased toward nuclear and against coal because they can make more profits from the former. This is hard to understand in view of the fact that many utilities have been badly hurt financially by their nuclear ventures. It would not explain the positions of nonprofit utilities like TVA or municipally owned power authorities. Most Western European power plants are built and owned by the government, with no possible profit motive in favoring the nuclear option, and of course the same is true of Eastern European nations, the Soviet Union, and China, all of whom are actively pursuing nuclear power. The last two, at least, also have large reserves of coal.
In Western Europe, where there has been no regulatory turbulence comparable to that in the United States, most nuclear power plants have been relatively cheap. A 1982 study by the European Economic Community14 estimated that for projects started at that time, electricity from nuclear plants would be less costly than that from coal-fired plants by a factor of 1.7 in France, 1.7 in Italy, 1.3 in Belgium, and 1.4 in Germany. These results are based on only a very slow rise in the price of coal, and do not include scrubbers for removal of sulfur. Scrubbers are usually required on new coal-burning plants in the United States and add very substantially to their cost.
With safety problems hopefully behind us, and with cost considerations looking favorable, it truly seems like the United States is now ready for Nuclear Power: Act II.