Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
TECHNOLOGICAL CONSIDERATIONS OF D-T FUSION In this discussion the technological requirements for fusion power will be divided into three principal areas of concern: (l) the power balance, that is, the unique power-handling requirements associated with the production of electrical power by fusion; (2) reactor design, focusing primarily on the requirements imposed by a tritium-based fuel cycle, thermal-hydraulic considerations, and magnet systems; and (3) materials considerations, including surface erosion, radiation effects, and materials compatibility. The requirements discussed are based upon specific concepts, using current physics and engineering assumptions, the intent is to point out directions for improved or new plasma- confinement and engineering concepts. Therefore, the following discus- sions should be considered only as illustrative of the types of problems fusion will have to overcome or circumvent in the future. THE POWER BALANCE In contrast to fission reactors, fusion reactors will inherently require input power to establish the fuel conditions necessary for appreciable nuclear power release. Thus, fusion reactors can be viewed as energy- amplifying devices. The energy amplification achieved in a fusion reac- tor depends on both the plasma performance and the blanket design. The power-producing potential of a fusion reactor depends on the magnitude of the amplification, the efficiencies of the power-handling systems and the allowable capital cost of the power-handling systems. The technolog- ical requirements associated with power handling in a fusion reactor depend on the confinement scheme. The Tokamak Reactor Concept The Tokamak reactor operates in an ignited, quasi-steady-state mode. That is, plasma fueling and spent-fuel removal are required during cyclic burning phases. In current design studies it is generally assumed that: (a) fueling would be accomplished by injecting solid fuel pellets (with negligible power requirements) into the plasma; and (b) spent-fuel removal would be accomplished by guiding charged particles l7
l8 out of the plasma chamber along diverted magnetic fuel lines generated by coils called "divertor" coils. The feasibility of pellet fueling in Tokamak plasmas is currently being investigated at ORNL and GA. Stel- lerator confinement schemes have operated with divertors and experiments are now being constructed to investigate the behavior of Tokamaks with divertors, e.g., the PDX experiment at PPPL. Preliminary results ob- tained on a small Tokamak divertor experiment (DITE) at Culham, England, have been encouraging. In the Tokamak, a changing magnetic flux induces an axial current in the plasma to provide: (a) a pulsed poloidal magnet field that works together with a steady-state toroidal field to confine the plasma; and (b) initial plasma heating that arises from the associated ohmic heating within the plasma. The power to establish and drive the axial current could be delivered by an energy storage and transfer system or by the combination of an energy storage system and a power supply taking power from the line. The energy required to establish the axial current in a Tokamak reactor will be in the vicinity of a few GJ (l GJ = l billion joules), and the power requirements may approach l000 MW at the peak power level. The TFTR at Princeton will employ motor-generator-flywheel sets to drive the axial current. Inertial energy storage could also be used in reactors, but inductive energy storage and homopolar generators are also being investigated for Tokamak applications. It is generally assumed that the intrinsic ohmic heating process in Tokamaks will not provide sufficient heating to bring the plasma up to the ignition temperature and therefore auxiliary heating will be required to achieve ignition. A number of auxiliary techniques are cur- rently being investigated, including neutral beam injection and radio- frequency (RF) heating. At present, neutral beam injection is considered the most promising method for reactor applications; however, more work needs to be done in the area of RF heating. For reactor applications it appears that injection powers between 50 and l00 MW will be required, with injection energies in the range of l00 keV to 500 keV in order to achieve sufficient beam penetration to heat the plasma center preferen- tially. The neutral beams will be on for about 5-l0 seconds every burn- ing cycle. In order to achieve good beam-system efficiencies it will be necessary to develop schemes based on negative ion acceleration and neu- tralization. Current Tokamak and mirror experiments employ neutral beam systems based on positive ion acceleration and neutralization. The production and acceleration of negative ions at high current density represents a new development area for fusion technology. The power-producing potential of the Tokamak reactor depends on the fueling power requirements and on the achievable duty factor of the burn cycle. When the fueling power requirements are low (about l percent of the total nuclear power release) and the cycle duty factor is high (about 90 percent), the overall plant efficiency approaches that of the thermal converter, and the economic constraints on the power-handling systems are modest. It appears that it will be difficult to achieve Tokamak downtimes of less than about l minute for reactors. Therefore, it would also appear that burn times of the order of about l0 minutes must be achieved in order to get acceptable economics for Tokamak reac- tors. The length of the burning phase will be limited by either: (a)
l9 plasma-quenching resulting from enhanced plasma radiation associated with impurity buildup within the plasma; or (b) the available magnetic flux through the center of the torus, which determines the duration of the axial current. The magnetic flux limitation would easily allow burn times of l0 minutes and greater. Therefore, impurity control is essen- tial to the economic viability of Tokamaks. It should be noted that with sufficient impurity control, it is in principle possible to operate a Tokamak in steady state, using beam injection and the bootstrap effect (the self-heating of plasma by alpha particles) to sustain the axial current. Early Tokamak reactor studies were usually based on systems producing ~ l000-2000 MWe. It has been mistakenly assumed by many people outside of the fusion community that these large output powers are required for Tokamaks. The electrical output of a Tokamak should, at this point, be considered a design variable whose limits have yet to be set by physics experiments and then optimized. There is no evidence indicating that the power of the Tokamak must be in excess of a l000 MWe in order for Tokamaks to be economical power systems. The Mirror Reactor Concept The mirror reactor operates in a driven steady-state mode. That is, continuous energy input is required to sustain the plasma burn, because the mirror plasma does not ignite. The fusion power density is main- tained at a steady state by continuously feeding fuel into the reacting plasma and by continuously removing spent fuel. Note that, in the mirror concept, plasma end losses provide the advantage of an inherent spent-fuel-removal mechanism. The energy required to start up and sus- tain the plasma burn would be provided by neutral beam injectors that would also serve as the fuel source in reactors. For mirror reactors operating with classical confinement (Q* ~ l), the injected power re- quirements are comparable to the fusion thermal output of the device. Thus, a reactor releasing about l000 megawatts of fusion power would require approximately l000 megawatts of injected beam power. If end- stoppering research is successful and higher Q values become possible, the injected power requirements would be significantly reduced. The power associated with fusion neutrons appears as heat in the blanket and converts to electricity by means of a thermal energy- conversion system. The power leaving the plasma in the form of ener- getic charged particles (this includes essentially all the power associated with the injected neutral beams) is fed to a direct energy- conversion system based on electrostatic concepts. The mirror operating with a Q ~ l requires that direct energy conversion be employed in order to achieve an acceptable overall power balance. The average energy of the injected particles would have to be in the range of about 300-600 Fusion Power Release *O = Power Input Required
20 keV. It is emphasized that in a mirror reactor the neutral beam injectors must operate continuously during the burn as opposed to a Tokamak reactor in which the beams must be on only until ignition is achieved. The power-producing potential of mirrors operating with classical confinement (that is, Q ~ l) depends on the performance of the neutral beam injection and direct energy conversion systems. Even for relatively optimistic assumptions concerning the efficiencies of these systems, the overall plant efficiency of the classical mirror system is relatively low and the allowable capital costs for the power-handling systems appear to be extremely stringent. The power balance performance of mirror- based systems would be significantly enhanced if higher Q could be achieved. The enhancement of Q in mirror-based devices has become a major objective of the mirror program and is being pursued vigorously both at the Lawrence Livermore Laboratory and at the Oak Ridge National Laboratory. The Inertial-Confinement Reactor Concept In this concept, plasma heating energy in the form of laser beams, elec- tron beams, or ion beams is delivered to a fuel pellet in about l nano- second (l0~9s) within a cavity surrounded by a blanket. For a given fusion energy release per pellet microexplosion, the system output power depends on the achievable cavity pulse rate and on the number of reactor cavities. Cavity pulse rates in the range l shot per l0 sec per cavity to l00 shots per sec per cavity have been considered in reactor studies. The allowable pulse rate will be determined, to a large extent, by the time required to reestablish the necessary cavity environment permitting subsequent pellet injection and efficient beam penetration following each microexplosion. The number of reactor cavities that a single beam system can serve will be determined by pulse-rate capabilities and optical considerations. The power-producing potential of inertial confinement reactors depends on the performance of the beam system. For example, current estimates indicate that attractive power production, assuming a pellet gain of l00, will require laser systems with an energy output of ~ l MJ, an efficiency approaching l0 percent, and a capital cost of ~ $200/J of laser energy output. The performance of available laser systems is currently far below that necessary for reactor applications, and new laser media may have to be identified to achieve acceptable performance. Also note that the energy storage equipment of the beam system must satisfy very stringent design requirements with regard to energy transfer times (~ 3 to 5 psec) and repetition rates (~ l00 million to l billion pulses/year). REACTOR DESIGN REQUIREMENTS A tritium-based power economy is not feasible unless the rate of tritium production in the blanket exceeds its rate of consumption in the plasma. The requirement that the blanket breed tritium implies that lithium in
2l some form must be present in the blanket. The tritium-breeding perfor- mance of the blanket must be such that the tritium doubling time is con- sistent with electrical-energy growth patterns, around 7 to l0 years by present standards. It appears that the required tritium-breeding per- formance can be satisfied by a variety of blanket configurations and material choices. It is also noted that the specific power (MW/kg of tritium inventory) anticipated in a fusion breeder reactor will be greater than the specific power (MW/kg of plutonium inventory) in fission breeder reactors. Therefore, the fuel-doubling times of interest require significantly lower breeding ratios with fusion reactors than with fission reactors. The design of the blanket tritium recovery system is intimately re- lated to the choice of materials, blanket cooling system, power- conversion system, and tritium-containment technology. The types of technology required for the blanket tritium recovery systems under cur- rent consideration are to a large extent available. However, specific schemes will require considerable development and demonstration. In addition there is need for data concerning: (a) permeability, with and without diffusion barriers, at low tritium partial pressures (in the range of l0~6 torr and lower); (b) diffusion coefficients for tritium in proposed breeding materials and coolants at low concentration (in the range of ~ l0 ppm and lower); and (c) equilibrium information on tritium- lithium systems and tritium-metal systems over a wide temperature range. The technological requirements associated with the recovery of tritium from the plasma exhaust seem less formidable than those associated with the recovery of bred tritium. On the other hand, the plasma-exhaust tritium-recovery system will be required for tritium-burning experiments that do not employ a breeding blanket. Thus, demonstration of a plasma- exhaust tritium-recovery system is a nearer term objective than demon- stration of a blanket tritium-recovering system. DOE is currently reviewing proposals for a tritium systems test facility, the major objective of which will be to demonstrate the plasma-exhaust tritium- recovering system for fusion devices. Energy deposition at the first wall of the blanket will result in design limitations based on thermal stress considerations for all reactor concepts. The potential effects of energy deposition at the first wall appear to be most severe in the case of inertial confinement. In these concepts, the very short time scales for energy deposition would result in ablation of an unprotected metal first wall and in severe thermal cycling effects. Several alternative cavity first wall designs are being considered to ameliorate this problem. Both mirror and Tokamak reactors will employ steady-state super- conducting magnet systems for plasma confinement. The stored energy associated with these magnet systems falls in the range of about l00,000 to 200,000 joules per kWe. If capital cost allotments of around $200 per kWe are allowed for the steady-state magnet system, then the magnet cost must be in the range of about l to 2 mills per joule of stored energy. This allowable cost is in close agreement with projected costs for large superconducting magnet systems. It appears that force con- tainment and mechanical design will be the limiting factors in the design of superconducting magnet systems for mirror and Tokamak reactors.
22 However, protection of the magnets in the event of a sudden plasma quenching also represents a major area of concern. DOE is currently involved in a superconducting magnet development program aimed at devel- oping large coils for Tokamak and mirror systems. MATERIALS CONSIDERATION Surface erosion associated with plasma-particle bombardment can signifi- cantly limit the useful lifetime of components in fusion reactors. Moreover, in Tokamak plasmas the presence of impurities arising from first-wall erosion can severely affect plasma performance. It appears that the erosion processes of potential concern in fusion reactors will be plasma-particle sputtering and exfoliation resulting from the bursting of radiation-induced blisters. Currently, these phenomena are poorly understood in the context of a fusion reactor environment and, therefore, it is not possible to calculate accurate surface erosion rates in fusion reactors at present. A number of schemes are being pursued to protect both the plasma and the first wall from the consequences of surface erosion. Fundamental studies on sputtering and blistering are being conducted within the fusion program. Neutron-induced atomic displacements and gas production can lead to deleterious changes in the structural properties of materials, and thereby limit the service life of structural components in fusion re- actors. These radiation effects are expected to be most severe at high temperatures and in the vincinity of the blanket first wall. Calcula- tions indicate that at the first wall (a) annual atomic displacement rates would be lower than those achieved in high-flux fission reactors, and (b) annual gas production rates would, with the exception of nickel- bearing materials, be substantially higher than those achieved in high- flux thermal fission reactors. Currently, intense sources of fusion- energy neutrons [~ lO^n/(cm^sec) ] are not available; therefore, materials testing for fusion reactors must rely heavily on fission neutron irradi- ations and ion bombardments to simulate the fusion reactor radiation environment. Recent data obtained on the radiation performance of stainless steel under simulated fusion conditions suggest that the structural lifetime of the blanket first wall may be substantially in- creased (about l0-20 megawatt years per square meter) beyond that esti- mated several years ago (about 2 megawatt years per square meter) providing the operating temperature is lowered to 450°C or less. The instantaneous displacement and gas-production rates during the plasma burn of inertial confinement reactor concepts are about six orders of magnitude greater than those associated with the mirror and Tokamak reactor concepts. The higher displacement rates represent a completely different damage regime than we are used to dealing with in fission reactors. Very little is known about the effects of enhanced recombin- ation of defects under those conditions, but preliminary theoretical estimates reveal that such effects may even reduce the residual damage in metals at high temperature. However, there is no significant exper- imental effort being conducted at this time to substantiate these
23 predictions. The rate dependence of the radiation effects must be investigated. At present, it appears that this rate dependence can be examined only by ion-bombardment techniques. In general, the theoretical understanding of corrosion phenomena and kinetics is inadequately developed. Therefore, an extensive experimental program will be required to define the corrosion behavior of potential structural materials in lithium, lithium salts, and helium under the operating conditions and radiation environment expected in fusion reactors.
