WILDCAT: a catalyzed D-D tokamak reactor
xvi 1. ACKNOWLEDGMENTS Contributions to this report were made by P. A. Finn in the area of tritium handling and by Y. Gohar in the area of neutronics. E. Singleton provided coordination of the drafting services and significant input to the engineering design. Cyrilla Hytry was responsible for the typing and final assembly of the report. xiv --• ABSTRACT WILDCAT is a conceptual design of a catalyzed D-D, tokamak, commercial, fusion reactor. WILDCAT utilizes the beneficial features of no tritium
... ures of no tritium breeding, while not extrapolating unnecessarily from existing D-T designs. The reactor is larger and has higher magnetic fields and plasma pressures than typical D-T devices. It is more costly, but eliminates problems associated with tritium breeding and has tritium inventories and throughputs approximately two orders of magnitude less than typical D-T reactors. There are both a steady-state'version with Alfven-wave current drive and a pulsed version. Extensive comparison with D-T devices has been made, and cost and safety analyses have been included. All of the major reactor systems have been worked out to a level of detail appropriate to a complete, conceptual design. xvi '• 1 .. 1 As a consequence, in order to have a resonable power output for WILDCAT, it is necessary to increase the size, the toroidal field, and/or the plasma beta relative to values for, say, STARFIRE. These are the three parameters which most influence the power apart from the plasma temperatures and the reactivities. The choice for WILDCAT has been to extend each of these parameters somewhat from the STARFIRE values and to also produce less thermal power. In this case no individual parameter is unreasonably extrapolated beyond a value considered viable for STARFIRE. A schematic comparison of WILDCAT and STARFIRE is shown in Fig. 1-1 . The increase in size is readily apparent. The thicker coils are an indication of the higher field. It can also be seen that the space between the plasma and the peak field position of the toroidal coils has been reduced in order to make more effective use of the 1-3 1-6 Table 1-2. A Summary of the WILDCAT Reference Parameters Parameter Major radius, m Aspect ratio, A Peak toroidal field, T Plasma beta Average electron temperature, keV Plasma current, MA Plasma elongation Safety factor Edge Axis Neutron wall load, MW/m 2 14.06 MeV 2.45 MeV Net heat load, MW/m 2 1497 3077 3788 62.8 Pulsed 849 1885 3844 4528 73.8 Since WILDCAT does not have to breed tritium, the blanket/shield can be optimized to have a thinner inboard extent (82 cm vs. 120 cm for STARFIRE) leading to more efficient use of the toroidal field and to increased neutron energy multiplication (2.02 vs. 1.14). These benefits help to overcome the reduced reaction rates and lead to a 60Z more efficient blanket in terms of 1-7 power generation. In addition to enjoying increased blanket energy multiplication, the blanket/shield has been designed for personnel access after 24 h and uses as ouch as possible materials which are not resource limited and which have lower activation. Ninety percent of the material in WILDCAT can be recycled after 40 y. About one-half of the neutrons produced in WILDCAT are 14-MeV, D-T neutrons. Their blanket neutron energy multiplication is 1.54 and they contribute about 33% of the thermal power. Even though WILDCAT is termed a D-D reactor, the largest single source of power (38%) is, in fact, the neutrons and heating from the D-T reaction." -The 2.5-MeV, D-D neutrons have a higher energy multiplication of 4.43, but only contribute 19% of the thermal power. The remainder of the thermal power comes from fusion heating (45%) and rf heating (<4%). A further breakdown of the power production is given in Table 2-4. The numbers given are for the steady-state case, but the pulsed case is not essentially different apart from having no rf heating. The first wall is PCA stainless steel consisting of a corrugated plate bonded to a backing plate. There is a 3-mm beryllium cladding bonded to the corrugated part, which faces the plasma. Light-water coolant flows in the closed part of the corrugations. The configuration is shown in Fig. 3-1. The lifetime is estimated to be 20 y, or half of the expected plant life, and is limited primarily by sputtering loss of the beryllium cladding. The longer lifetime (compared to STARFIRE with a 6-y replacement schedule for first-wall/ blanket sections) is due to the lower neutron flux of the D-D fuel cycle for a fixed heat load on the wall. The heat load is 1 MW/m 2 for both STARFIRE and the steady-state version of WILDCAT. The pulsed version is limited to less than this value because of increased materials damage resulting from the pulsed loading. The WILDCAT steady-state burn cycle is characterized by long start-up and shut-down times to minimize power supply requirements and extra tritium and 3 He injection to provide heating during startup. The burn cycle starts with a 19-s "ohmic heating" period during which enough current (1 MA) is induced for the rf current drive to take over. This is followed by a 20-min "current inducement" period with low deuterium density and rf heating from the Alfven waves at 107 MW. After a "fusion power ramp" period of 19 min with extra tritium and 3 He, the plasma is brought to full operating conditions, and 1-8 Iodine is added for burn control. The burn then continues for typically up to 6 mo. Shutdown is similar to startup. The small amount of extra tritium and 3 He for startup is recovered and stored during the rest of the cycle. The pulsed verson burn cycle is described in Ref. 2. The startup and shutdown are necessarily faster, and all of the current is generated by the poloidal coils. The power supply requirements for the ohmic heating and equilibrium field system are substantially larger. In addition, t.iermal storage is required to keep the power to the turbine constant during the dwell period between cycles. The D-D reactor requires a substantially higher value of nx, but the confinement is still compatible with empirical scaling laws. Impurity control is via a pumped limiter. The heat loads on the limiter are somewhat higher than for STARFIRE, but a limiter similar to that for STARFIRE is expected to be adequate. Disruptions present a potential problem for WILDCAT. Because of the large amount of energy stored in the plasma (8.3 GJ vs 1.1 GJ for STARFIRE and 240 MJ for INTOR ), disruption scenarios whicn>are marginal for other devices become deleterious for WILDCAT, involving more melting and vaporization of the wall. No solutions to this problem have been identified for WILDCAT except to operate the plasma in a mode where disruptions do not occur, except perhaps as very low probability accidents. It is not unreasonable to expect our understanding of plasma behavior to be sufficiently advanced for this to be possible by the time that one would consider building WILDCAT, and quite likely similar requirements would be necessary for other than near-term devices in any event. A small number of disruptions should not be catastrophic. The high toroidal fields (14.35 T for the steady-state version and 14.0 T for the pulsed version) present problems primarily related to the stresses, which increase as the square of the field. The conductor design iself is similar to that for STARFIRE, utilizing various amounts of Nb 3 Sn in the regions with different field strengths. Substantially more material, however, is required. The out-of-plane loads are supported by filling essentially all of the space between the outer legs of the toroidal field coils with reinforced concrete, as shown in Fig. 1-2. There are three blocks (upper, Middle, and lower) between each coil. The Middle block has a plug for access to the limiter, which can be removed as a drawer-like unit, and the rest of the interior of the machine, especially the current-drive antennas. This support 1-9 concept is relatively inexpensive and appears to adequately handle the large forces. A detailed structural analysis has not been performed, however. A small ohmic heating system has been supplied for the steady-state case, and a larger, conventional ohmic heating system with a solenoid, for the pulsed case. The pulsed reactor dasign is substantially constricted by the need to supply a large number of volt-seconds (695). The plasma has been made less D-shaped to reduce the requirements on the equilibrium field system. For the pulsed version these two systems repesent a large, additional cost item. The real advantages of WILDCAT lie in not having to breed tritium and in reduced tritium inventories and throughputs. Both factors should lead to increased safety. It should be noted, however, that the higher magnetic fields in WILDCAT would probably result in increased magnet safety issues compared to STARFIRE. This study has not made an in-depth safety comparison of DT-fueled and alternate-fueled fusion reactors, nor could this be done at this time.' The benefits of not breeding tritium, including not having to deal with liquid lithium or not having to extract tritium from solid breeders, are also difficult to quantify at this time. It is, however, most likely that the ease with which tritium can be bred will determine the desirability of D-D reactors. The reduced tritium inventories and throughputs in WILDCAT (approximately two orders of magnitude less than for STARFIRE) are, however, a significant and quantifiable advantage. The vulnerable inventory is 15 g vs. 397 g for STARFIRE, and the nonvulnerable inventory is 20 g (33 g for the pulsed version) vs. 11,000 g for STARFIRE. The tritium throughput is 10 g/day vs. 760 g/day for STARFIRE. Normal releases of tritium are reduced from 13 Ci/day to 0.31 Ci/day, and accidental releases are reduced from 10 g to 0.56 g. In addition, no significant inventory of the more toxic HTO or T 2 0 has been identified. Additional savings lie in longer-lived vacuum pump valves (plant life vs. 2 y for STARFIRE) and lack of necessity for a ventilation stack. Even with the reduced inventories, there is still enough tritium present that no major tritium handling systems could be eliminated, and in view of the higher gas loads, the tritium/vacuum/fuel system is roughly the same size as for STARFIRE. The power flow diagrams for WILDCAT are shown in Fig. 1-3 . It has been determined that the turbine could have a high efficiency (35.7%, same as for STARFIRE), helped in part by using the lower-grade heat from the limiter as 1-10 The principal disadvantage of a D-D reactor is that the plasma power density is less than 2% of that of a tritium-fueled reactor. As a consequence, a D-D reactor must necessarily be substantially larger and/or operate at substantially higher fields or higher plasma betas than a D-T reactor of comparable thermal power. The design of larger devices and higher-field magnets is, of course, more difficult. Moreover, since the auxiliary systems, such as plasma heating, current drive, magnet power supplies, and vacuum pumping are then typically larger, the parasitic power losses represent a larger fraction of the thermal power, resulting in lower efficiency and even further reduced net electric power. The larger energy stored in the plasma is a more serious problem in the event of a plasma disruption. A second disadvantage is that a D-D reactor must likely operate at higher temperatures (25-30 keV compared to 8-10 keV for a typical D-T system). Cyclotron and bremsstrahlung radiation losses both increase with temperature. It is not known if diffusion losses increase or decrease with temperature in these temperature ranges, but there are models such as ripple diffusion and trapped particle modes which show loses increasing strongly with temperature. These factors affect the achievement of ignition. Using the assumptions made in the present study, ignition in a D-D reactor appears to require an orderof-magnitude larger confinement parameter, nx, and an order-of-magnitude fewer impurities compared to a D-T, reactor. A third feature of a D-D reactor is that a larger fraction of the power coming out of the plasma is in the form of heat (charged particles or radiation) rather than neutrons. If neutron damage of the first-wall/blanket/shield system were the limiting factor, this would be an advantage. For the type of design considered in this report, this becomes a disadvantage for a D-D reactor. STARFIRE, for example, supports a total wall load of 4.8 MW/m 2 with a heat load of 1.0 MW/m 2 , while the WILDCAT steady-state version supports a total wall load of only 1.7 MW/m 2 for the same heat load. The reduced neutron flux does lead to longer life for the first-wall/blanket, however. It is especially difficult to overcome the disadvantage of lower power production, and it would seem that D-D reactors would not be built for power production if it were possible to utilize D-^T reactors. If, however, D-T reactors (because of problems associated with tritium fueling and/or breeding or lifetime limitations due to neutron damage effects) are not feasible, then D-D 1-13 reactors could likely be built in their place with reasonable extrapolations of parameters considered adequate for D-T reactors.