Meta-analysis of Tokamak Reactor Designs

The long-term interest in nuclear fusion using tokamaks has yielded many published reactor designs. This study performs the first meta-analysis of tokamak reactor designs in an attempt to unveil new understanding not available in the traditional bottom-up method of looking at each design individually. Forty tokamak designs intended to produce significant fusion power at gains above one were identified in the published literature. Thirty-three important parameters describing a tokamak design were compiled by examining the set. The parameters from each design were extracted and placed into a database and plotted against each other to identify trends and outliers. Major outliers include two low aspect ratio designs and two He-3 reactor designs. Two classes are apparent in the database indicating two design philosophies: large major radius (-7 m), high power (1.8 GW), and low density (I * 1020 m-3) designs utilizing superconducting magnets; and small major radius (-2.5 m), low power (-0.2 GW), and high density (~4* 1020 m-3) designs utilizing copper magnets. The former class has longer confinement times, higher plasma current, and lower magnetic field while the latter class tends to have lower gain, higher power per surface area, higher power per volume, and much smaller stored magnetic energies. Between the two sets, the non-dimensional plasma physics parameters are similar. These two basic design strategies have been in practice for the last 40 years. Since tokamak designs were first published, there has been little appreciable change in the mean and design envelope of the major parameters such as major radius, fusion power, magnetic field, and plasma current and of the plasma physics parameters such as beta, safety factor, temperature, density, and confinement time. The lack of significant change suggests that no major technological or physics breakthrough that could radically affect design philosophy has been discovered, and neither design philosophy has dominated. Trade-offs in triple product are apparent as all designs are at similar plasma temperature except He-3 reactors, while confinement time and density vary inversely among designs. The major dependencies on plasma current and size in the experimental confinement time scaling are also apparent. The conservatism inherent in reactor designs can be inferred from plasma physics parameters such as the confinement enhancement factor, normalized beta, and safety factor. The database indicates designers push all plasma physics limits simultaneously instead of individually. Thesis Supervisor: Dennis Whyte Title: Professor of Nuclear Science and Engineering