American Nuclear Society

Home / Store / Journals / Electronic Articles / Fusion Science and Technology / Volume 8 / Number 3

A Nonlinear, Multivariable Method for Fusion Reactor Blanket Optimization

Wayne R. Meier, Edward C. Morse

Fusion Science and Technology

Volume 8 / Number 3 / November 1985 / Pages 2665-2680


Member Price:$27.00
Member Savings:$3.00

A method for optimizing the design of a fusion reactor blanket as a function of several design variables is described. Applications of the method are described elsewhere. The optimization problem consists of four key elements: a figure of merit (FOM) for the reactor, a technique for estimating the neutronic performance of the blanket as a function of the design variables, constraints on the design variables and neutronic performance, and a method for optimizing the FOM subject to the constraints. The FOM and constraints depend on the application and design objectives of the particular reactor concept. In general, they may be functions of the design variables and of the neutronic performance. A direct search, nonlinear simplex method is used to optimize the FOM subject to the constraints. The optimization algorithm requires the evaluation and comparison of the FOM at many different points in the search for the most attractive point. An evaluation of the neutronic performance is required each time a new point (i.e., a new set of design parameters) is chosen for comparison. The neutronic performance is evaluated by successive variational interpolation. With this approach, analytical expressions can be written for the neutronics performance as a function of the design variables based on only a limited number of reference point, neutron transport calculations. Hence, the FOM can be evaluated at any intermediate point without the need for additional transport calculations.

Questions or comments about the site? Contact the ANS Webmaster.