Optimization of Process Variables for Co-decontamination Process Using Genetic Algorithms
The process variables of the co-decontamination processes, such as flow rates and feed concentrations of elements involved, were optimized by using Genetic Algorithms (GAs). The concentration profiles and the objective function for a given flowsheet were calculated with the SOLE code. Because the mass balance of elements involved in a contactor is expressed in nonlinear simultaneous equations, derivatives of the objective function cannot be calculated explicitly. On the other hand, GAs require no derivatives of objective function and can be well applied to such a problem of optimization.<BR>In the present study, two types of spent fuel for the optimization were considered. One was the spent fuel of LWR, and the other was that of FBR. In the latter case, because the concern for criticality and third phase formation becomes more important, a penalty term, which was a function of the maximum Pu concentration in aqueous phase, was incorporated into the objective function. In both the cases, optimal values obtained were well within the range of practicability and the average fitness of a generation increased steadily as the generation proceeded, which means that GAs are appropriate methods for these problems of optimization.
- Journal of nuclear science and technology
Journal of nuclear science and technology 32(12), 1243-1249, 1995-12-25
Atomic Energy Society of Japan