Parameters Optimization of a Counter-Current Cascade Based on Using a Real Coded Genetic Algorithm

Abstract

Analysis of economical aspects of centrifuge-based separation shows that the bulk of the cost is proportional to the number of centrifuges in a cascade. In this paper a program which is called MAKNO is used to obtain the velocity field in a centrifuge by solving popular purely axial flow in a gas centrifuge. Through using MAKNO and solving concentration equation in a single gas centrifuge a realistic function for α, separation factor, in relation to θ, cut, and f, feed flow rate is achieved. To minimize the number of centrifuges in a cascade one needs to determine optimal parameters, α, θ, and f. Finding the optimum solution requires a huge amount of calculation in classical methods. Recently, a genetic algorithm (GA) technique has attracted considerable attention among various modern heuristic optimization techniques. To optimize the parameters and the number of centrifuges for a given cascade a Real Coded Genetic Algorithm program, RCGA, is implemented and developed. It has been shown that the application of RCGA to this problem not only reduces the amount of calculation but also guarantees finding the best solution. It is found that, from an evolutionary point of view, the performance of the GA is excellent.