Performance evaluation of uranium enrichment cascades using fuzzy based harmony search algorithm

Abstract

The production of energy in nuclear reactors needs enrichment of fuels. There is some interest in taking the fuel enrichment level to 3–5% by cascades. Optimization of the isotopic cascades is essential to make this process economic. This study presents a Fuzzy-based Harmony Search (FHS) algorithm aimed at dynamic parameter adaptation as well as establishing a balance between exploration and exploitation, which significantly increases the convergence speed of the algorithm. Accelerating the convergence of the algorithm is demonstrated in the Sphere, Schwefel, Ackley, and Drop-Waves benchmarks at first. This approach also enhances performance in several test cases of optimum cascade problems, with results validated through comparisons with conventional methods. According to the results, the total number of centrifuges using FHS reached 6306 in test case 1, which was reduced 44 pieces compared to the method used by Palkin, and 55 pieces compared to the real coded genetic algorithm. The total number of centrifuges using FHS reached 2808 in test case 4 with a different type of gas centrifuge, which decreased 27 pieces compared to the direct search method. Similar results were obtained in other test cases, indicating the effectiveness of the FHS algorithm in minimizing the total number of centrifuges and total flow rates.