Multiobjective Core Reloading Pattern Optimization of PARR-1 Using Modified Genetic Algorithm Coupled with Monte Carlo Methods. (19th April 2021)

• Record Type: Journal Article
• Authors: Shaukat, Nadeem; Ahmad, Ammar; Mohsin, Bukhtiar; Khan, Rustam; Khan, Salah Ud-Din; Khan, Shahab Ud-Din
• Other Names: Deinert Mark Academic Editor.
• Is Part Of: Science and technology of nuclear installations. Volume 2021(2021)
• Publication Date: 2021-04-19
• Subjects: Nuclear engineering -- Periodicals; Nuclear facilities -- Periodicals; Nuclear engineering; Nuclear facilities; Electronic journals; Periodicals; 621.48
• DOI: 10.1155/2021/1802492
• Languages: English
• ISSNs: 1687-6075

Abstract

Abstract : In order to maximize both the life cycle and efficiency of a reactor core, it is essential to find the optimum loading pattern. In the case of research reactors, a loading pattern can also be optimized for flux at an irradiation site. Therefore, the development of a general-use methodology for core loading optimization would be very valuable. In this paper, general-use multiobjective core reloading pattern optimization is performed using modified genetic algorithms (MGA). The developed strategy can be applied for the constrained optimization of research and power reactor cores. For an optimal reactor core reloading design strategy, an intelligent technique GA is coupled with the Monte Carlo (MC) code SuperMC developed by the FDS team in China for nuclear reactor physics calculations. An optimal loading pattern can be depicted as a configuration that has the maximum k eff and maximum thermal fluxes in the core of the given fuel inventory keeping in view the safety constraints such as limitation on power peaking factor. The optimized loading patterns for Pakistan Research Reactor-1 (PARR-1) have been recommended using the implemented strategy by considering the constraint optimization, i.e., to maximize the k eff or maximum thermal neutron flux while maintaining low power peaking factor. It has been observed that the developed intelligent strategy performs these tasks with a reasonable computational cost.