Genetic algorithms for nuclear reactor fuel load and reload optimization problems

Abstract

Approaches are examined in the present paper to the application of genetic algorithms for optimization of initial reactor load and subsequent reloading and reshuffling of fuel assemblies in the nuclear reactor core. The issues associated with selection of the optimization criterion, which was chosen to be the nuclear fuel burnup depth, are discussed. The burnup depth is estimated after the fuel assembly is unloaded from the core, i.e. after residence in the reactor core during 3 fuel irradiation campaigns. An important aspect determining the efficiency of the use of the genetic algorithm in the problem under examination is that the neutronics calculation of the reactor core is to be performed in sufficient details allowing "feeling" the change in the location of the fuel assemblies relative to each other. The use of low-precision instrument results in the uselessness of the proposed approach to the optimization of reactor core loading. The opposite extreme, i.e. the excessive degree of details, is associated with significant increase of expended computer CPU time. In the presented paper, the TRIGEX [1,2] application software package was used in the analysis of neutronics characteristics of the reactor core providing acceptable degree of details and capable to demonstrate sensitivity of the results to the changes in the reactor load arrangement. The genetic algorithm incorporates the use of at least two basic procedures—selection and mutation. One of the most important issues in the application of the genetic algorithm is the definition of the basic concepts, namely the concepts of mutation, crossing, and specimen. The answers to these questions as applicable to the problem under discussion are provided in the present paper. In addition, the main recommendations for the organization of crossing and mutation procedures are also given. The efficiency of use of the developed model of the genetic algorithm is demonstrated by the test example of a BN type reactor. The results of the test run demonstrated that the use of the proposed approach allows searching for optimal reactor load mapping for each separate core reshuffling operation. The main objective of the performed study was to demonstrate the applicability and efficiency of the new up-to-date approach to solving the problem of fuel loading into a nuclear reactor.

Keywords

• Optimization of fuel initial loading and reshuffling
• Burnup value
• Genetic algorithm
• Nuclear reactor