Teaching-learning based optimization algorithm for core reload pattern optimization of a research reactor

Abstract

A software design package has been developed for optimization of in-core fuel loading pattern of material test reactor PARR-1 (Pakistan Research Reactor-1) by using a nature inspired population based search algorithm called Teaching-Learning based optimization (TLBO) algorithm, which mimics the teaching and learning process of a classroom. This software package of TLBO and diffusion theory code PRIDE has been coupled to search for an optimal core by maximizing the effective multiplication factor (k-eff). Moreover as a second step, sensitivity analysis has been performed for teaching factor of TLBO and it is found that adaptive teaching factor is giving better results. Without any check on power peaking, by using adaptive teaching factor, the optimal core configuration with enhanced k-eff value of 1.0330490 was achieved as compared to the optimal core with k-eff 1.0304509 found without using adaptive teaching factor. As a third step, the penalty related to power peaking factor (ppf) was introduced, and TLBO with adaptive teaching factor is applied to optimize the core configuration. In the next step, the elitism is applied in TLBO and the worst solutions in each generation are replaced by elite solutions. Elitism has not only improved the results (i.e. increase in k-eff from 1.0174527 to 1.0207885) but convergence speed of the algorithm has also been increased. As a last and final step, an enhanced multi-objective optimization problem was formulated with proper weighting factors and ETLBO algorithm is applied to search optimal core. In this case, the multiplication factor for the searched core is further increased to 1.0212827 with a decreased power peaking factor of 1.29.