ADORE-GA: Genetic algorithm variant of the ADORE algorithm for ROP detector layout optimization in CANDU reactors

详细信息    查看全文

• 作者：Doddy Kastanya ; ^{Doddy.Kastanya@candu.com}
• 关键词：CANDU ; Regional overpower protection ; ROVER-F code ; Genetic algorithm ; Detector layout optimization
• 刊名：Annals of Nuclear Energy
• 出版年：2012
• 出版时间：August, 2012
• 年：2012
• 卷：46
• 期：Complete
• 页码：160-168
• 全文大小：1695 K

文摘

The regional overpower protection (ROP) systems protect CANDU? reactors against overpower in the fuel that could reduce the safety margin-to-dryout. The overpower could originate from a localized power peaking within the core or a general increase in the global core power level. The design of the detector layout for ROP systems is a challenging discrete optimization problem. In recent years, two algorithms have been developed to find a quasi optimal solution to this detector layout optimization problem. Both of these algorithms utilize the simulated annealing (SA) algorithm as their optimization engine. In the present paper, an alternative optimization algorithm, namely the genetic algorithm (GA), has been implemented as the optimization engine. The implementation is done within the ADORE algorithm. Results from evaluating the effects of using various mutation rates and crossover parameters are presented in this paper. It has been demonstrated that the algorithm is sufficiently robust in producing similar quality solutions.