A preliminary study on applicability of artificial neural network for optimized reflector designs. (December 2017)

• Record Type: Journal Article
• Authors: Kim, Song Hyun; Vu, Thanh Mai; Pyeon, Cheol Ho
• Is Part Of: Energy procedia. Volume 131(2017)
• Page Start: 77
• Page End: 85
• Publication Date: 2017-12
• Subjects: Artificial Neural Network -- ANN -- Reflector Design -- Fuel Pattern -- Arrangement -- Optimization -- Nuclear Reactor -- Criticality; Power resources -- Congresses; Power resources -- Periodicals; Power resources; Conference proceedings; Periodicals; 333.7905
• DOI: 10.1016/j.egypro.2017.09.478
• Languages: English
• ISSNs: 1876-6102

Abstract

Abstract: The neutron reflector is a material to reflect neutrons into reactor cores. The reflectors are designed with their one purpose such as increasing the criticality, specific flux distribution, and others. Generally, the reflector design has been conducted by the experiences of designers due to the lots of design variables such as material selection and arrangement. In this study, the applicability of the artificial neural network is preliminarily studied for the optimization of the reflector arrangement. For the research, a system of artificial neural network was developed using C++ program language. The feedforward neural network was used with three layers which are input, hidden, and output layers. The back-propagation algorithm was adopted for the training of the neural network. After the construction of the neural network system, the optimization and auto machine learning algorithms was developed by C++ programing language for the preliminary study on the applicability of artificial neural network into the reflector design. The results show that the reflector gives a good performance to obtain the goal responses. It is expected that this system can contribute to dramatically increase the efficiency of the reflector designs.