An artificial neural network based neutron field reconstruction method for reactor

作者：Cao Pei; Gan Quan; Song Jing; Long Pengcheng; Wang Fang; Hu Liqin; Wu Yican

来源：Annals of Nuclear Energy, 2020, 138: 107195.

DOI：10.1016/j.anucene.2019.107195

• Neutron spectrum
• Advanced reactors
• Ex-core measurement
• Truncation plane
• Artificial neural networks

Nowadays, there are many reconstruction methods base on the ex-core detectors to monitor the state of the core in real-time, but none of these methods can calculate the neutron spectrum which has always been an essential parameter in advanced reactors design, performance diagnoses and safe operation. Thus, by utilizing ex-core measurements, a reconstruction neutron field with energy distributions method based on artificial neural networks was proposed in this paper. Furthermore, in this study, in order to amplify the contribution of low-energy neutrons to the ex-core detectors, the idea of truncation plane is used to calculate the response matrix library. To validate the method, neutron field of pressure vessel regions and core regions in a homogenized assembly-level VENUS-3 reactor model were obtained through transport calculation and reconstruction calculation. The comparison of results demonstrated that the reconstruction method was effective and available to reconstruct the neutron spectrum in reactors.