Deep learning-interface flow coupling-based pressurized water reactor neutron transport method

Abstract

The invention discloses a pressurized water reactor neutron transport method based on deep learning-interface flow coupling, and the method comprises the steps: simulating a neutron migration process after the capture, fission or scattering of neutrons and materials in a local simple geometric problem through a Monte Carlo method; a simulation result of a Monte Carlo method and a deep learning method are utilized to learn neutron incidence-emergence distribution rules of different materials and geometric units, and a connection relation between simple geometric units and a neutron transport calculation process are established based on interface flow conservation. According to the method, the traditional process that in a deterministic theory method, resonance calculation must be carried out firstly to obtain a multi-group cross section, and then complex geometric multi-group transport calculation is carried out is subverted, and a novel neutron transport calculation method integrating a Monte Carlo method, deep learning and the deterministic theory method is formed.

Description

technical field [0001] The invention relates to the field of nuclear reactor core design and safety, in particular to a pressurized water reactor neutron transport method based on deep learning-interface flow coupling. Background technique [0002] In the current development of nuclear energy, the Monte Carlo method has high computational accuracy and geometric adaptability for the calculation of neutron transport in a pressurized water reactor with complex geometry. However, a large number of particle sampling calculations lead to low computational efficiency. Compared with the Monte Carlo method, the deterministic method has high computational efficiency, but the current resonance calculation method and transport calculation method cannot meet the high-precision and high-efficiency calculation requirements of the complex geometry PWR. The deep learning method has a strong ability to learn complex laws. It can use the deep learning method to learn the migration laws of a la...

AI Technical Summary

Problems solved by technology

[0002] In the current development of nuclear energy, for the calculation of neutron transport in PWR with complex geometry, the Monte Carlo method has high calculation accuracy and geometric adaptability, but a large number of particle sampling calculations lead to low calculation efficiency

Compared with the Monte Carlo method, the deterministic method has high computational efficiency, but the current resonance calculation method and transport calculation method cannot meet the high-precision and high-efficiency calculation requirements of complex geometric PWRs.

Images