Optimization design of passive residual heat removal system based on improved genetic algorithm

Abstract

The passive residual heat removal system can improve the inherent safety of the nuclear power plant (NPP). Adopting the passive residual heat removal system is an important trend of marine nuclear power plant (MNPP). The MNPP is Limited by space size and weight, and lightweight and safety must be fully considered in the passive system design process. In the process of designing and optimizing the passive residual heat removal system, many thermal–hydraulic analysis programs need to be called to calculate the constraint functions, resulting in high calculation costs. The surrogate model technology can be used to reduce the calculation costs and improve the efficiency of the optimization process. Aiming at the shortcomings of the traditional Back Propagation (BP) neural network, this paper adopts the two-layer adaptive genetic algorithm (AGA) to optimize the BP neural network to construct the AGA-BP model. Based on the AGA-BP model and the two-layer adaptive genetic algorithm based on the adaptive penalty function, an optimal design method of the passive residual heat removal system is formed, and it is applied to the optimal design of a marine passive residual heat removal system. The results show that the weight of the passive residual heat removal system is reduced by 3.58% after the optimization scheme is adopted. Finally, the feasibility of the optimization results and the safety of the nuclear power plant in the accident was verified by the thermal–hydraulic analysis program RELAP5. The proposed method provides direction for reducing the weight of the passive residual heat removal system.