Application of Improved PSO-BP Neural Network in Fault Detection of Liquid-Propellant Rocket Engine

Because liquid-propellant rocket engine (LRE) is working in complex environment, the reliability of some traditional fault detection methods is low due to the characteristics of limited fault samples, small samples, high coupling and nonlinear variation of the main engine. In order to improve the shortcomings of fault detection in traditional method, this paper presents a fault detection algorithm (PSO-BP) based on improved particle swarm optimization (PSO) algorithm and BP neural network. The structure of BP neural network is established, the collected samples are input into PSO algorithm to optimize, and the optimal BP parameters are found, and the fault detection model of the main engine is established. Finally, the established model is used to simulate the fault. The simulation results show that the proposed algorithm can improves the accuracy of fault detection in the health management of LRE and provides a guarantee for the market demand of commercial launch.