改进PSO优化WNN的液体火箭发动机故障检测

Improved PSO and Optimized WNN for Liquid Rocket Engine Fault Detection

提出一种改进粒子群优化的小波神经网络模型,将其应用于火箭发动机的故障检测研究。针对传统粒子群算法初期容易陷入局部最优的问题,改进粒子群算法的惯性权重和学习因子,采用逐渐递减的选取方式。进行动态调整后的粒子群算法有利于在初始迭代时寻找满足条件的局部最优值,在寻找到局部最优值之后能够快速地收敛逼近于全局最优值,提高运算效率。此外,为了提高小波神经网络的学习速率,对所采用的小波神经网络权值和小波基函数参数增加了动量项。两种算法相结合,最终提出一种改进粒子群算法(IPSO)与小波神经网络(WNN)结合的模型。最后根据MATLAB仿真和数据分析表明,新算法可以很好地用于液体火箭发动机的故障检测研究,并且IPSO-WNN模型比BPSO-WNN和WNN模型后期具有较快的局部收敛能力,预测更加准确。

A wavelet neural network model with improved particle swarm optimization is proposed and applied to the research of fault detection of rocket engines.In view of the problem that the traditional particle swarm algorithm is easy to fall into the local optimum in the early stage,the inertia weight and learning factor of the particle swarm algorithm are improved,and the selection method of gradually decreasing is adopted.The particle swarm algorithm after dynamic adjustment is beneficial to find the local optimal value that satisfies the conditions in the initial iteration.After finding the local optimal value,it can be quickly converged to the global optimal value and improve the computational efficiency.In addition,in order to improve the learning rate of the wavelet neural network,a momentum term is added to the wavelet neural network weights and wavelet basis function parameters used.The two algorithms are combined to finally form a proposed model that combines improved particle swarm optimization(IPSO)with wavelet neural network(WNN).Finally,the MATLAB simulation and data analysis show that the new algorithm can be used for fault detection research of liquid rocket engines and the IPSO-WNN model has faster local convergence ability and more accurate prediction than BPSO-WNN and WNN models in the later stage.