基于PSO神经网络的故障诊断方法研究

将粒子群优化算法和BP算法相结合,提出了一种基于粒子群神经网络的故障诊断方法.该方法分阶段实施神经网络的训练,有效地加强了算法的全局搜索能力,采用PSO优化了传播中的权值、阈值以及网络结构.这不仅弥补了BP算法的不足,而且删除了冗余连接,提高了故障模式识别的能力.仿真结果表明该方法加快了神经网络的学习收敛速度,提高了故障模式的识别正确率,可以有效地应用于设备的故障诊断.

Research on fault diagnosis method based on PSO neural network

A new fault diagnosis method based on PSO neural network is proposed by combining PSO and BP algorithm.This method trained the neural network with two steps,the global search ability is enhanced,and the connecting weighs,thresholds and structure of neural network are optimized.Thus,it makes up the lacks of BP algorithm,eliminates the redundancy connecting,and improves the ability of fault pattern recognition.The experimental result shows that new method speeds up the convergence process,enhances the accurate rate in fault pattern recognition,and will be applied to equipments fault diagnosis effectively.