基于邻域粒子群混合算法的齿轮故障诊断

Gear Fault Diagnosis Based on Hybrid Algorithm of Particle Swarm Optimization with Neighborhood Operator

为了提高齿轮故障诊断的准确性,采用了一种邻域粒子群混合方法。即根据齿轮的故障特征量,利用邻域粒子群算法来优化BP神经网络的权值,并用优化好的BP网络进行故障诊断。实例仿真结果表明,该方法具有较高的故障诊断准确度,具有一定的实用性。

In order to improve accuracy of gear fault diagnosis,the hybrid algorithm of particle swarm optimization with neighborhood operator is applied. According to fault feature vectors, PSO is applied tooptimize weight of BP neural network,then fault diagnosis is accomplished via optimized neural network. The simulation results show that method has high accuracy of fault diagnosis and has a certainpracticality.