粒子群优化神经网络在高压断路器机械故障诊断中的应用

Application of Neural Network Based on Particle Swarm Optimization in Mechanical Fault Diagnosis of High-Voltage Circuit Breaker

提出了一种以振动信号小波包特征熵为特征向量的高压断路器机械故障诊断的智能算法,该算法利用小波包分解原理将高压断路器振动信号分解到不同的频段中,计算各频段的能量熵值,并将其作为神经网络的输入向量,同时利用粒子群算法对神经网络进行优化,以提高故障诊断的精度。试验结果表明:该方法不仅能够取得良好的分类效果,而且诊断速度与精度均高于传统神经网络算法。

An intelligent algorithm based on wavelet packet- energyentropy（WP-EE）formechanicalfaultdiagnosisofhigh-voltage circuit breaker is presented, in which wavelet packet was used to decompress the vibration signal into different frequency bands. WP-EEwasthen extracted to construct characteristic vectors of signals andisusedasaninputofneuralnetwork, whichwasoptimizedbyparticle swarm optimization （PSO）. The experimental results show that the algorithm can obtain satisfied classification result, and diagnosis speed and accuracy is betterthan traditional nenral network algorithm, thus the proposed algorithm is suitable for mechanical fault diagnosis of HV circuitbreaker.