基于小波包能谱熵与自组织RBF神经网络的低压断路器机械故障诊断

Mechanical Fault Diagnosis for Low Voltage Circuit Breaker Based on Energy Spectrum Entropy of Wavelet Packet and Self-Organized RBF Neural Networks

利用小波包分解技术分析断路器故障时的振动信号,提取小波包的能谱熵,将其作为断路器故障模式的特征向量。然后,建立基于K-均值聚类方法的自组织径向基神经网络,对断路器的几种模拟故障进行识别分析,证明了算法的收敛性,给出收敛速度计算公式。通过仿真实验,验证了该方法的有效性,且较之传统BP神经网络有更快的收敛速度和更高的准确度。

By analyzing the vibration signal in the decomposition of wavelet packet when circuit breaker（CB） failed,the energy spectrum entropy of wavelet packet was extracted as the feature vector of failure patterns,Then,self-organized radia basis function（RBF） neural network was established based on K-means clustering method and the simulated faults of the CB was identified,the convergence of the algorithm was proved and the formula of convergence rate was provided.By simulating,the efficiency of the method was verified which was faster in convergence and higher in accuracy,compared with the traditional BP neural networks.