小波包滤波在齿轮故障诊断中的应用

Application of Wavelet Packet Filtering in Fault Diagnosis for Gears

为快速准确地判断齿轮故障的类型,提出了小波包滤波和神经网络相结合进行齿轮故障分类的方法。介绍了小波包去噪的原理和神经网络的设计方法,对阈值算法和神经网络优化算法作了改进,得到了不含噪声的信号和准确的故障分类方法。仿真结果表明,基于小波包滤波的神经网络方法具有更高的准确性和稳定性,可以满足工业故障诊断的要求。

In order to rapidly classify the types of faults for gears,the method combining wavelet packet filtering and neural network for classifying gear faults is proposed.The principle of wavelet de-noising and the design method of neural network are introduced.The threshold algorithm and neural network optimization algorithm are improved to obtain the signals without noise,and precise fault classifying method.The result of simulation indicates that the neural network method based on wavelet packet features high accuracy and stability,and it satisfies the requirements of industrial fault diagnosis.