提升小波包和神经网络在轴承故障诊断中的应用

Application of Lifting Wavepack Transform and Neural Networks to Fault Diagnosis of Bearing

针对轴承故障信号的特点,采用9/7提升小波包和概率神经网络(Probabilistic Neural Networks)相结合的算法对轴承故障进行诊断。首先对原始数据进行小波变换,并对其进行特征提取。然后利用概率神经网络对得到的特征向量进行类别判定。在VB和Matlab设计的故障诊断仿真实验平台上,验证了9/7提升小波包和概率神经网络混合的故障诊断方法满足实验要求.

According to the features of the vibration signals that come from the bearings, 9/7 lifting wavepack transform and Probabilistic Neural Networks （PNN） were used to analyse the fault data. Firstly, wavelet transform on the original data is performed and its feature vector is calculated. Secondly, the PNN is used to classify the vector. The fault diagnosis method of the 9/7 lifting wavepack and PNN is proved to meet the requirements by the VB and Mat- lab platform.