基于小波包的机械故障信号消噪处理

De-noising of Sign in Machinery Fault Diagnosis Based on Wavelet Packet

在旋转机械故障诊断系统中需要获得采样信号用于系统的故障辨识或判断,而这些采样信号将不可避免地受到噪声污染,使系统辨识精度得不到提高。小波包变换的方法可以降低系统噪声影响,通过变换分解出高频噪声部分,利用小波包收缩的阈值量化方法去掉高频部分,再重构信号从而达到消噪目的。

In many real fault diagnosis systems, sampling signals are inevitably influenced by noise, which makes fault signals identified from these signals exist error and disturbs improving identified precision. The wavelet packet transform is used to reduce the high frequency noise of fault diagnosis system signal. This transform decomposes the signal, thus separating the high frequencies due to noise effects. After the decomposition, the wavelet shrinkage is performed by thresholding to eliminate the high frequency component. Then the signal can be reconstructed without the high frequency component.