基于冗余第二代小波变换的振动信号降噪

The noise reduction of vibration signals on the basis of redundant second-generation wavelet

为了能够提取隐藏在振动信号中的冲击信号,设计了冗余第二代小波变换。采用拉格朗日插值细分原理设计各层的初始预测器和初始更新器,然后通过冗余方案对原始信号进行分解,并对各层细节信号进行阈值处理,将处理后的信号进行重构,即可获得降噪后的信号。对仿真信号和工程实际信号的降噪表明,用冗余第二代小波对原始信号降噪能获得更高的信噪比和更小的均方差,且能完整地保留信号中的冲击成分。在某油田往复压缩机状态监测和故障诊断中,采用该方法有效地提取了曲轴振动信号中的冲击信号。

To extract the impact signal hidden in the vibration signal,the redundant second-generation wavelet transformation was designed.Adopting the Lagranges interpolation subdivision principle,the initial predictor and updater of the layers were designed.Then,the redundancy scheme was utilized to decompose the original signal and the threshold was adopted to deal with every stage's detail signal.The processed signals were reconstructed and the signal after noise reduction was obtained.The noise reduction of simulation signal and actual engineering signal indicated that utilization of redundant second-generation wavelet can achieve higher signal-to-noise ratio and smaller mean square deviation in noise reduction of the original signal and the impact element of the signal can be preserved completely.In the condition monitoring and fault diagnosis in a certain oilfield,the utilization of this method achieved an effective extraction of the impact signal of the vibration signal of the reciprocating compressor crankshaft.