基于Bayes阈值估计的小波在转子故障信号降噪中的运用

提出了一种有效的转子故障信号的降噪方法.该方法利用概率统计的最优估计原理对样本信号进行估计,增加了小波阈值参数的可靠性.根据Bayes估计风险最小的原则,对不同子带的小波系数估计小波阈值,通过软阈值函数进行逐层降噪,对降噪后的信号进行重构,达到对转子故障信号降噪.把从多功能振动实验台上取得的转子故障振动信号,利用该方法进行降噪.并将降噪结果与Birge-Massart小波阈值降噪法和Donoho小波阈值降噪法的处理结果进行比较,结果显示使用Bayes(贝叶斯)阈值估计降噪法处理的信号能够很好地保留信号的细节部分,而且信噪比(SNR)也有明显提高.

Signal Denoising of Rotor Fault via Sym6 Wavelet Based on Bayes Estimation

A kind of effective rotor fault signal de noising method is proposed,optimal estimation of the probability and statistics theory is adopted to this method,the measuring signal is decomposed by using wavelet in the method,the reliability of wavelet threshold is increased.According to the Bayes risk minimum principle,wavelet threshold is estimated in wavelet coefficients of different sub bands.Then the de noised signal is obtained and reconstructed layer by layer.Rotor fault vibration signals obtained from the multi stage experiment console are decomposed down via the method.Comparing it with the results of Birge Massart threshold method and Donoho threshold algorithm shows that the signals disposed via the method mentioned above are better to retain some details of the original signals,and signal to noise ratio has been improved significantly.