采用经验小波变换的风力发电机振动信号消噪

针对风力发电机振动信号非线性特征及恶劣监测环境，分析经验小波变换理论（EWT）及自适应分解特性，提出基于经验小波变换的振动信号消噪方法.采用带噪声leleccum和轴承故障仿真信号对该方法进行消噪效果检验；在同信号源下，与基于db1强制消噪方法、db1软阈值消噪方法和sym5消噪方法分析比较消除噪声效果.针对真实的风力发电机振动信号，验证了基于经验小波变换方法的消噪效果，对同样信号采用其他3种方法进行消噪分析和比较.仿真和实验分析结果表明，基于EWT小波消噪方法，与基于db1强制消噪方法、db1软阈值消噪方法和sym5消噪方法能够达到同样的消噪效果和目的，甚至更优；不损耗原振动信号能量，在自适应模态分解层数方面甚至优于经验模态分解，并且具有较强的鲁棒性.
Abstract: The empirical wavelet transform theory (EWT) and its adaptive characteristics were analyzed according to the nonlinear characteristics of wind power generator vibration signal and the poor monitoring environment. Then a de-noising method was proposed based on EWT. The proposed method was tested by using the leleccum and bearing fault simulation signal with noises, compared with de-noising method based on db1 wavelet with compulsion, db1 wavelet with soft threshold, and sym5 wavelet. The de-noising effect based on EWT was verified for the actual vibration signals of wind power generator. The other three methods were used to eliminate noises analysis and compared with the same signals. The simulation and experimental results show that the de-noising method based on EWT can achieve the same de-noising effect, or even better than the methods based on db1 wavelet with compulsion, db1 wavelet with soft threshold, and sym5 wavelet. The proposed method does not loss the original vibration signals energy, and it is better than the empirical mode decomposition in the adaptive mode decomposition with strong robustness.