基于改进EMD和小波阈值法的风机数据降噪研究

风机轴承振动信号中混杂着噪声,会对后期的故障诊断造成困难。提出了一种改进经验模态分解(EMD)与小波阈值降噪相结合的风机数据降噪方法。首先,采用EMD对原始振动信号进行分解得到信号的固有模态函数,考虑到各个分量中都含有噪声和信号从而出现模态混叠现象,故采用相关系数法筛选出信号分量和噪声分量对EMD进行改进。然后,采用小波阈值法对噪声分量进行降噪,并将信号分量与处理后的噪声分量进行重构,最终完成信号降噪。最后,选取某风机轴承振动数据作为实验数据,有效地去除了原始信号中的噪声,并得出轴承内圈故障的结论,与实际结果一致,验证了所提方法的有效性。

Noise Reduction of Wind Turbine Data Based on Improved EMD and Wavelet Threshold

A new noise elimination method is improved to solve the problem of fault diagnosis due to the noise mixed in the vibration signal of wind turbine bearing.Firstly,the original signal is decomposed by empirical mode to get the components of each mode function.Because the noise and signal exist at the same time,the modal overlap will occur,the correlation coefficient is used to screen out the signal component and noise component;then wavelet threshold is used to reduce noise;finally the signal component and the processed noise component are reconstructed to reduce the noise and diagnose the fault.The simulation results show its effectiveness.