改进小波阈值降噪算法在水电机组信号处理中的应用

The Application of an Improved Wavelet Threshold Denoising Algorithm in Signal Processing of Hydropower Units

小波阈值降噪是近年来水电机组状态监测信号去噪常用的算法,经典的阈值函数为硬、软阈值两种函数。在硬阈值方法中,处理过的小波系数在阈值处是间断的,得到的估计信号在重构时易产生局部附加振荡;而经软阈值函数降噪的信号虽然整体连续性好,但与原始信号之间总是存在着恒定的偏差,影响重构精度。因此在软、硬阈值函数基础上提出了一种改进阈值函数的小波降噪算法,通过Matlab仿真和实际采集数据的实验结果表明:该方法克服了传统软、硬阈值函数算法的缺点,不仅连续性好,而且提高了信噪比。通过对比降噪后对振动信号特征能量的保留程度,说明改进阈值函数较传统经典去噪方法更为优越,是一种有效的降噪方法。

Threshold denoising based on wavelet is a commonly used algorithm in the denoising of hydropower unit status monitoring signal in recent years.The classic threshold functions can be divided into hard threshold function and soft threshold function.The wavelet coefficients processed by the hard threshold function are discontinues at the threshold value,which are prone to cause additional partial oscillation on the reconstructed signal.The signal denoised by soft threshold function has a better overall continuity,but there is always a constant deviation between the reconstructed signal and the original one,which is prone to affect the reconstruction accuracy.Based on this situation,a wavelet denoising algorithm with an improved threshold function is put forward.The Matlab simulation with the actual collected data indicate that the new method overcomes the drawbacks of classic ones,not only does it have a good continuity,but also raises the Signal to Noise Ratio（SNR）.By comparing the preservation degree of characteristic energy of vibration signals after denoising,the improved threshold function is superior to the traditional de-noising methods.In summary,it is an effective denoising method.