A denoising method for blasting vibration signals based on improved EMD
Here,aiming at problems of end effect and poor denoising effect in empirical mode decomposition(EMD)algorithm,an improved EMD-based blasting vibration signal denoising method was proposed according to idea of extension-decomposition-clustering-denoising-reconstruction.This method could combine characteristics of comprehensive similarity index to give consideration to both shape and amplitude similarity of extended signals,clustering characteristics of K-means algorithm and denoising advantages of wavelet packet,so it could not only effectively suppress end effect,but also have good denoising effect.The study results showed that in simulated signal's end effect suppression tests,compared with polynomial fitting and boundary local feature extension methods,improved EMD method has the smallest energy error and mean square error;in actually measuring blasting vibration signal denoising,compared with EMD and variational mode decomposition(VMD)methods,improved EMD method has the maximum signal-to-noise ratio of 20.94 dB and the minimum RMS error of 0.003 1;improved EMD method can not only better preserve signal energy with mid-low frequency of 0-200 Hz,but also have a good filtering effect on high-frequency noise above 200 Hz.
万方数据
维普
