电磁辐射信号的小波变换去噪研究

Research on Removing Noise of Electromagnetic Emission Signals with Wavelet Transform

重点分析了实验室实验过程噪声的来源和特征,并利用小波变换的多尺度分析和非线性软门限方法对噪声进行了滤除。结果表明,实验室的噪声主要有固定频带的干扰和随机白噪声两种;对固定频带的噪声可以通过小波变换的多尺度分析进行有效滤除,对白噪声可以通过非线性软门限方法进行有效滤除,得到真正有用的信号。噪声滤除对于分析煤岩电磁辐射信号规律具有重要的意义,可以为现场利用电磁辐射监测技术预测煤岩动力灾害现象提供更加准确可靠的信号。

Electromagnetic emission （EME） can be produced during the deformation ＆ fracture process of coal or rock and EME signals may easily be disturbed by electromagnetic noises. In this paper the sources and characters of electromagnetic emission （EME） noises are analyzed and these noises are removed with multi-resolution analysis of wavelet transform and non-linear soft threshold method. The results show that there are two kinds of noises： interference of fixed frequency band and random white noise. The former can be thrown off by multi-resolution analysis of wavelet transform and the latter can by non-linear soft threshold method effectively. Therefore the useful and true EMR signals can be attained. The removing of noises is very important for analyzing EME rules of coal or rock and the accuracy of forecasting the dynamical catastrophes with EMR monitoring technology.