Weak fault detection based on amplification-compression function and sparse spike deconvolution

Abstract

Weak fault detection is still a hotspot in these years. The key of this work is to enhance the ratio of signal to noise (SNR). Then a new extraction method of weak fault impact based on amplification algorithm is proposed. Since the fault impact is related to bigger derivative, the amplification function combined with the derivative of the vibration signal is used to amplify the amplitude of the impact. Then Sparse Spike Deconvolution (SSD) which has been widely used in earthquake for impact detection is able to extract small impact from the signal processed by amplification function. Then, in order to enhance the periodicity of fault impacts, the compression function is used to compress the amplitude of bigger impacts according to different scales. At last, the experiment result shows that the proposed method is more effective than the common envelope analysis.