SWD data preprocessing using wavelet transform of correlation domain
利用相关域小波变换进行SWD资料预处理

Wave fields of Seismic While Drilling (SWD) are very complex, and it is an important preprocessing to analyze the signal from bit and the noise from machines on ground. In the paper, we use temporal structures, such as periods and time-delay of signal and noise, and propose an algorithm based on wavelet transform of correlation domain to analyze SWD wave fields. Signal propagated and reflected multiply in drill-string, so it has strict periodicity. Noise from machines such as pump has periodicity too. In auto-correlation domain, these periodic components are emerged. Components of SWD wave fields have different time delays in sensors, can emerge in cross-correlation domain at certain delay. Wavelet transform is used to analyze these components emerging in correlation domain in multi-scales, and their frequency contents, periods, attenuation can be known. Based on this information, a proper SWD data processing is carried out and the direct wave of signal from bit is gained. It is indicated that the method of wavelet transform of correlation domain is a useful preprocessing.