Delay corrected independent component analysis algorithm and its key parameters based on tunnel seismic prediction periodic interference suppression

The process of tunnel seismic prediction (TSP) is often accompanied by near-periodic interference such as power frequency interference and pump vibration, which seriously affects the recognition and extraction of target signals in TSP, resulting in the failure to accurately obtain the geological structure information in front of the tunnel face. At present, the commonly used method to suppress periodic noise is notch filter, which is simple and fast, but its application range is limited. When the frequency spectrum of the periodic interference overlaps with that of the effective signal, the effective signal will be seriously damaged while the periodic interference is suppressed. To solve this problem, a delay-corrected independent component analysis (DC-ICA) algorithm is proposed, which can separate the periodic interference from the effective signal by using the statistical characteristics. The processing effect of DC-ICA algorithm is closely related to the key parameters such as the number of input seismic records, the delay time of target signal, the accuracy of frequency estimation and the frequency range. In this paper, the key parameters of DC-ICA algorithm will be deeply studied and analyzed, and then, the appropriate range of key parameters will be determined to achieve the suppression of periodic interference and high-quality extraction of target signals, so as to improve the distance and accuracy of tunnel geological advance detection.