A method for detecting crack wave arrival time and crack localization in a tunnel by using moving window technique

Source localization in a dispersive medium has been carried out based on the time-of-arrival-differences (TOADs) method: a triangulation method and a circle intersection technique. Recent signal processing advances have led to calculation TOAD using a joint time-frequency analysis of the signal, where a short-time Fourier transform(STFT) and wavelet transform can be included as popular algorithms. The time-frequency analysis method is able to provide various information and more reliable results such as seismic-attenuation estimation, dispersive characteristics, a wave mode analysis, and temporal energy distribution of signals compared with previous methods. These algorithms, however, have their own limitations for signal processing. In this paper, the effective use of proposed algorithm in detecting crack wave arrival time and source localization in rock masses suggest that the evaluation and real-time monitoring on the intensity of damages related to the tunnels or other underground facilities is possible. Calculation of variances resulted from moving windows as a function of their size differentiates the signature from noise and from crack signal, which lead us to determine the crack wave arrival time. Then, the source localization is determined to be where the variance of crack wave velocities from real and virtual crack localization becomes a minimum. To validate our algorithm, we have performed experiments at the tunnel, which resulted in successful determination of the wave arrival time and crack localization.