Improvement of Autoregressive Model-Based Algorithms for Picking the Arrival Times of the P-Wave of Rock Acoustic Emission

• Published in: Geotechnical and geological engineering Vol. 41; no. 2; pp. 707 - 719
• Main Authors: Wang, Guilin, He, Chenhao, Liang, Feng, Gong, Sheng
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.03.2023; Springer Nature B.V
• Subjects: Accuracy; Acoustic emission; Acoustics; Algorithms; Autoregressive models; Autoregressive processes; Bayesian analysis; Civil Engineering; Criteria; Earth and Environmental Science; Earth Sciences; Efficiency; Errors; Geotechnical Engineering & Applied Earth Sciences; Hydrogeology; Methods; Original Paper; P waves; Probability theory; Rock; Rock mechanics; Rocks; Terrestrial Pollution; Time series; Waste Management/Waste Technology; Waveforms; Wavelet transforms; Acoustic emission; Autoregressive model; P-wave arrival time picking; Algorithm improvement
• ISSN: 0960-3182; 1573-1529
• DOI: 10.1007/s10706-022-02296-2

Acoustic emission (AE) monitoring technology has been widely used in rock engineering. In addition, the accurate picking of P-wave arrival times is the key to in-depth rock mechanics and AE research. Autoregressive (AR) model-based algorithms such as AR-Akaike information criterion (AR-AIC) and AR-Bayesian information criterion (AR-BIC) are efficient methods currently adopted for P-wave arrival time picking; however, their picking results sometimes have large errors, owing to the complexity of amplitude distribution in the time series of AE waveform data. To minimize these errors, this study improved the adopted algorithms by leveraging the surge phenomenon in AR model variance. Specifically, a single-segment AR-BIC algorithm was developed by pre-processing the time series of AE waveform data. It was verified that the improved AR-BIC algorithm achieved a 50% higher efficiency than the conventional algorithm, including a picking accuracy above 98.5%.