A Multi-Indicator Fusion-Based Technique for the Identification of Acoustic Emission Signals During Rock Failure

With the widespread application of acoustic emission (AE) technology in geotechnical engineering, effectively separating and identifying dense AE signals generated during rock fracturing remains a critical challenge. This study proposes an AE event identification technique based on waveform energy envelopes and multi-indicator characteristic parameters. First, the waveform energy envelope is used to adaptively segment dense and partially overlapping AE waveforms without relying on fixed timing parameters. Then, a template sliding-window scan integrating waveform correlation, ring count, rise time, and signal energy is performed to identify candidate AE events. In addition, a time-difference correction and window-stacking strategy is adopted to improve multi-channel arrival picking. Experimental validation on representative single-peak single-event and double-peak multi-waveform cases extracted from laboratory rock-failure tests demonstrates that the proposed method can effectively separate and identify AE waveforms under the tested conditions. Compared with conventional timing-parameter-based segmentation and correlation-dominated matching, the proposed workflow is more robust to waveform attenuation and distortion. The method provides a methodological basis for AE waveform identification and arrival-time extraction in rock-failure monitoring and has potential to support early warning after further validation.