Location method of microseismic source and its engineering application: anisotropic velocity model and local optimization

Accurately locating microseismic (MS) sources and delineating damaged zones are crucial for understanding rock mass damage. In tunnel engineering projects, MS monitoring faces significant challenges due to complex wave velocity structures and intense construction disturbances . This study proposes a hybrid positioning workflow comprising three steps: waveform denoising, regional gridding, and inversion-based localisation. A residual convolutional neural network is employed to precisely extract the arrival time of MS P-waves. Subsequently, an anisotropy layered velocity model is constructed using advanced geological prediction to global search potential MS source locations. The improved artificial bee colony (IABC) algorithm is then applied for efficient local optimisation final MS event. To validate the proposed method’s accuracy, blasting and MS events were analysed from a tunnel in Southwest China. Results indicate that, compared to the simulated annealing-particle swarm optimisationalgorithm, the IABC algorithm, and the particle swarm optimisation algorithm, the proposed method enhances blasting event localisation accuracy by 60.75%, 29.23%, and 57.85%, respectively. Furthermore, analysis of real tunnel MS data reveals a strong correlation between the identified MS aggregation zones and the observed damaged areas.