基于声发射的岩石损伤三维成像传感器动态优化方法研究

Research on Dynamic Optimization Method of Sensors for Rock Damage 3D Imaging Based on Acoustic Emission

摘要: 针对传统岩石损伤探测中传感器静态布设带来的信息不全、难以获取三维空间精细损伤分布的问题，本文提出了一种基于声发射技术的动态优化布设与三维损伤成像方法。具体来说，先利用初至波走时反演构建初始波场矩阵，再通过KD树聚类把损伤区域划分出来。对于每个聚类出来的损伤子区，融合RANSAC和HOUGH变换两种算法去拟合出概率较高的损伤路径，然后动态调整传感器的位置，做第二次探测，并把新得到的波速场矩阵和初始矩阵合并到一起。最后用克里金插值算法生成完整的波速场矩阵，进而反演出岩石内部损伤的三维分布情况。总的来说，这套方法为岩石内部损伤的精细化探测提供了一种动态优化传感器布设的新思路。通过分区聚焦探测加上多源数据融合，有望在降低探测成本的同时，提高成像的完整性和定位精度，也能为地质工程中的灾害预警提供一些实用的技术参考。

Abstract: Aiming at the problem of incomplete information and difficulty in obtaining fine damage distribution in three-dimensional space caused by static layout of sensors in traditional rock damage detection, this paper proposes a dynamic optimization layout and three-dimensional damage imaging method based on acoustic emission technology. Specifically, the initial wave field matrix is constructed by using the first arrival travel time inversion, and then the damage area is divided by KD tree clustering. For each clustered damage sub-region, the two algorithms of RANSAC and HOUGH transform are fused to fit the damage path with higher probability, and then the position of the sensor is dynamically adjusted for the second detection, and the newly obtained wave velocity field matrix and the initial matrix are merged together. Finally, the Kriging interpolation algorithm is used to generate a complete wave velocity field matrix, and then the three-dimensional distribution of rock internal damage is inverted. In general, this method provides a new idea for dynamic optimization of sensor layout for fine detection of rock internal damage. Through the partition focusing detection combined with multi-source data fusion, it is expected to improve the integrity and positioning accuracy of imaging while reducing the detection cost, and also provide some practical technical reference for disaster warning in geological engineering.