(1. State Key Laboratory of Hydraulics and Mountain River Engineering，Sichuan University，Chengdu，Sichuan 610065，China；
2. China Three Gorges Corporation，Beijing 100038，China)
摘要快速精准的微震震源定位方法是微震监测技术更好地发挥岩体稳定性预测预警作用的基础。精确的初至波走时计算是提高微震定位精度的关键。针对实际工程中带有空洞和速度分区的复杂波速岩体，引入速度快、精度高和对复杂模型适应性强的多模板快速行进法multi-stencils fast marching methods(MSFM)用于初至波走时计算。该方法利用坐标旋转的方式生成新的计算模板，使网格对角邻点参与计算，提高了the Fast Marching Method(FMM)在对角方向的计算精度。首先，对比分析了一、二阶FMM和MSFM计算精度和效率，表明二阶MSFM具有更高的计算精度；其次，在分层速度岩体和带空洞岩体中分别采用单一速度模型和二阶MSFM计算初至波走时，通过与解析解对比发现，二阶MSFM较单一速度模型的计算绝对误差平均减小了97.65%和95.18%；然后，建立速度分层且带有隧洞的岩体模型，验证了二阶MSFM算法对复杂速度模型适应性极强的特点；最后，提出了基于MSFM的复杂速度岩体微震定位算法，并将其应用到白鹤滩水电站左岸边坡微震定位。通过对现场采集的4个爆破事件进行定位试算，得到使定位误差平均值最小(9.6 m)的相对最优层速度组合并用于考虑空洞的速度模型构建，爆破事件走时正演验证了本速度模型的可靠性。对2015年5月的128个微震事件分别采用单一速度模型和提出的定位算法进行定位，通过对比定位效果，发现后者定位的事件较前者在岩体主要损伤区(层内错动带LS331和LS337)附近有更明显呈条带状分布的聚集现象。研究表明MSFM算法作为走时正演方法具有很好的应用价值，提出的基于MSFM的微震定位方法能够有效地提高震源定位精度。
Abstract：A rapid and accurate microseismic(MS) source location method is the foundation of MS monitoring to play a better forecasting and early-warning role for rock mass stability. The accurate travel time calculation of first arrival wave is the key to improve the location accuracy of MS source. Aiming at rock mass with complex velocities in actual engineering，like voids filled with air and partitioned velocities，the multi-stencils fast marching methods(MSFM) was introduced to calculate the travel time of first arrival wave，which is fast and with high precision and strong applicability for complex model. In this method，a new calculation stencil is generated by means of coordinate rotation to make the diagonal adjacent points of the mesh calculated，and the precision of the fast marching method(FMM) is improved in diagonal direction. Firstly，the accuracy and efficiency of the first or second order FMM and MSFM were compared and analyzed，showing that the second order MSFM has higher accuracy. Secondly，for rock mass with partitioned velocities model and one with a void，both single velocity model and the second order MSFM were used to calculate the travel time. Through comparing them to analytical solution，the average absolute errors of the second order MSFM were reduced by 97.65% and 95.18% compared with the single velocity model. Then，the rock mass model with partitioned velocities model and tunnels was built，and it was proved that the second order MSFM has strong applicability for complex model. Finally，the MSFM-based MS source location algorithm of rock mass with complex velocities was proposed，and applied to Baihetan hydropower station left bank slope of MS location. By trial location of four blasting events collected in fields，the relative optimal layer velocities combination was obtained，which had the minimum mean value(9.6 m) of location errors and was used to establish the void-considered velocity model. The reliability of this velocity model was validated by the travel time forward calculation of a blasting event. The 128 MS events in May 2015 were located by both the single velocity model and the proposed location algorithm，by comparing the location results，the located events by the latter were distributed as belts more obviously near the main damage zone(LS331 and LS337) than those by the former. The study concludes that the MSFM has very valuable application as a travel time calculation method，the proposed MSFM-based MS source location method can improve the source location precision effectively.
2017, Vol. 36 
