Matching the velocity model to the actual engineering and geological conditions and improving the accuracy and stability of the microseismic (MS) source location remain challenges for scientists. An MS source location method based on a velocity model database and statistical analysis, named LM-VMD-SA, is proposed in this study. The method firstly divides the monitoring area into different subareas based on four influencing factors and creates an initial velocity model database by assigning an initial velocity to each sensor combination. Secondly, blasting tests are carried out in each subarea, where the velocity model database is inverted using a location error optimization method based on the pattern search algorithm (LEOM-PSA). The initial velocity model database for each subarea is updated by the velocity model database of the blasting events in the same subarea, and a velocity model database is constructed. Then, the velocity models for all sensor combinations of an MS event are called from the velocity model database for the corresponding subarea by matching the sensor combination of the MS event, and all corresponding solutions of the MS event are solved by the ND-N method. Finally, the three-dimensional coordinates of MS source are identified by utilizing the log-logistic (3P) distribution probability density function. According to blasting tests in the Beiminghe Iron Mine, the location accuracy of the proposed method is 20.88% and 18.24% higher than that of the traditional method and subarea method, respectively. The application of the proposed method to the Beiminghe Iron Mine revealed the illegal mining activities at −125 m and −155 m level, providing effective technical support for mineral resources protection and mining safety.

将速度模型与实际工程和地质条件相匹配并提高微震 (MS) 震源位置的准确性和稳定性仍然是科学家面临的挑战。本研究提出了一种基于速度模型数据库和统计分析的MS源定位方法，称为LM-VMD-SA。该方法首先根据四个影响因素将监测区域划分为不同的子区域，并通过为每个传感器组合分配初始速度来创建初始速度模型数据库。其次，在每个分区进行爆破试验，使用基于模式搜索算法（LEOM-PSA）的定位误差优化方法对速度模型数据库进行反演。每个分区的初始速度模型数据库由同一分区内爆破事件的速度模型数据库更新，并构建速度模型数据库。然后，通过匹配MS事件的传感器组合，从相应分区的速度模型数据库中调用MS事件所有传感器组合的速度模型，并通过ND-N方法求解MS事件的所有对应解. 最后，利用log-logistic(3P)分布概率密度函数识别MS源的三维坐标。北鸣河铁矿爆破试验表明，所提方法的定位精度比传统方法和分区法分别提高了20.88%和18.24%。