影响微震震源定位精度的主要因素分析

 微震震源坐标是微震监测最基本和最重要的参数之一,定位精度的高低直接影响着微震分析的结果。研究分析了微震震源定位精度的影响因素,台站误差,速度误差及由优化算法带来的误差。采用优化算法确定震源的坐标及震源事件的发生时间时,不同算法求解的源坐标值相差很大,为此,如何选取合理的优化算法意义重大。将麦夸特法(LM)、简面体爬山法(SM)、准牛顿法(QN)、最大继承法(MIO)、自组织群移法(SOMA)、麦夸特法与全局优化算法(LM-GO)、简面体爬山法与全局优化算法(SM-GO)、准牛顿法与全局优化算法(QN-GO)、最大继承法与全局优化算法(MIO-GO)应用到震源定位中,分析不同算法对震源定位精度的影响。研究发现单纯使用MIO、QN、LM很容易陷入局部最优值,选用时应慎重,联合使用全局最优算法可以显著提高定位精度。SOMA、SM-GO、MIO-GO、LM-GO计算结果较为稳定,定位精度高,建议在声发射源与震源定位时优先选用。

Main Influencing Factors for the Accuracy of Microseismic Source Location

The microseismic source location is one of the most important and foundational parameters in seismic analysis. The positioning accuracy directly affects the results of the microseismic results. The influencing factors for the accuracy of the microseismic source location, including the station error, the velocity error and the error caused by the optimization algorithm, are analyzed and discussed. The source location and the number of source events are determined by an optimization algorithm, and the source coordinates vary greatly with the use of different algorithms. It is important to select a reasonable optimization algorithm with a high accuracy. The famous optimization algorithms, such as the Levenberg-Marquardt Method (LM), the Simplex Method (SM), the Quasi-Newton Method (QN), the Max Inherit Optimization (MIO), the Self-Organizing Migrating Algorithms (SOMA), as well as the Global Optimization (GO) coupled with MIO, QN, MIO and LM, were used to locate and compare the seismic source coordinates. Results show that it is easy to be trapped into a local optimum value using MIO, QN and LM alone. Therefore, the selection of MIO, QN and LM should be careful. The joint use of the global optimal algorithm combined with MIO, QN and LM can significantly improve the positioning accuracy. The SOMA, SM coupled with GO, the MIO coupled with GO and the LM coupled with GO, which are more stable with a high positioning accuracy, should be the recommended methods to locate the seismic or acoustic emission source coordinates.