摘要
微震源定位问题是页岩气勘探、矿山微震研究的重要内容之一。为减小微震定位误差、解决粒子群算法求解微震源反演模型不稳健的情况,采用线性定位法、Geiger法、粒子群优化算法等组成的三步反演法进行微震源定位求解,并研究了粒子群优化算法随机粒子分布类型等对定位结果的影响。第一步,建立线性定位方程组,求解得到初始震源近似解,将其作为Geiger法的初值,计算得到第一步震源近似解结果;第二步,结合第一步定位结果和阈值,设计循环判定准则,执行PSO算法的多次循环,直至达到最大迭代次数后停止,得到第2步的震源近似解;第三步,重复执行第二步若干次得到近似解集合,利用质心法得到最终结果。研究结果表明:三步反演方法相对于单一遗传算法、模拟退火算法或者粒子群算法的定位精度更高,算法更加稳健;群体优化算法的参数对定位效果影响较大,基于F分布的PSO算法更加适合微震源定位,相对于传统正态分布和均匀分布,具有更快的定位求解速度和更精确的定位结果;微震源的三步反演方法在矿山微震源定位精度和算法稳健方面有一定参考价值。
The microseismic source localization problem is one of the important elements of shale gas exploration and mining microseismic research.In order to reduce the microseismic localization error and solve the unstable microseismic source inversion model solved by particle swarm algorithm,a three-step inversion method consisting of linear localization method,Geiger method and particle swarm optimization algorithm is used to solve the microseismic source localization,and the effect of the type for random particle distribution is investigated in the particle swarm algorithm.In the first step,a linear system of localization equations is established and solved to obtain the initial source approximation solution,which is used as the initial value of Geiger’s method to calculate the source approximation solution result in the first step.In the second step,combining the localization result and threshold values in the first step,a cyclic determination criterion is designed to execute multiple cycles of the PSO algorithm until it stops after reaching the maximum number of iterations to obtain the source approximation solution in the second step.In the third step,the execution is repeated step 2 several times to obtain the set of approximate solutions,and the final results are obtained using the center-of-mass method.The results show that:the three-step inversion method has higher localization accuracy and more robuster algorithm compared with single genetic algorithm,simulated annealing algorithm or particle swarm algorithm.The parameters of the population optimization algorithm have a greater influence on the localization effect,and the PSO algorithm based on F distribution is more suitable for microseismic source localization,with faster localization solution speed and more accurate localization results compared with the traditional normal and uniform distribution.The microseismic source The three-step inversion method of microseismic sources has some reference value in terms of localization accuracy and algorithm robustness in mines.
作者
庞聪
李查玮
马武刚
程诚
江勇
廖成旺
PANG Cong;LI Chawei;MA Wugang;CHENG Cheng;JIANG Yong;LIAO Chengwang(Institute of Seismology,CEA,Wuhan 430071,China;Hubei Key Laboratory of Earthquake Early Warning,Wuhan 430071,China;School of Mathematics and Information Technology,Yuncheng University,Yuncheng 044000,China)
出处
《高原地震》
2022年第4期35-40,共6页
Plateau Earthquake Research
基金
中国地震局地震研究所和应急管理部国家自然灾害防治研究院基本科研业务费专项资助项目(项目编号:IS202236328、IS202226321)
中国地震局检测预报科研三结合课题(项目编号:3JH-202201024)
河北省地震动力学重点实验室开放基金资质(项目编号:FZ202212)
湖北省自然科学基金(项目编号:WHYWZ202208)联合资助。
关键词
粒子群算法
微震定位
三步反演
随机粒子
GEIGER
异常值检测
Partide swarm algorithm
Micro-seismic Loacalization
Three-step inversion
Random particles
Geier
Outlier detection
