解域约束下的微地震事件网格搜索法、遗传算法联合反演  

A joint inversion combining the grid-search algorithm and the genetic algorithm under solution-domain constraints for microseismic events

作　　者：宋维琪[1] 杨晓东[1] 

机构地区：[1]中国石油大学(华东)地球资源与信息学院,山东东营257061

出　　处：《石油地球物理勘探》2011年第2期259-266,160,共8页Oil Geophysical Prospecting

摘　　要：网格搜索法具有简单、快速的特点,能够快速确定真解的大概位置,但是真解的确定又易受初至、速度模型等因素的影响,很难掌握。遗传算法对初至的适应程度很高,且全局搜索能力强,但是任意设定搜索范围,则搜索过程会繁琐,费时,且解的可靠性也大大降低。文中针对网格搜索法与遗传算法的上述特点,结合微地震资料的实际情况,在研究利用网格搜索反演方法、遗传算法及解域评价的基础上,提出了解域约束下的微地震事件网格搜索法、遗传算法联合反演方法。讨论了微地震反演中模型建立、初至敏感性、射孔资料等与反演计算相关的问题,提出了解决方案。理论和实际资料对文中反演方法的验证结果表明,无论在精度、速度及稳定性方面,文中方法较单一的搜索法或遗传算法都有明显的改善。The grid-search method is simple,fast and can quickly determine a rough solution.But it is difficult to confirm the real solution due to the impact of first arrivals,velocity model and so on.The genetic algorithm adapts to first arrival completely and has strong global search ability.The search process can be complex,time-consuming and the accuracy of solution will be reduced greatly if search domain is set arbitrarily.Taking the advantages of these two methods,we put forward the grid-search and genetic algorithm joint inversion under solution-domain constraints for microseismic events.We discuss the forward modeling,the first arrival sensitivity and perforation information closely related to the inversion calculation.Applications of this method in both synthetic datasets and real datasets prove that the accuracy,the efficiency and the stability are obviously improved by the joint inversion method than the search inversion or the genetic algorithm inversion only.

关 键 词：网格搜索法 解域约束 遗传算法 联合反演 初至敏感性 数值计算 

分 类 号：P631[天文地球—地质矿产勘探]