基于自适应遗传算法的CSAMT一维反演  

1D inversion of CSAMT data based on adaptive genetic algorithm

作　　者：孙彩堂[1,2] 李玲[1,2] 黄维宁[1,2] 刘真[1,2] 周逢道[1,2] 

机构地区：[1]国家地球物理探测仪器工程技术研究中心(吉林大学),吉林长春130061 [2]吉林大学仪器科学与电气工程学院,吉林长春130026

出　　处：《石油地球物理勘探》2017年第2期392-397,401,共7页Oil Geophysical Prospecting

基　　金：吉林省科技厅发展计划重点项目(20100349);国家潜在油气资源产学研用合作创新研究项目(OSR-02)联合资助

摘　　要：传统的CSAMT反演方法大多依赖于初始模型的选择,并且在反演过程中可能会因为出现病态矩阵而导致反演失败。本文提出了一种基于自适应遗传算法的CSAMT一维反演方法,该方法具有不依赖于初始模型的优点,并且反演过程中不会出现病态矩阵。首先,通过水平层状模型对标准遗传算法和自适应遗传算法进行比较,证明后者的改进效果;然后,在数据中加入随机噪声,证明其具有抗噪性;最后,将其运用到实测数据中,证明了该方法的实用性。Most of conventional controlled source audio-magnetotelluric (CSAMT) data inversions depend on the initial model, and they may fail due to the presence of ill matrix. We propose an adaptive genetic algorithm to 1D inversion of CSAMT data in this paper. This algorithm does neither depend on the initial model nor on generate ill matrix in the process. First with a horizontal layered model we test standard genetic algorithm and adaptive genetic algorithm (AGA), and prove its improvement. Then adding some random noise to data, we test the algorithm and find its good anti-noise ability. Finally, applications in real data prove its practicability. © 2017, Editorial Department OIL GEOPHYSICAL PROSPECTING. All right reserved.

关 键 词：CSAMT 一维反演 自适应遗传算法 

分 类 号：P631[天文地球—地质矿产勘探]