基于经验学习算法的跨孔电阻率法溶洞探测研究  

Identification of Karst Cave Using Cross-hole Resistivity and Experience-based Algorithm

作　　者：雷静 王珊 李宏涛 罗盈 艾姣姣 LEI Jing;WANG Shan;LI Hongtao;LUO Ying;AI Jiaojiao(、The First Construction Co.,Ltd.of China Fourth Construction Bureau Guangzhou 510800,China;、Guangzhou University Guangzhou 510006,China)

机构地区：[1]中建四局第一建筑工程有限公司,广州510800 [2]广州大学,广州510006

出　　处：《广东土木与建筑》2021年第6期117-120,共4页Guangdong Architecture Civil Engineering

基　　金：国家自然科学基金项目(51808147)。

摘　　要：跨孔电阻率法作为一种常见的地质预报方法,常用来探明地下溶洞。为了改善传统的基于Tikhonov的灵敏度方法易陷入局部最优的缺陷。基于经验学习算法构建了反演方法。通过Matlab建立三种形态(包括裂隙状、袋状和锥状)的三种充水状态(充水、半充水和充气)的溶洞模型。分析不同状态的溶洞的反演成像结果。结果表明,基于经验学习算法的跨孔电阻率法可以清晰的反演出溶洞的形态和充水状态。Cross-hole resistivity method is commonly used to explore karst caves.In order to solve the traditional Tikhonov-based sensitivity method defects that it is easy to fall into local optimal.A new inversion method based on the experience-based algorithm is proposed to improve the performance of the traditional Tikhonov regularization based sensitivity iteration method.A comparison was made between the Tikhonov-based sensitivity method and the experience-based algorithm using nine numerical examples of three water-filled states(water-filled,semi-charged and inflatable)in three forms(fissure-like,bag-shaped,and conical)caves.Results show that the cross-hole resistivity method based on the experience based-algorithm can clearly inverse the shape and water-filled status of the caves.

关 键 词：电阻率法 跨孔 经验学习算法 溶洞 

分 类 号：TU91[建筑科学—建筑理论]