基于模糊C均值聚类的隧道电阻率反演方法及现场试验

Tunnel Resistivity Inversion Method Utilizing FuzzyC-means Clustering and Its Field Application

突涌水灾害是隧道建设面临的主要威胁之一,地下含水体构造的定位和成像是灾害防治的关键.隧道电阻率探测方法对水体响应敏感,是一种有效的含水构造探测方法,能够对含水体位置与规模进行探测.近年来,隧道工程建设面临的地质条件愈加复杂,对探测精度的需求不断提高,亟需开展更高精度的隧道电阻率反演方法研究.基于模糊C均值聚类的隧道电阻率反演方法通过将聚类算法引入到隧道电阻率反演算法中,建立基于模糊C均值聚类算法的隧道电阻率反演目标函数和反演方程,实现了电阻率聚类反演成像.针对隧道前方典型含水体开展了数值模拟,并依托滇中引水工程香炉山隧洞开展了现场应用试验,研究结果表明:相较于传统电阻率反演结果,基于模糊C均值聚类的隧道电阻率方法能够较准确刻画不同尺度含水体的规模,提高了反演结果的分辨能力和物性值准确度.

Water inrush disasters pose a significant challenge to tunnel construction.To mitigate this threat,it is crucial to employ location imaging techniques for underground water-bearing structures,as they play a pivotal role in disaster prevention.Due to its sensitivity to water body response,the tunnel resistivity detection method emerges as a practical approach for detecting water-bearing structures.The successful detection of water bodies has been achieved in determining their location and scale.However,in recent years,the geological conditions for tunnel construction have become increasingly complex,necessitating a higher level of detection accuracy.Therefore,there is an urgent need to research tunnel resistivity inversion methods that offer enhanced precision.By introducing the clustering algorithm into the tunnel resistivity inversion algorithm,the objective function and inversion equation of tunnel resistivity inversion based on fuzzy C-means clustering algorithm are established,and the tunnel resistivity inversion method based on fuzzy C-means clustering is proposed to realize the resistivity cluster inversion imaging.Additionally,numerical simulations were conducted for a representative water body located in front of the tunnel,and a field application test was performed at the Xianglushan tunnel of the Central Yunnan diversion project.The findings suggest that the tunnel resistivity method,utilizing fuzzy C-means clustering,offers a more precise depiction of water size at various scales compared to conventional resistivity inversion results and enhances the resolution and accuracy of the inversion outcomes.