Tunnel Full Time Domain Induced Polarization Advanced Detection Clustering Inversion Imaging Method

With the focus of tunnel construction in China shifting to the western mountainous areas,karst areas and other areas with complex geological conditions,the number of deep-buried long tunnels is increasing,and geological disasters such as water and mud inrush are frequent.Finding out the bad geological conditions in front of the tunnel face in advance can effectively avoid the occurrence of geological disasters and guide the safety of tunnel construction.Tunnel induced polarization method is sensitive to water body,and it is an effective method for advanced prediction of water exploration.It has been widely used in tunnel geological exploration.With the increasing requirements of tunnel engineering on the imaging accuracy and interpretation effect of unfavorable geological bodies such as faults and karst caves,the imaging and identification effect of water-bearing structures in front of tunnel face needs to be further improved.In this paper,through theoretical analysis,numerical test and field test,aiming at the problem of detection and imaging identification of unfavorable geological bodies in the process of tunnel construction,the resistivity inversion and full-time induced polarization multi-parameter imaging method of disaster water body are studied.Firstly,the singleparameter inversion method of tunnel resistivity based on fuzzy C-means clustering is carried out,which improves the imaging and positioning accuracy of water-bearing structures.Then,the selection of clustering center and inversion weight is optimized,and the inversion effect is improved.Finally,a multi-parameter clustering joint inversion method for tunnel induced polarization in full time domain is proposed,which realizes multi-parameter imaging and multiangle interpretation.The effectiveness and feasibility of the method are verified by field tests.The main research work and achievements of this paper are as follows:(1)In view of the obvious volume effect of the traditional resistivity method and the large deviation between the shape of the abnormal body in the inversion results and the actual situation,the clustering algorithm is introduced into the inversion equation,and a tunnel resistivity inversion method based on fuzzy C-means clustering is proposed.The volume effect of the inversion imaging results is effectively suppressed,the resolution of the inversion is improved,and the accurate imaging and positioning of the water-bearing structure in front of the tunnel face is realized.(2)Aiming at the problem that inaccurate parameter selection in the inversion process can easily lead to local optimum and poor inversion results,a fuzzy C-means clustering inversion method guided by discrete values is formed by introducing a priori clustering center into the inversion process.Furthermore,numerical comparison experiments under different clustering centers are carried out to clarify the influence characteristics of clustering centers on inversion results,and a clustering center selection method suitable for tunnel induced polarization clustering inversion is established.Finally,based on the weight parameter optimization strategy,a semi-automatic parameter optimization process suitable for tunnel environment is established,which solves the problem of unstable clustering inversion results caused by random initialization.(3)Aiming at the problem that there are many false anomalies and poor consistency in the inversion results of the full-time domain induced polarization inversion,a multi-parameter clustering joint inversion method for the full-time domain induced polarization of the tunnel is proposed.On this basis,the multi-parameter normalization strategy and weight adaptation strategy of the induced polarization are proposed.The full-time domain induced polarization frequency correlation coefficient,zero-frequency resistivity,intrinsic polarization and relaxation time of the disaster water body are realized.The four parameter imaging provides a feasible means for the multi-parameter comprehensive positioning of the water body in front of the tunnel face.On the basis of the above research,in order to verify the applicability of the method in this paper,numerical simulation tests under different working conditions were carried out.The resistivity method and the full time domain induced polarization method were used to detect the Yellow River Diversion Project in central Shanxi and the Xianglushan Tunnel in central Yunnan.The detection results of these two methods can accurately reveal the spatial distribution of unfavorable geological bodies in front of the tunnel,and verify the feasibility and effectiveness of the method in this paper.