Detection method of karst features around tunnel construction by multi-resistivity data-fusion pseudo-3D-imaging based on the PCA approach
Maoxin Su,Yimin Liu,Yiguo Xue,Kai Cheng,Zexu Ning,Guangkun Li,Kai Zhang
DOI: https://doi.org/10.1016/j.enggeo.2021.106127
IF: 6.902
2021-01-01
Engineering Geology
Abstract:During tunnel construction, large groundwater influx poses serious safety hazards. To ensure construction safety, effective measures such as early warning, grouting, and framework support are needed to counteract potential water hazards. It is also critical to clarify the spatial location and size of the water-bearing geological structures before construction to avoid catastrophe. Electrical resistivity tomography (ERT) operates at a high resolution and is sensitive to low-resistivity targets such as water. Thus, it has strong application potential for the detection of karst geomorphology around tunnel construction. However, the use of a single-ERT method may not be sufficient to accurately locate the spatial distribution of karst features. Therefore, it is necessary to carry out multiple-ERT data fusion to generate a comprehensive interpretation of karst geohazards. According to the spatial distribution of different karst caves, we established four geoelectric models for numerical simulation based on the method of principal component analysis. Studies have shown that data fusion is superior to conventional comparative analysis of geophysical data results and creates a more accurate comprehensive description of the same target. For example, inversion data fusion improves poor imaging of the cross-hole ERT near the borehole and reduces the distribution of pseudo-anomalies near the electrode. In this study, the vertical resolution of the surface ERT, and the horizontal resolution of the cross-hole ERT were both improved, while the imaging ability of the borehole-surface ERT in the peripheral low-sensitivity area was enhanced. Moreover, a three-dimensional (3D) conversion method of two-dimensional (2D) detection data in the tunnel environment was employed. An interpolation 3D-imaging of the 2D detection data was carried out, rendering the interpretation of the results more intuitive and comprehensive. Finally, we applied this method in the Daba Tunnel Water Governance Project in Hunan Province, China, and demonstrated that it could overcome the limitations of the single-ERT survey. We were able to accurately delineate the distribution range of the main water-rich bodies around the tunnel. Our findings provide key guidance for the successful implementation of tunnel water-inrush treatment.