Water inrush during tunnel excavation severally threatens the mining safety as blind water-bearing structures may develop in front of the working face. The transient electromagnetic method (TEM) has been widely applied in the advanced detection of tunnel water-bearing structures. However, the metal interference of both supports and tools in the tunnel has become a bottleneck that reduces the forecast accuracy of this method. In this paper, we analyse the effect of metal interference on TEM data and propose a novel set of an observation and correction method under metal interference based on the ratio of anomalous and background apparent resistivity. Flume model experiments both with and without metal interference are carried out, showing that this interference can affect TEM measurements significantly and result in false anomalies, and that our proposed method can remove this ambient noise caused by metal interference appropriately. The practical application further proves that this method can effectively reduce low-resistivity interference introduced by the support and other metal tools inside the tunnel. By applying this correction method, the location of water-rich anomalies can be detected more precisely during the excavation process of the same tunnel, which is of high application value of reducing exploration difficulty and tunneling risk.

During tunnel excavation, water hazards in faults, especially steep water rich faults, pose a serious threat to safe construction in some complex mountains, which leads to low economic growth and development in these areas. Direct current resistivity method, which has high resolution and sensitivity to the low resistivity body, is widely used to predict the water-bearing structures in the front of tunnel face. The current prediction models are based on the resistivity isotropic medium, however, the resistivity of water bearing fault is often anisotropic due to rock fracture. The prediction model neglecting the anisotropy is obviously inaccurate, which brings potential threats to safe construction. We develop a three-dimensional resistivity modeling for anisotropic media using unstructured finite element method. The algorithm is proved to be accurate by comparison of numerical results and analytical solutions for a whole-space model. Another classical anisotropic model also demonstrated the reliability of our code from a physical point of view. Then we propose a prediction equation to predict the position of a vertical fault with anisotropic resistivity in the front of tunnel face by the finite element simulations. The parallel Monte Carlo method is used to test and evaluate the quality of our prediction equation by simulations of 10000 random vertical fault models, results counted by the histogram showed 85.36% of the results are predicted within 10% of the error. Besides, 93.17% of the results are predicted within 15% of the error using the equation for random faults with 75 degree dip angle, which shows that our prediction model can effectively forecast steeply dipping water-rich faults or fracture zones. INDEX TERMS Finite element method, Monte Carlo method, Advanced detection in tunneling, steep waterbearing faults, resistivity anisotropy.

No abstract

3D direct current long electrode modelling 591 electrode source is closer to the anomaly. This new method using the long electrode source greatly improves the resolution of the anomaly, which is of great significance for the safety of the tunnel construction.