Application of NLGC inversion method in exploration of extra long railway tunnels

Because of the complex geological conditions of railway extra-long tunnels, the accuracy and speed of identifying complex lithology and fracture structures in conventional linear inversion of magnetotelluric data can not always meet the requirements of investigation. In this paper, the nonlinear conjugate gradient method was used for inversion imaging, which overcame the limitation of linear iterative inversion. For the non-quadratic minimization problem, the direct solution was carried out, which significantly reduced the computational workload and improves the imaging speed. Through practical application in Dajinshan Tunnel of Yumo Line Extra-long Tunnel, the locations of many faults in the tunnel were quickly and efficiently identified, which played a good guiding role in the later geological work. The results show that the inversion of stochastic non-linear conjugate gradient method can play an ideal role in the exploration of extra-long railway tunnels. © 2020, Central South University Press. All rights reserved.