Subsidence estimation utilizing various approaches – A case study: Tehran No. 3 subway line

[Display omitted] ► Find a fitting model to predict Subsidence and convergence in sections of the tunnel in Tehran No. 3 subway line. ► Applying empirical, numerical and SPSS Statistical Equation methods, also the neural network method. ► The neural network and statistical equations were obtained based on data from 51 subway tunnels worldwide. ► Suggesting neural network (SPSS software) methods to predict. The aim of this study is to analyze the subsidence and convergence and also to find an appropriate model to predict the behavior of the tunnel in Tehran No. 3 subway line. Empirical methods are employed to determine the variation of radial displacements along the longitudinal direction of a tunnel when subjected to a hydrostatic in situ stress field. The deformation in these sections is also determined by using numerical analyses. In addition the neural network method is utilized by two forms of advancing and back-propagation (BP) approaches. The data pertinent to the optimum network were obtained from 50 subway tunnel in Iran and Turkey which have been constructed by the NATM method with similar soil properties. The obtained result of empirical relationship of Peck (1969), Ranken (1987), Attewell et al. (1986) and Statistical Package for Social Sciences (SPSSs) compared with monitoring data indicate a very good agreement. In both SPSS and neural network methods the actual error and correlation coefficients are suitable.