An approximate ANN-based solution for convergence of lined circular tunnels in elasto-plastic rock masses with anisotropic stresses

An approximate solution is presented for predicting the convergence of lined circular tunnels in elasto-plastic rock masses with anisotropic in situ stresses. The basic idea which led to derivation of the solution is function approximation using Artificial Neural Networks (ANNs). A suitable data base including 2500 convergence values is prepared by innovative combination of Design of Experiments technique and Finite Difference Method. Then, an ANN with optimum architecture and appropriate training algorithm is used to learn the underlying phenomena involved in the problem from the data. The explicit form of the solution for calculation of convergence is elicited from the trained ANN. Subsequently, the ANN-based solution is validated against analytical and numerical solutions. Finally, the limitations associated with the solution are discussed.

► ANNs were successfully used to derive an explicit solution for the tunnel problem.
► The solution was validated against available analytical and numerical solutions.
► It is possible to solve currently un-solved problems with the presented methodology.