Feedback analysis of tunnel construction using a hybrid arithmetic based on Support Vector Machine and Particle Swarm Optimisation

Feedback controls for tunnel construction require identifying geomechanical parameters and adjusting supporting parameters, both of which are optimisation problems. This paper presents an integrated optimisation method for the feedback control of tunnel displacement; it combines the Support Vector Machine (SVM), particle swarm optimisation (PSO) and numerical analysis methods. Initially, the nonlinear relationship between parameters and displacements is efficiently represented by SVM. Numerical analysis is then used to create training and testing samples for SVM recognition. PSO is used to search on the parameters of SVM and to obtain the geomechanical and support parameters. A case study is provided to verify the proposed methodology. This study provides an alternative means for adjusting tunnel construction schemes based on field observations and the optimisation of quantitative integration.

Research Highlights
► The method includes both parameter identification and scheme optimization.
► It is a real time, quantitative method for tunnel construction.
► SVM was capable of accurately describing the nonlinear relation of tunnel system.
► The global optimization based on PSO-SVM has quick convergence speed.