Investigation on the deformation mechanism of the full-section tunnel excavation in the complex geological environment based on the PSO-BP neural network.

In full-section excavated tunnel construction, the abnormal deformation of the surrounding rock, such as arch cracking and surrounding rock collapse, is often caused by insufficient timely support and over-excavation, which brings serious consequences to the project. In this paper, the geomechanical model is established to investigate the influence of different indexes on surrounding rock deformation. The deformation of surrounding rock can be divided into three stages, among which the earlier deformation caused by the construction in front of the monitoring section in the 1st stage accounts for 18% of the total deformation. The deformation caused by tunnel excavation in the second stage occupies 72%, and the lagging deformation in the third stage accounts for 10% of the total deformation. Furthermore, the critical indicators are screened out by the rough set algorithm. Finally, the prediction model of surrounding rock deformation of the full-section excavated tunnel is established based on the PSO-BP neural network. The validation results indicated that the prediction results from the integration model are in good agreement with the field observation results, and the prediction efficiency and accuracy are better than those of the numerical simulation and the single BP neural network. This study fully used tunnel monitoring data and proposed a deformation prediction model by investigating the correlation characteristics of the deformation data, which provides an essential reference for similar engineering research. [ABSTRACT FROM AUTHOR]