• A rheological solution embedding damage tensor of layered rock mass was proposed. • A multi-level deformation prediction model for multi-layer supporting was established. • Theory and intelligent decision technologies for installation timing were proposed. Multi-layer supporting structures have excellent flexibility in dynamically regulating the stress release of surrounding rock. Therefore, accurately determining the installation timing and combination scheme is of great significance for preventing and controlling severe large squeezing deformation. To systematically sort out the principles for determining the installation timing, the evolution law of mechanical mechanism of the double-layer primary supports in Maoxian tunnel was analyzed. A theoretical analytical method for determining installation timing was proposed by combining the viscoelastic-plastic creep constitutive model, which considers the damage tensor of layered rock mass, with the elastoplastic model during the construction influence period and the failure model of the supporting structure. Based on the continuity requirement of deformation under a multi-layer supporting system, an artificial intelligence algorithm was used to establish a multi-level deformation prediction model, assisting in determining the installation timing. The results indicated that for continuous medium surrounding rock and large squeezing deformation, the concept of flexible support followed by strong support and appropriately delaying the installation timing, can significantly reduce the surrounding rock pressure. By adopting the above concept, the maximum surrounding rock pressure of Maoxian tunnel can be reduced from 950.7 kPa to 665.9 kPa. The applicability of the proposed theoretical analysis method and multi-level deformation prediction model was verified through engineering examples. The maximum analytical solution error of the average surrounding rock pressure in Maoxian tunnel was only 41.2 kPa, while the average deformation prediction error was only 8.3%.
Zhou et al. (Tue,) studied this question.