关键词: 深埋软岩隧道, TBM隧道, 超前管棚支护, 支护参数优化, Kriging算法, 多目标遗传算法

Abstract: To address the problems of TBM jamming, support failure, and tunnel clearance violation caused by large deformation of surrounding rock in deep-buried soft rock TBM tunnels, a case study is conducted on the main tunnel upstream of the Branch No.7 of the Xianglushan Tunnel in the Central Yunnan Water Diversion Project, and the nonlinear influence characteristics of four pipe roof parameters—pipe diameter, pipe length, circumferential spacing, and arrangement angle range—on crown settlement and maximum axial stress of the pipe roof are systematically quantified and analyzed. Next, a multi-objective optimization model is established to achieve the dual objectives of minimizing crown settlement and pipe roof stress. The central composite design is adopted to create the sample space, and 25 sets of sample points are designed for the four support parameters of the advance pipe roof. The finite difference method is used to conduct numerical simulations on crown settlement and maximum axial stress of the pipe roof, and a response surface model relating pipe roof support parameters to crown settlement and maximum axial stress is established. The results show that crown settlement presents different nonlinear variation characteristics at different distances from the tunnel face, and pipe diameter and circumferential spacing slightly affects the crown settlement. In the near-tunnel-face zone (0－2 m), crown settlement first increases and then decreases with the growing pipe length, while it decreases initially and then rises as the arrangement angle range expands. In the far-tunnel-face zone (8－10 m), crown settlement decreases first and then increases with growing pipe length, and rises initially before declining with the expansion of arrangement angle range. Taking the minimum crown settlement and axial stress as the optimization objectives, the multi-objective genetic algorithm is applied to optimize the support parameters. The optimal parameters are determined as follows: pipe diameter of 180 mm, pipe length of 24 m, circumferential spacing of 30 cm, and arrangement angle range of 180°. Compared with the original support scheme, the crown settlement is reduced by 19.37%, and the maximum axial stress of the pipe roof is decreased by 66.97%. Field monitoring data are used to verify the response surface model. The relative error of numerical simulation for crown settlement is less than 8.75%, and the relative error of the response surface model is below 2.95%, indicating that the proposed response surface model possesses high calculation accuracy and favorable applicability.

Key words: deep-buried soft rock tunnels, TBM tunnels, advance pipe roof support, support parameters optimization, Kriging algorithm, multi-objective genetic algorithms