Reliability Analysis And Design Optimization Of Tunnel Excavation Considering The Ground-Support Interaction

With the sustained and rapid development of the national economy,a variety of major infrastructure construction urgently need to build a large number of tunnels,underground caverns and other rock underground engineering.Due to its environmental complexity and subject to a variety of factors,there exist a great deal of uncertainty and variability in underground rock engineering.The reliability method based on probability statistics can quantitatively consider the impact of various uncertain factors and the variability of rock mechanical parameters,the actual load and support structure etc.,and thus has increasingly gained attention and concern in tunnels and other underground engineering.In this paper,the reliability analysis and evaluation of the deformation and stability of the surrounding rock in the underground engineering have been studied;the multi-failure modes and its corresponding performance functions of the tunnel considering the ground-support interaction were established;The first-order reliability method and second-order reliability method based on the moving least squares method for tunnel reliability analysis were proposed;The system reliability analysis method for tunnel system based on conditional probability in Excel spreadsheet environment was developed;The method for tunnel reliability analysis considering the spatial variability of rock mass parameters was established;The probabilistic optimization design model of tunnel support system based on response surface method was established.According to the above researches,the following main achievements and understandings are obtained:(1)Based on the convergence-confinement method,the interaction mechanism between surrounding rock and support structure in tunnel excavation was systematically expounded.According to the possible failure modes of the tunnel,a variety of performance functions for the reliability analysis of the stability of the tunnel considering the ground-support interaction were established.(2)A new algorithm for tunnel reliability analysis based on moving least squares method was proposed by using uniform design.By sampling with uniform design,the moving least squares based response surface was constructed directly in U-space(standard Gaussian space).The accurate reliability results can be then obtained by the iterative procedure,where the response surface is continuously updated.Compared with other response surface models,the new algorithm not only has high computational accuracy,but also requires little computation and simple determination of internal parameters,which makes the reliability analysis easy and convenient.(3)A new method for calculating the system probability of failure based on conditional probability was proposed for the tunnel system with multiple correlated failure modes.The "common source random variables" can be identified and defined by introducing "DS class model",and the correlated failure modes event is transformed into conditional independent failure modes event.With the aid of the conditional probability,the system reliability analysis was transformed into the reliability analysis of the single failure mode,and a simple,convenient and transparent procedure for system failure probability calculation based on Excel spreadsheet was developed.(4)Based on the interpolation autocorrelation and response surface method,a new reliability analysis algorithm,which can consider the spatial variability of rock mass parameters was proposed.The random field was discretized into a series of random variables at control points by interpolation autocorrelation,and the stochastic reliability analysis was then carried out by using the first-order reliability method and response surface.The impact of the spatial variability of the rock mass parameters on the failure probability of the tunnel was explored.Moreover,the influence of the scale of fluctuation of the rock mass parameters on the failure probability of the tunnel convergence was investigated,and finally the influence of the interval between the control points in the interpolation autocorrelation on the results of reliability analysis was also discussed.(5)The reliability-based probabilistic design optimization model was proposed for the tunnel support system with multiple failure modes.The reliability index approach(RIA)and the performance measure approach(PMA)were employed separately to solve the reliability-based design optimization by using the response surface,and the efficiency and accuracy of the two methods were compared.The influences of the quality of rock mass,the coefficient of variation of the design variables and the target reliability indices on the probabilistic design optimization were investigated.