Analysis On Support Optimization And Reliability In Expansion Excavation Of Extra-large Section And Small Interval Tunnels From Two To Four Lanes

With the rapid development of transportation construction, and the constraints offactors such as the geological and topographical conditions, and general route type,tunnels with large section and small interval are gradually increasing, thus bringingabout the emergence of tunnel expansion excavation projects. At present, researcheson the expansion excavation of large section and small interval tunnels from twolanes to four lanes are very limited, the majority of which mainly concentrate on theconstruction mechanics. While those on the analysis of the optimization andreliability of tunnel support are rarely seen. On one hand, being faced with thecomplicated engineering geological conditions and complex stress capacities ofstructures, designers often artificially increase support parameters in order to ensurethe tunnel safety, which results in relatively higher safety coefficient and seriouseconomic waste so as to make it difficult to achieve safe and economic optimizationdesigns. On the other hand, the optimization design, to a certain extent, would reducethe stability of the tunnel. Therefore, the research project on the analysis of supportoptimization and support reliability in the expansion excavation of large section andsmall interval Tai Mo Shan tunnel from two to four lanes is of great significance.Taking the expansion excavation project of Tai Mo Shan tunnel from two to fourlanes in QuanXia section of ShenHai Highway as the engineering background, thispaper, based on the optimization back analysis and reliability theory, adopts themethods of field monitoring and numerical simulation to study in depth about thesurrounding rock deformation law, surrounding rock parameter inversion, initialsupport system optimization and support reliability during the expansion excavationof extra-large section and super-small interval tunnel by means of test design, finiteelement method, and intelligent analysis method. The main research contents andresults are as follows:1) Analyzing the measurement data from the site monitoring and studying thesurrounding rock deformation law when enlarging the two-lane tunnel with large section and super-small interval into four lanes so as to provide evaluating basis forthe accuracy of the numerical simulation.2) Developing the parameters inversion method of tunnel surrounding rock basedon genetic algorithm support vector machine (SVM) and identifying surrounding rockparameters for the purpose of providing references for numerical simulationparameters.3) Building optimization schemes of initial tunnel support by the combination ofthe GASVM-based response surface model and genetic algorithm and studying theoptimal supporting parameters of Tai Mo Shan tunnel with safety as the constraint,and capital-saving as the goal. The cost of the initial support in optimization schemewould be reduced by about32.5%compared with that in the original plan, so that theoptimization effect is obvious.4) Setting up three-dimensional finite element model of two-to-four-lane tunnelexcavation and comparing mechanical response properties of tunnel surrounding rockand support before and after the support optimization. Research shows that themechanical response laws of the surrounding rock and the stress distribution laws ofthe initial lining and the anchor and their influence to the tunnel remain roughly thesame. Besides, after the optimization of the tunnel support, the critical points in thetunnel deform vertically while the von miss stress of some increases (especially thatof abuts); the influence radius of the surrounding rock stress, the maximum von missstress of the initial lining and the maximum axial force of the anchor all improve ineach step sequence while the maximum function strength of the surrounding rock insome step sequences also heightens, and each corresponding growth is within5%.5) Establishing calculation method of tunnel support reliability based on thecombination of genetic algorithm support vector machine (SVM) and checking pointmethod to further analyze the reliability of optimized Tai Mo Shan tunnel support,which is5.62with displacement as criterion, hence indicating that the tunnel is safeand reliable after initial support optimization.