Study On Construction Influences Of Shield Tunnel Excavation To The Neighboring Buildings

With the accelerating of urbanization speed in China, more and more cities buildunderground railways to reduce the traffic problems. Every method in tunnel excavatingcan inevitably cause soil layer deformation. In big cities, there are all kinds of buildingsstanding on the ground, and all types of pipelines laying underground. Therefore, how topredict and control ground settlements and building deformations caused by runnelexcavation is the key problem in tunnel design and construction.To discuss the influences on the neighboring structure during shield tunnelingconstruction, a 3D numeric model which includes a frame structure, piles, soil and tunnelis established. This paper studies the changing trend of the internal forces (axial forces,shear forces and moments of beams, columns and piles) of the structure in shield tunnelexcavation. And The sensitivities of the key parameters, such as depth of the tunnel,distance between structure and tunnel, and length of the piles to maximal surfacesettlement and differential settlement between piles are also analyzed in this thesis.Finally, using the theories of Artificial Neural Network (ANN), this paper studies areasonable prediction method of surface and foundation displacement. The main work ofthe paper are listed as follows,1. The influences on the neighboring building caused by shield tunneling. In thispaper, a 3D numeric model which includs shield tunneling and neighboring structurewas established to discuss the influences on the internal forces and deformations of thestructure caused by dynamic construction process.2. Study on the sensitivity of the key parameters. In tunnel excavating, different projecthas different tunnel burying depth, different distance between building and tunnel anddifferent foundation type, this paper studies the sensitivity of these parameters on themaximal ground settlement and differential settlement between piles.3. Prediction of the surface and structure deformation based on ANN. Modelestablishment and calculating process of FLAC^3D is complex and specified. Using theadvantages of Back Propagation (BP) Neural Network in solving the nonlinear anduncertain problems, a BP NN is made in this paper to predict the deformations of the ground and structure, which provides usable references to engineering practice.