Study On Construction Monitoring And Stability Of Surrounding Rock For Neighborhood Highway Tunnel Of Large Span

With the fast development of highway in our country, highway tunnels are constructed more and more. As a result of special geological and topographical limitations, space of tunnels cannot satisfy with general technical codes. When we can not solve the problems above with separate tunnels, multiple–arch tunnels appear. But the construction of multiple–arch tunnels is comparatively complex and needs a long time. Also, waterproofing of multiple–arch tunnels is difficult and their investment is large. For solving this problem, neighborhood tunnels which are a new structure style appear. Compared with railway tunnels, the span of highway tunnels is large and their flat ratio is low. Method of excavation is branch excavating. From the point of safe, the risk is large with the space of tunnels smaller and smaller and the span larger and larger. Therefore, it is important to carry on the systematic advanced research on the surrounding rocks stability of neighborhood highway tunnels of large span.Firstly, the results of monitoring measurement in-situ are analyzed and the laws of peripheral convergence displacement, velocity of rock surrounding displacement, ground surface settlement, contact stress, axial force of anchor bolt and the monitoring of blasting vibration are carried out. The results of measurement show that: peripheral convergence displacement, velocity of rock surrounding displacement and velocity of blasting vibration are in the range of codes; Stress in the rock pile is comparatively complex. Contact stress and axial force of anchor bolt in the rock pile are larger distinctly. But there is large safe accumulation in the anchor bolt of designingSecondly, based on displacement back-analysis approach of numerical simulation and neural network, Orthogonal experimental design is adopted to constitute study examples. Accuracy of forecast of net is ensured and time of test is reduced. Artificial neural network is trained and simulated by ANN TOOL in MATLAB6.5. Errors are controlled in 5% by test. The trained network is used to intelligent recognition of surrounding rocks parameters in cross section K1+760. This can provide surrounding rocks parameters which accord with engineering for finite element analysis of neighborhood tunnels of large span.Thirdly, the paper at first studies on the stability of tunnel with no support by numerical simulation for researching stress characteristic and stability of neighborhood tunnels of large span. In the condition ofⅢ～Ⅴclass surrounding rocks, depth from...