3D Geological Visualization Of Tunnel And Construction Decision Method For Inrush Water Based On Multi-Source Information Fusion And Its Application

With the rapid infrastructure development in the whole country,tens of thousands of kilometers of new tunnels in the next 1 0 years will be built in the field of highway and railway and dozens of large-scale water diversion projects.The water inrush disaster can be regarded as a world-class engineering problem.Therefore,in this paper,a prediction technology of tunnel water inrush and construction decision system are presented based on the Yuelongmen Tunnel.It can solve some key problems such as fusion analysis of seismic wave data,coupled modeling of geological bodies,three-dimensional fracture network simulation.The main research results of this paper are as follows:(1)A Kalman filtering data fusion technology is proposed based on tunnel seismic wave detection data.Combining with investigation data of the study area,borehole data and aerial data,regularization of multi?source data in three-dimensional geological modeling is realized.It provides data-based support for solving the contradiction between complex geological structure,diversification of modeling data and high accuracy requirements for three-dimensional modeling.(2)For the geological hazard of water inrush in Yuelongmen tunnel,constrained Delaunay triangulation technology is adopted to construct and couple the basic model of regional geological body;According to differential geometry and numerical discrete method,a geologic interface reconstruction technology based on cubic Multi-quadric radial basis interpolation function is proposed.It realizes the fast and accurate construction of geological interface model.Finally,a technology of coupling tunnel model with regional geological model based on surface-volume boolean operational algorithms.According to the results of modeling,the flooded area of Yuelongmen Tunnel is analyzed.(3)According to detection data of structural plane in the research area,a new stochastic-deterministic discrete fracture network model is constructed by combining with Monte-Carlo method,"parent-daughter"and "step-structure" modified model.A full-space parallel search algorithm with the advantages of depth-first search and breadth-first search is innovatively proposed.And it is applied successfully to the stochastic-deterministic DFN model.It solves the problem of fast and accurate search of the seepage path of the fracture network with numerous data and complicated logic structure.(4)The numerical calculation method is used to solve the water inflow of each section of Yuelongmen Tunnel based on stochastic-deterministic discrete fracture network and the principle of water flow balance.According to the monitoring data such as water inflow information,surrounding rock grade,water inflow pressure,horizontal convergence,vault settlement,and 3D geological model and construction database information,the extensional neural network is used to analyze and predict the corresponding construction decision.Through the on-site excavation construction measures,the accuracy of the method described in this paper is verified.which provides an effective way for the intelligent assistant decision-making research of tunnel construction.