The Study On The Intelligent Evaluation Of Stability Of The Whole Tunnel Surrounding Rock Based On Numerical Experiments

With the implementation of the national western development strategy, national highway and railway construction focus on the westward, the western mountainous landform has created lots of complex geological situation of super long tunnel. As the "New Austrian method" has the characteristic of supporting structure is reliable, low cost and simple construction technology,it is widely used in tunnel engineering construction. the core of "new Austrian method" is using the carrying capacity of surrounding rock to design and construction the tunnel support structure. This thesis using the uniform design method, the finite difference theory and strength reduction method, the neural network theory and GIS, etc, based on the Baishiyi tunnel geological survey data to evaluate the stability of surrounding rock of the tunnel through, then using GIS for visual expression. This thesis made a explore of using GIS combined with neural network to evaluate the whole tunnel stability, the tunnel through stability evaluation chart has certain reference value for the tunnel construction with "new Austrian method", the main progress of work done in this paper are as follows:(1) Based on the study of the theory of the surrounding rock stability, to get the factors which influence the stability of surrounding rock, and ultimately selected seven parameters to express the impact factor, and the seven parameters as the input of neural network. The seven parameters are: elastic modulus, poisson ratio, internal friction Angle, cohesive force, porosity, permeability coefficient and tunnel buried depth. After study select the safety factor as the evaluation index of the stability of tunnel surrounding rock, as well as the output of the neural network.(2) Based on uniform design method evenly divide the seven parameter to 30 level, query U30 *(3013) selection table known as 1, 2, 3, 4, 6, 12, 13, to lookup U30 *(3013) design table to get the sample of impact factor, and the error meets the requirements.(3) Baishiyi tunnel model was established by applying the FLAC3 D, according each set of sample table data input to the related parameters of FLAC3 D, considering the influence of seepage on the stability of surrounding rock, to calculate the seepage and seepage not conditions, and to study the performance status of the tunnel after the two conditions, it is concluded that the seepage has a great influence on the tunnel vault displacement. Using dichotomy to make the surrounding rock strength reduction theory come true, calculated 30 groups of tunnel surrounding rock stability safety factors under the action of seepage, and got the neural network training sample table.(4) Study of BP neural network algorithm, to establish three layers BP neural network structure, and designed of the seven number of hidden layer nodes structure of neural network, using the MATLAB to realize algorithm of neural network, after normalization process of the training sample import neural network computation to calculate, the establishment of a reasonable mapping rules. Regression fitting curve shows that the simulation results is ideal, and meets the requirements.(5) In ARCGIS established seven factors of raster data layer in the study area, and reclassified to form the stability zoning map, through the spatial analysis function of GIS in the study area set up 52272 fish outlets, each dot represents the location of research area, using a dot file enterprising seven impact factor value, and extract the results of the export, import to the designed BP neural network to forecast stability of surrounding rock after data processing, the forecasted value import the dot file form new dot files, then converted the safety factor in the dot file to raster layer, after reclassification can get stability evaluation figure of surrounding rock in the study area. Finally through the grid cutting command can get the figure of stability of surrounding rock in the total length tunnel evaluation.(6) Using the fuzzy hierarchy comprehensive evaluation method to study the stability of surrounding rock at coal seam, the results are consistent with the evaluation results, demonstrates the reliability of intelligent evaluation results.Through research, this paper combines neural network with the GIS study the stability of surrounding rock has bigger advantage, the whole stability evaluation figure which gained form this study make the engineers for the stability of surrounding rock condition be clear at a glance, it has certain guiding significance to the design and construction of the tunnel with the "new Austrian method".