The Study Based On Artificial Neural Network Of Classification Of Surrounding Rock And Optimization Of Supporting Parameters

The job will work at highland stress and very poor rock conditions environment with the increase of mining depth, which may lead to the whole tunnel crush, floor heave, dome subsidence, two sidewalls convergence, large deformation damage like saddle-shaped and so on, all of these require new technique in tunnel supporting. This thesis takes the roadway supporting of QianYingzi coal mine in WanBei coal group as the engineering background, the study of rock classification of deep well and optimized design of supporting were carried out, it have important Practical significance.Many factors influence the stability of surrounding rock classification, mainly include natural and exploitation technical these two factors. Considered the characteristics of coal measure strata. The author select BQ classification as a basis for the samples of classification of surrounding rock. The software system of rock classification of deep well and optimized design of supporting was developped, which based on BP neural network and used MATLAB neural network toolbox function, The basis of collecting much engineering geological information HuaiNan, HuaiBei mining area, as well as the corresponding support parameters.The software system mainly consists of two functions. First, it can achieve classification of stability of surrounding rock; Second, it can implmentation realize optimization and decision-making of supporting parameter in mine roadway. At the XiYi return aircourse and the 3212 tunnel of QianYingzi coal mine, the outcome of the decision-making of application of the software design, after the scene by the bolt force, and deep displacement and broken zone monitoring, it shows that this software is reasonable to decision-making. The software in the successful application of the QianYingzi coal mine, can obviously shorten the design supporting time, reduce the percentage of products sent back for repair and destabilization, The software is humanized, easy to operate, and the required information, most of the information were inputed by the drop-down selection; we can see that the software have some application value.