Rheology Parameters Back Analysis Of Soft Rocks Of Coal Deeply Buried Chamber Based On BP Neural Network

In deep mining project, rock in the complex environment of high temperature and pressure, and then mine roadway excavation, support and maintenance issues even more prominent. Surrounding rock of deep mine roadway show the characteristics of large deformation, high field stress and long-term sustainable rheology. Construction operations face more difficult tasks. So the exploration of creep damage law of surrounding rock in high-stress environment is the urgent need for solving important technical problems in deep mining engineering, developing coal construction and exploitation.The timely deformation have major impact on stability of surrounding rock in the construction and safety of the linning material in operation.Therefore, the rheological parameters of rock mass and the initial stress parameters are important mechanical parameters in the design and construction.Because of the discontinuing, heterogeneity and size effects of rock mass.The rheological parameters obtained by conventional laboratory and field test have certain limitations.So displacement back analysis are expected to address these deficiencies.This paper,based on the background of the actual coal mine roadway engineering,difference method and evolutionary BP neural network, conbined with field test dataes,carried out study on displacement back analysis. This paper simulated excavation by steps.Based on orthogonal design, The corresponding neural network samples are obtained. Based on Matlab neural network toolbox, this paper learn and train rock mechanical parameters, optimize network structure and parameters, eventually generalization ability is improved.Training and testing the network showed that system retrieval accuracy up to higher precision. Finally, an engineering example verify the accuracy and reliability of the displacement back analysis method.Figure [36] table [16] reference [71]...