Study On Rockburst Prediction And Control In Jiangbian Hydropower Station

In the study progress nearly a century, scholars put forward a variety of theories and research methods from the mechanism of rock burst. The mechanism of rock burst was very complex; any kind of current theory could not explain the mechanism enough. The occurrence of rock burst was closely related with engineering geology, so the suitable theory and research methods should be selected for the different projectsIn this paper, set the diversion tunnel of Jiangbian Hydropower Station for the study object, with the results of previous studies, adopted system analysis, the combination of on-site investigation, laboratory tests, numerical analysis and mathematical evaluation, conducted the study of the rock burst prediction and prevention.1. Summarized that the geo-stress condition was the most important factor of rock burst through geological collected data and investigation; through the comparison of geo-stress conditions, theoretical criteria and the actual rockburst, determined the applicability of theoretical criteria.2. Took advantage of laboratory and field tests to determine the relevant mechanical parameters of rock samples near the typical section of the tunnel, including:rock samples uniaxial compression failure mode, the elastic deformation energy index and the rock integrity factor; the experiments proved that the diversion tunnel surrounding rock is hard and brittle, with a strong rock burst tendency.3. Combined with survey data and geo-stress measurement data, numerical model of the study area was established by using the ANSYS and flac3D; through Radial Basis Function(RBF) Neural Network, conducted the initial geo-stress inversion of study area, to obtain data related to initial geo-stress; the secondary stress field data was obtained through numerical calculation. And then the rock burst tendency of typical sections was evaluated by using the stress criteria.4. Rock burst is a highly nonlinear problem, so non-linear ways and means are needed. In view of limitations of the single factor criteria for predicting rock burst, the extension comprehensive evaluation model was adopted to predict rockburst of typical sections in division tunnel, and got rock burst intensity. Based on RBF neural network prediction model, the rock burst occurrence probability had been got and the prediction results were in line with the actual situation5. Using of rock burst prediction results of typical sections in division tunnel, the corresponding support programs were put forward to guide construction.