Rock Mechanics And Engineering Intelligent Analysis Method In Mine Surrounding Rock Control Application

Based on the rock mechanics and engineering intelligent method in the analysis of surrounding rock control application for the subject, first elaborated the intelligent computation algorithm of the BP neural networks and genetic algorithm, and combined with BP neural network genetic algorithm local search and global optimization characteristics of evolutionary neural network algorithm. Then by using the evolutionary neural network algorithm theory of mine surrounding rock control in roadway surrounding rock deformation in the roof pressure prediction, time series forecasting, and bolt support parameters design has carried on research and discussion, obtained the practical methods to solve the problem:According to the effect of roadway surrounding rock deformation factors analysis, the roadway surrounding rock deformation and influence factors between complex nonlinear relation, namely the roadway surrounding rock deformation in prediction of the problem can be transformed to determine the effects of roadway surrounding rock deformation in factors and moved closer to the nonlinear relationship between the amount of. On the influence of roadway surrounding rock deformation of the main factors were analyzed quantitatively, and the collection of a large number of roadway surrounding rock deformation in samples of data, using the evolutionary neural network to deal with nonlinear relation, through a network of learning and training, to obtain a stable forecasting roadway surrounding rock deformation in network structure, after the test sample test, network the structure has good practical value.Based on the artificial neural network for time series prediction theory of nonlinear analysis, the coal mine roof pressure using artificial neural network for time series prediction of roof pressure. The coal mine of Sanyuan wang zhuang working surface working resistance of support for time weighted processing, obtained the time series of pressure on the working face data, according to the time series nonlinear prediction theory, the time sequence of the working surface pressure data into the network training samples, namely through the front of a few data to infer the current data, based on the sample data for the learning and training, to obtain stable evolutionary neural network prediction network structure of roof pressure, through the test sample test, network structure on roof weighting prediction has better practical value.According to the present bolt support design methods and theoretical analysis, the current conventional bolting design method has great limitation. Through the analysis of bolt supporting in actual mining roadway parameters and influence of roadway surrounding rock stability factors of complex nonlinear relationship between, through the collection of a large number of bolting design successful engineering case as learning and training samples to train, using the evolutionary neural network to determine the support parameter and influencing factors of the stability of roadway surrounding rock nonlinear relationship between mining, finally get the roadway bolt support parameters and stable network structure.