Roadway Support Optimization by Improved BP Neural Network and Numerical Simulation

摘要

Based on the analysis of the influence factors of the stability of roadway, we firstly collected the roadway support parameters of some roadways with good supporting effect and used them as the training samples of BP neural network. Then, we simulated the deformations of the forecasting samples and compared with the actual results to examine the accuracy of the supports scheme. At last, the optimal support scheme of No.3306 haulage roadway at Xinan Coal Mine, China was predicted by using improved BP neural network and it was verified by using the FLAC3D numerical simulation. The results showed that the model established by improved BP neural network has fast convergence, high accuracy and good stability, and it could effectively predicted the roadway deformation and provided scientific basis for the supporting design of roadway.