巷道围岩变形的神经网络模型

巷道围岩变形是巷道稳定性最直接判据,预测巷道围岩变形为选择支护方案提供有力的依据.巷道围岩变形的影响因素非常复杂,具有不确定性,而神经网络不仅能考虑定量因素,而且能考虑定性因素,因此神经网络适用于解决变形预测问题.系统分析了影响巷道围岩变形的因素,构建了巷道围岩变形预测的神经网络模型,经过改进的网络模型具有较好的收敛性和稳定性.

Model of Neural Network Forecasting the Deformation of Surrounding Rock in Tunnel

The deformation of the surrounding rock in tunnel directorly indicates the stability of tunnel. forecasting of the deformation can provide us a powerful evidence for tunnel support. The factors affecting the deformation of surrounding rock in tunnel are complicated and uncertain, however, neural network can not allow for the quantitative but the qualitative factors, this ability of neural network is suitable to forecast the deformation of the surrounding rock in tunnel. In this paper, factors affecting the deformation of the surrounding rock in tunnel are analyzed, the model of neural network forecasting the deformation of the surrounding rock in tunnel constructed, and improving learning algorithms has made it achieve a fast convergence speed and better stability.