The Application of BP Neural Network in Rock Parameter Determination of Mining Engineers

To guarantee the security of the ground construction in mine area, it is very important to forecast the deformation of the underground rock masses, especially the surface settlement. The deformation analysis result is affected by the determination of rock parameters greatly. The parameters can be obtained from lab tests of some soil and rock samples taken out from the stratum. But the most of the test samples are often get from upper layer stratum, the deep rock sample is difficult to get. Then it is hard to get an exact forecast result. The determination of rock parameters from the back calculation was taken here. Back-propagation (BP) neural network combined with finite element method (FEM) was adopted to obtain the rock parameters. Numerical simulation and analysis of the rock mass deformation and the ground settlement of the mine area were carried out in three kinds of exploitation state. The result shows that determination parameter from the back analysis of BP combined FEM is an efficient way.