The Parameter Optimization Mehod For Underground Runoff Modulus

Forecasting of tunnel inflow is a difficult problem. This paper summarized some traditional methods for forecasting the long tunnel inflow, and analyzed their advantages and disadvantages. In addition, this paper compared six long tunnel water inflow forecasting results with measured results. The results show that the underground runoff modulus method is more accuracy, simpler and easier to implement than other traditional methods; but because of great subjective preferences, and sometimes also expected a larger deviation.The parameters of underground runoff modulus include:underground runoff modulus and impact width. Underground runoff modulus is usually obtained by field measurement; this paper proposed some optimization methods for determination of underground runoff modulus based on experience in engineering, as a selection of underground runoff modulus reference for future. The determination of impact width usually relies on the experience; the paper selected typical sections of the tunnel that had been built, then inversed the actual impact width. By comparing the inversion results with the predicted results, this papershows quite a difference. Therefore, to optimize the choice of the width is a priority in parameter optimization.This paper selected three factors of impact width, including surface conditions, the rock properties and geological structure. This paper analyzed the AHP weight of each factor, and established a grading score model to evaluate these factors. The surface condition, rock properties and geological structure of the selected sections were evaluated. This paper established a BP neural network optimization model for the width of select, which used the above evaluation results as samples.The application results show that the BP neural network optimization has a certain reference value, as it can improve the accuracy of the width selection.