基于GA-BP神经网络的采空区地表沉降预测模型

A Surface Subsidence Prediction Model for Underground Goafs Based on GA-BP Neural Network

为了提高采空区地表沉降预测准确性,选择上覆岩层弹性模量、泊松比、内聚力、内摩擦角、开采深度、采高、矿体倾角和采场尺寸共8项影响采空区沉降的指标进行研究,通过遗传算法(GA)优化BP神经网络,构建了GA-BP神经网络采空区地表沉降预测模型,对采空区地表沉降趋势初步预测与分析。模型解算结果表明,相比传统BP神经网络预测模型,GA-BP神经网络预测模型在预测精度、拟合性能和收敛速度方面都有所提高。

In order to improve the accuracy of surface subsidence prediction for underground goafs,eight indicators,such as elastic modulus of overlying strata,Poisson′s ratio,cohesion,internal friction angle,mining depth,mining height,dip angle of ore body and stope size,were selected for studing.By optimizing BP neural network with genetic algorithm(GA),a prediction model for goaf surface subsidence was constructed based on the GA-BP neural network,and was then adopted to preliminarily predict and analyze goaf surface subsidence.Model solving result shows that compared with the traditional BP neural network model,the GA-BP neural network model can have better prediction accuracy,fitting performance and convergence speed.