开采地面沉陷预测的神经网络方法研究  

Studies of An Artificial Neural Network Approach to Predicting Mining Induced Surface Subsidence

作　　者：丁德馨[1] 毕忠伟[1] 王卫华[1] 

机构地区：[1]南华大学建筑工程与资源环境学院,湖南衡阳421001

出　　处：《南华大学学报（理工版）》2002年第3期1-5,共5页Journal of Nanhua University(Science & Engineering)

基　　金：湖南省自然科学基金重点资助项目 (0 1JJY10 0 4);湖南省教育厅重点资助项目 (0 1A0 15 ) .

摘　　要：应用神经网络理论 ,建立了开采地面沉陷预测的BP神经网络模型 ,采用开采地面沉陷实测数据对网络进行了训练 ,以此训练好的BP神经网络模型来描述开采地面沉陷与其影响因素之间的非线性映射关系 ,并采用测试样本对模型进行了测试 ,结果表明 ,网络的预测性能是令人满意的 .An artificial neural network approach is proposed to predict mining induced surface subsidence in this paper. The artificial neural network model is trained by using practical mining induced surface subsidence data. The trained network is used as a nonlinear image relationship between mining induced surface subsidence and its influential factors. Furthermore,the network is used to make prediction for the existed examples and the prediction was compared with the measurements. The comparison results show that the proposed approach is valid and applicable in predicting mining induced surface subsidence.

关 键 词：开采地面沉陷 神经网络 BP算法 非线性映射 模型 预测性能 

分 类 号：TD325.4[矿业工程—矿井建设]