基于Adam优化算法的深度神经网络岩爆预测模型  

Rockburst Prediction Model of Deep Neural Network Based on Adam Optimization Algorithm

作　　者：田睿 孟海东 TIAN Rui;MENG Haidong(Institute of Mining Engineering,Inner Mongolia University of Science and Technology,Baotou,Inner Mongolia 014010,China)

机构地区：[1]内蒙古科技大学矿业研究院,内蒙古包头市014010

出　　处：《矿业研究与开发》2020年第11期40-46,共7页Mining Research and Development

基　　金：国家自然科学基金项目(51564038,51464036);内蒙古自治区自然科学基金项目(2018MS05037);内蒙古自治区高等学校科学研究项目(NJZY19127);内蒙古自治区博士研究生科研创新资助项目(B20171012702)。

摘　　要：岩爆是未来深部资源开采过程中必须要解决的关键科学问题之一。为准确可靠地预测岩爆灾害,提出一种基于Adam优化算法的深度神经网络(DNN)岩爆预测模型(Adam-DNN)。该方法利用国内外318例已有工程实例数据,通过深度学习技术建立预测模型。在考虑岩爆产生的内外因基础上,选取洞壁围岩最大切向应力与岩石单轴抗压强度比σθ/σc、岩石单轴抗压强度与抗拉强度比σc/σt和岩石弹性能量指数Wet组成岩爆预测指标体系。通过对锦屏二级水电站和冬瓜山铜矿进行岩爆预测的工程实例分析,验证了模型的有效性和正确性。研究结果表明:所提出的模型预测准确率达95%以上,可为类似工程的岩爆预测提供科学依据。Rockburst is one of the key scientific problems that must be solved in the future deep resource exploitation.In order to accurately and reliably predict rockburst disasters,a rockburst prediction model using deep neural network(DNN)based on Adam optimization algorithm(Adam-DNN)was proposed.Based on the data of 318 engineering cases at home and abroad,the prediction model was established by deep learning technology.On the basis of the internal and external factors of rockburst generation,the ratio of the maximum induced tangential stress on the boundaries of tunnels or caverns and the uniaxial compressive strength(σθ,σc),the ratio of the uniaxial compressive strength and tensile strength(σc,σt),and the elastic energy index of rock(Wet),were chosen to form the rockburst prediction indexs system.The validity and correctness of the model were verified by analyzing the engineering examples of the rockburst prediction of Jinping II Hydropower Station and Dongguashan Copper Mine.The research results show that the prediction accuracy of the proposed model is over 95%,which can provide a scientific basis for rockburst prediction of similar projects.

关 键 词：岩石力学 岩爆 岩石地下开挖 深度神经网络 ADAM 

分 类 号：TD311[矿业工程—矿井建设]