摘要
以新疆金川矿业京希-巴拉克采区南帮边坡为工程背景,针对震后边坡力学参数弱化及余震下边坡稳定性评估问题,提出了基于乌鸦算法(CSA)优化的BP神经网络模型(CSA-BP),用于震后边坡力学参数反演,并结合离散元法对余震状态下危岩体边坡稳定性进行评价。结果表明,CSA-BP模型反演震后边坡力学参数,定量揭示了岩体弱化特征;5级余震下边坡中上部凝灰质砂岩位移显著,x方向位移远超竖向(z方向)位移,边坡失稳以水平滑移为主。CSA-BP模型能通过参数-动力耦合机制精准定位高风险区,可为震后边坡防护提供理论支撑。
The south slope of the Jingxi⁃Barak mining area of Xinjiang Jinchuan Mining Industry was taken for study.As for the issues of weakening mechanical parameters of slopes after an earthquake and the stability assessment of slopes under aftershocks,a BP neural network model optimized by the crow search algorithm(CSA⁃BP)was proposed for the inversion of mechanical parameters of slopes after an earthquake.The stability of slopes with dangerous rock mass under aftershocks was evaluated by using the discrete element method.The results show that the CSA⁃BP model can quantitatively reveal the weakening characteristics of rock masses through the inversion of mechanical parameters.Under aftershock of a magnitude 5 earthquake,significant displacement occurs in the tuffaceous sandstone in the middle and upper parts of the slope,with horizontal(x⁃direction)displacement far exceeding vertical(z⁃direction)displacement,indicating that slope instability is dominated by horizontal sliding.The CSA⁃BP model can accurately identify high⁃risk zones through a parameter⁃dynamic coupling mechanism,providing theoretical support for slope protection after earthquakes.
作者
卢栋
富国凯
孙正军
代吉才
高晨翔
侯钦宽
LU Dong;FU Guokai;SUN Zhengjun;DAI Jicai;GAO Chenxiang;HOU Qinkuan(Xinjiang Jinchuan Mining Co.,Ltd.,Ili Kazakh Autonomous Prefecture 835000,Xinjiang,China;School of Resources and Civil Engineering,Northeastern University,Shenyang 110819,Liaoning,China)
出处
《矿冶工程》
北大核心
2025年第2期34-40,46,共8页
Mining and Metallurgical Engineering
基金
中国—中东欧国家高校联合教育项目(2021111)
国家自然科学基金(U1602232,52374157)
辽宁省重点科技计划项目(2019JH2-10100035)
中央高校基本科研业务专项资金(N2301005,N2301006)。
关键词
危岩体
边坡稳定性
岩石力学
参数反演
乌鸦算法
BP神经网络
地震响应
机器学习
unstable rock mass
slope stability
rock mechanics
parameter inversion
crow search algorithm
BP neural network
seismic response
machine learning
作者简介
卢栋(1977-),男,山西霍州人,高级工程师,主要从事基础地矿与矿产勘查工作。E-mail:18953932033@189.cn;通信作者:侯钦宽(1995-),男,河南安阳人,博士研究生,主要从事岩体稳定性评价与分析研究。E-mail:houqinkuan@qq.com。
