摘要
隧洞开挖后为了得到准确的围岩参数来指导支护设计,利用现场隧洞收敛监测数据,采用Phase2联合MATLAB的BP神经网络工具箱进行围岩参数反演。结果表明:神经网络训练误差较小,有限元反演方法得到的参数能较好地反应围岩开挖后位移变化的规律,变化规律与实测收敛数据一致,利用BP神经网络反演隧洞围岩参数方法可行。
After tunnel excavation,in order to obtain accurate surrounding rock parameters to guide the support design,Phase2 combined with MATLAB's BP neural network toolbox is used to carry out the back analysis of surrounding rock parameters based on the on-site tunnel convergence monitoring data.The results show that the training error of the neural network is low,and the parameters obtained by the finite element back analysis method can reflect the law of the displacement change after the excavation of the surrounding rock,and the variation is consistent with the measured convergence data.The back analysis of surrounding rock parameter with BP neural network is feasible.
作者
张争
马杰
刘永智
石广斌
范新宇
ZHANG Zheng;LIU Yongzhi;SHI Guangbin;FAN Xinyu(PowerChina Northwest Engineering Corporation Limited, Xi'an 710065,China;Xi'an University of Architecture and Technology, College of Civil Engineering,Xi'an 710055,China)
出处
《西北水电》
2021年第2期55-58,共4页
Northwest Hydropower
关键词
开挖
蠕变
收敛观测
围岩稳定
BP神经网络
excavation
creep
convergence monitoring
surrounding rock stability
BP neural network method
作者简介
张争(1990-),男,陕西省渭南市人,工程师,主要从事水工结构和地下硐室稳定分析工作.
