Accurate dynamic modeling of landslides could help understand the movement mechanisms and guide disaster mitigation and prevention.Discontinuous deformation analysis(DDA)is an effective approach for investigating land...Accurate dynamic modeling of landslides could help understand the movement mechanisms and guide disaster mitigation and prevention.Discontinuous deformation analysis(DDA)is an effective approach for investigating landslides.However,DDA fails to accurately capture the degradation in shear strength of rock joints commonly observed in high-speed landslides.In this study,DDA is modified by incorporating simplified joint shear strength degradation.Based on the modified DDA,the kinematics of the Baige landslide that occurred along the Jinsha River in China on 10 October 2018 are reproduced.The violent starting velocity of the landslide is considered explicitly.Three cases with different violent starting velocities are investigated to show their effect on the landslide movement process.Subsequently,the landslide movement process and the final accumulation characteristics are analyzed from multiple perspectives.The results show that the violent starting velocity affects the landslide motion characteristics,which is found to be about 4 m/s in the Baige landslide.The movement process of the Baige landslide involves four stages:initiation,high-speed sliding,impact-climbing,low-speed motion and accumulation.The accumulation states of sliding masses in different zones are different,which essentially corresponds to reality.The research results suggest that the modified DDA is applicable to similar high-level rock landslides.展开更多
为解决危大工程中吊装作业安全管理的问题,基于深度学习构建目标检测算法(You Only Look Once version 5,YOLOv5)网络模型,针对进入吊装作业区域内人员的防护装备进行多目标融合检测,并对吊钩在施工过程中的状态进行检测。在原始的检测...为解决危大工程中吊装作业安全管理的问题,基于深度学习构建目标检测算法(You Only Look Once version 5,YOLOv5)网络模型,针对进入吊装作业区域内人员的防护装备进行多目标融合检测,并对吊钩在施工过程中的状态进行检测。在原始的检测网络模型中引入4种注意力机制,并通过5种训练模型的结果对比分析,进而选择卷积块注意力模块(Convolutional Block Attention Module,CBAM)最优模型。优化后的检测模型对安全帽的平均识别精度达86.5%,对反光衣的平均识别精度达83.0%,对吊钩的状态识别精度达92.0%。将训练好的人员检测模型和吊钩检测模型打包成exe执行文件,应用到施工安全管理人员的中控平台,可帮助管理人员更好地判断吊装作业的工作情况,进而及时进行风险管控。展开更多
基金supported by the National Natural Science Foundations of China(grant numbers U22A20601 and 52209142)the Opening fund of State Key Laboratory of Geohazard Prevention and Geoenvironment Protection(Chengdu University of Technology)(grant number SKLGP2022K018)+1 种基金the Science&Technology Department of Sichuan Province(grant number 2023NSFSC0284)the Science and Technology Major Project of Tibetan Autonomous Region of China(grant number XZ202201ZD0003G)。
文摘Accurate dynamic modeling of landslides could help understand the movement mechanisms and guide disaster mitigation and prevention.Discontinuous deformation analysis(DDA)is an effective approach for investigating landslides.However,DDA fails to accurately capture the degradation in shear strength of rock joints commonly observed in high-speed landslides.In this study,DDA is modified by incorporating simplified joint shear strength degradation.Based on the modified DDA,the kinematics of the Baige landslide that occurred along the Jinsha River in China on 10 October 2018 are reproduced.The violent starting velocity of the landslide is considered explicitly.Three cases with different violent starting velocities are investigated to show their effect on the landslide movement process.Subsequently,the landslide movement process and the final accumulation characteristics are analyzed from multiple perspectives.The results show that the violent starting velocity affects the landslide motion characteristics,which is found to be about 4 m/s in the Baige landslide.The movement process of the Baige landslide involves four stages:initiation,high-speed sliding,impact-climbing,low-speed motion and accumulation.The accumulation states of sliding masses in different zones are different,which essentially corresponds to reality.The research results suggest that the modified DDA is applicable to similar high-level rock landslides.
文摘为解决危大工程中吊装作业安全管理的问题,基于深度学习构建目标检测算法(You Only Look Once version 5,YOLOv5)网络模型,针对进入吊装作业区域内人员的防护装备进行多目标融合检测,并对吊钩在施工过程中的状态进行检测。在原始的检测网络模型中引入4种注意力机制,并通过5种训练模型的结果对比分析,进而选择卷积块注意力模块(Convolutional Block Attention Module,CBAM)最优模型。优化后的检测模型对安全帽的平均识别精度达86.5%,对反光衣的平均识别精度达83.0%,对吊钩的状态识别精度达92.0%。将训练好的人员检测模型和吊钩检测模型打包成exe执行文件,应用到施工安全管理人员的中控平台,可帮助管理人员更好地判断吊装作业的工作情况,进而及时进行风险管控。
