软岩隧道失稳塌方过程微震场与位移场透彻感知及综合预警理论研究

The overbreak disasters in the construction of soft rock tunnel has induced adverse consequences, such as casualties, financial losses and social influences. It has raise extensively concerns and hindered the safe construction of transportation and hydraulic tunnels. To prevent the overbreak disaster in the soft rock tunnel is critical but challenging due to the lack of the effective theory and technology for monitoring the instability process and identifying the overbreak disaster in the soft rock tunnel. In order to solve the issue, an academic method has been proposed, which is “to monitor the failure process of internal rocks using microseismic field, to sense the deformation status of surrounding rocks using deformation field, and to identify the overbreak disaster based on the multiple information and intelligent algorithms”. Several correlated tasks will be conducted, utilizing several approaches including theoretical study, numerical simulation, pilot plant testing and field testing. Firstly, the photonic crystal waveguide MOEMS based microseismic sensing method will be studied to accurately capture the broadband micro-seismic fields in the soft rock tunnel. Additionally, the deformation field sensing and correction theory based on the optic fiber embedded smart Geogrid will be investigated in detail so as to detect the deformation status of the surrounding rocks in the tunnel. Finally, the deep learning algorithms based overbreak disaster identification model will be built for the early-warning of disasters. The research will provide comprehensive theories, methods and techniques for the monitoring and early-warning of overbreak disasters in the soft rock tunnels.

软岩隧道塌方灾害造成了重大的人员伤亡、严重的经济损失与恶劣的社会影响，已成为交通与水工隧道安全建设的巨大挑战。本项目针对软岩隧道微震场信息难以精准获取、隧道围岩位移场信息难以全方位刻画、软岩隧道塌方灾害难以准确预警的关键问题，提出“以微震场全局监测岩体破裂过程，以位移场全面刻画围岩变形状态，以多场信息智能辨识软岩隧道塌方灾害”的学术思想，采用理论研究、数值模拟、模型试验、现场试验相结合的方法，提出基于光子晶体波导MOEMS的微震场传感新方法，攻克软岩隧道微震场高精度、宽频带传感难题，提出基于光纤内置式智能格栅材料的隧道围岩位移场立体感知新模式，突破隧道围岩位移场“点-线-面-体”全方位刻画理论与修正方法，最终，建立基于深度对抗学习的塌方灾害智能辨识与预警模型，形成软岩隧道失稳塌方过程透彻感知与塌方灾害综合预警成套理论、方法和技术，为软岩隧道塌方灾害监测与准确预警提供理论支撑与实用方法。

项目针对软岩隧道微震场信息难以精准获取、隧道围岩位移场信息难以全方位刻画、软岩隧道塌方灾害难以准确预警的关键问题，研究了梯形隔层双分量MOEMS振动传感机理与方法，研究了微震信号降噪与初至时刻提取技术，构建了微震传播速度智能反演与微震震源准确定位方法，实现了微震场信息的高效传感与准确获取。阐明了隧道围岩与光纤内置式智能材料的协同变形机制，研究了基于改进曲率递推的隧道围岩位移场重建方法，创建了基于形态自聚类分区分段的隧道位移场智能修正与高精度传感方法，实现了软岩隧道位移场全方位刻画。揭示了软岩隧道失稳塌方过程中微震场、位移场等多场信息的时空响应规律，研究了基于深度学习隧道异常目标智能识别方法，建立了基于神经网络算法的软岩隧道变形失稳预警模型，形成了多场信息深度融合的隧道灾害辨识与综合预警系统，并开展了现场试验，为软岩等隧道失稳过程透彻感知与灾害准确预警提供了支撑。项目发表SCI检索论文13篇，授权国家发明专利4项，协助培养博士研究生2名、硕士研究生8名，项目成果获中国科技产业化促进会科学技术奖科技创新奖1项，中国自动化学会技术发明奖1项。