隧道出现大变形现象的原因,一方面是膨胀性矿物成分的遇水膨胀作用,从而使隧道或洞室的周边产生大变形。另一方面是软弱围岩在高地应力作用下发生的挤压性变形,由于应力复杂,而岩石的强度较低,洞室在开挖后,周边将出现大范围的塑性破坏区,塑性区内的岩体发生剪切和挤压作用,迫使围岩中的质点向开挖空间进行移动,从而出现了大变形。因地质条件存在不确定性以及施工中开挖方法和支护措施不合理等各种原因,在挤压性软弱围岩中隧道施工易发生塌方,从而影响施工安全、工期和工程投入。故在挤压性软弱围岩隧道施工中,选择合理的支护措施和施工方案对控制围岩大变形是十分重要的,本文以宜万铁路堡镇隧道软弱围岩大变形段施工为依托工程,综合应用现场实测、理论分析、数值模拟和室内试验等手段,并引入人工智能技术,进行了如下研究：(1)现场测试试验方法的研究及结果分析,测试内容包括围岩变形监测和各项选测项目；(2)堡镇隧道施工期软弱围岩段大变形预测的智能方法；(3)堡镇隧道软弱围岩大变形的机理及结构参数优化；(4)堡镇隧道软弱围岩物理力学参数与初始地应力场的三维弹塑性反分析；(5)高地应力软弱围岩预留变形量的确定；(6)高地应力软弱围岩大变形分级标准及相应工程措施；(7)堡镇隧道高地应力大变形段位移控制基准；经对现场实测数据的统计分析,并结合理论分析和数值模拟,得出如下成果：(1)采用十进制遗传算法在BP网络训练过程中自动搜索训练效果最优的网络拓扑结构参数,形成进化神经元算法,并编制了相应的计算程序,将该算法和程序应用于堡镇隧道施工期软弱围岩大变形的多步滚动预测,应用结果表明该算法具有极高的预测精度。(2)采用进化神经元算法进行堡镇隧道施工大变形段的围岩物理力学参数和初始地应力场反演,将获得的反演结果用于隧道后续施工的围岩变形短期预测,使用结果表明,该方法有较高的短期围岩变形预测精度。(3)提出了不同地层结构的围岩预留变形量计算方法,以此确定了不同岩性、岩层组合发生大变形的预留变形量控制标准。(4)综合考虑围岩抗压强度、地应力、弹性模量及侧压力系数等因素,引入综合系数α提出了大变形等级划分的综合指标判定方法,并明确给出了不同大变形等级下的防治技术措施。(5)制定了堡镇隧道高地应力大变形段位移管理基准,并给出了不同管理等级和管理位移下的施工状态和工程措施。以上研究成果在堡镇隧道软弱围岩大变形段施工中得到应用,有效地保证了该段围岩隧道的施工安全和工程质量,并为类似工程设计和施工提供了一定的指导作用。
The cause for large-scale deformation in tunnel lies in that:on the one hand, expansion effect of expansive material ingredient leads to large deformation in tunnel; on the other hand, weak soft rock occurs extruded deformation under the effect of complicated stress. Due to complexity of stress and low strength of rock, once the tunnel is excavated, extensive plastic damage area will appear. The shearing and extruding effect occurs on rock in plastic area, and forces the mass to move toward excavation space, thus, large deformation is resulted in.In view of the various reasons, such as uncertainty of geological condition, unreasonable excavation method and supporting measures, collapse is easy to take place during tunnel construction in extruded weak and soft rock area, which affects construction safety, duration and investment. Therefore, selection of reasonable supporting measures and construction plan are critical for control over large deformation of extruded weak rockmass.In this article, combined with construction of Baozhen tunnel in soft, weak and large deformation rock area on Yiwan railway, many measures, such as site measurement, theoretical analysis, numerical simulation, indoor test and artificial intelligence, were integrated to execute the following studies:(1) Study on site measurement method, data analysis and testing content includes rock deformation monitoring and each optional test item.(2) Intelligent forecasting method of large deformation in soft and weak rock area during construction of Baozhen tunnel.(3) Large deformation mechanism and structural parameter optimization for soft and weak rock of Baozhen tunnel.(4) Elasto-plastic back analysis of physical & mechanical parameters and initial crustal stress field for soft and weak rock of Baozhen tunnel in three-dimension.(5) Determination of reserved deformation for soft and weak rock with high crustal stress.(6) Large deformation classification standard of soft and weak rock with high crustal stress and corresponding engineering measure.(7) Displacement control benchmark for large deformation section with high crustal stress in Baozhen tunnel. Through statistics and analysis on site measurement data, and in combination with theoretical analysis and numerical simulation, following results can be obtained:(1) The decimal genetic algorithm was used to search the optimal topological structure parameters of BP network during it’s training process and corresponding calculation code was programmed. This algorithm and program were applied in large deformation prediction of Baozhen tunnel by multi-step rolling method. The application result indicates such algorithm has extremely high forecast accuracy, which provide reliable guarantee for predicting final deformation and determining best time of secondary lining.(2) Evolutionary neural algorithm was used to inverse the physical and mechanical parameters of rockmass and initial crustal stress in large deformation area, and the inversion results are used to forecast short-term rock deformation in subsequent excavation. From the applying result, it can be concluded that this method possess high forecasting accuracy.(3) Calculation formula for rock reserved deformation in different stratum structure is proposed to determine reserved deformation control standard for large deformation of rock stratum composition with different rock properties.(4) The factors, such as compressive strength, crustal stress, plastic modulus and side pressure coefficient are considered as a whole to determine the a classification standard of large deformation in Baozhen tunnel. The comprehensive index judgment method for deformation level classification was put forwarded, and the prevention and treatment measures for different-level large deformation were presented clearly.(5) Displacement control benchmark in large deformation area with high crustal stress is determined. The construction condition and engineering measures under different management level and managed displacement are proposed.Above mentioned study results had been applied in construction of Baozhen Tunnel in soft and weak rock area with large deformation, which ensures construction safety and project quality of the tunnel, and provides guidance for similar engineering design and construction effectively.
