【作者】 李涛；

【导师】 仇文革；

【作者基本信息】 西南交通大学 ， 桥梁与隧道工程， 2019， 博士

【摘要】 铁路隧道向“更高密度、更长隧道、更大断面”发展,导致工程建设风险更大,对隧道建设管理提出了更高要求。相比其它非地下工程,隧道安全事故、质量事故、工期延误、成本失控的概率较高,后果更严重。若能借助信息化技术实现全过程目标精细化管控,将有效降低以上隧道建设风险,故这方面的研究颇具价值。通过现场调研与咨询,在不断凝练用户需求的基础上,深入研究了隧道信息化平台的构建及基于数据驱动的隧道多目标精细化管理问题,得出以下结论与并取得以下成果:1.隧道管理必须及时、准确掌握隧道地质。故针对现阶段隧道超前地质预报系统管理粗放、数据利用率低等问题,结合数据驱动重塑地质预报流程,研发了隧道地质信息系统TGIS。其高效集成了隧道勘察设计阶段地质、施工阶段超前预报地质、掌子面开挖揭示地质信息一体化,并预留接口延续至隧道运维阶段;实现了基于掌子面地质后评估的预报优选与动态调整,基于施工进度的实时提醒预报施做,基于地质异常的实时分级预警推送及基于机器学习的实现辅助预警处置等功能。达到了隧道地质信息精细化管理目的,为隧道安全、工期、成本乃至环境目标的有效控制提供有力支撑;2.掌子面开挖揭示的地质是最准确、最直观的,也是最具有价值的。为克服传统人工素描工作流于形式及掌子面图像识别准确率低等问题,基于多图像立体重建技术或三维激光扫描技术,构建并解决了点云重构掌子面三维地质算法,真实3D再现隧道开挖地质及自动提取结构面。并据此研发了隧道掌子面地质信息系统FGIS,基本实现了自动三维地质重构、自动结构面参数提取、自动围岩分级、自动报表及三维成果展示等功能;3.隧道支护安全、经济与否可通过监控量测数据予以准确、定量评估。故结合数据驱动重塑监测流程,研发了隧道监控量测系统TMMS,既支持传统监测手段(必测项目+选测项目),又兼容新型量测手段(自动化监测+全息变形监测),实现了基于地质的监测方案实时调整更新、测点布设实时提醒与测点延误布设统计,监测施做实时提醒与监测延误施做统计,基于变形(加)速度的实时预警与支护参数动态调整,基于经验数据类比的支护参数优化等功能;4.针对传统隧道监控量测存在的问题,引入三维激光扫描进行全息变形监测。无需标靶埋设,整体与局部变形一览无遗,克服了传统有限断面有限点监测的局限性,真实3D再现了隧道施工现场,避免了数据造假的可能。并据此构建了基于全息变形监测的隧道支护评估体系,即变形加速度控制安全、累计变形控制限界及经济,达到了安全与经济统一量化评估,避免了传统监测数据假安全的可能,又规避了监测数据有限导致无法经济性评估的局限性;5.针对检测参数多、检测频次高、资料管理难度大等问题,结合数据驱动重塑质量检测流程,研发了隧道质量检测系统TQIS。基本实现了材料、初支质量、衬砌质量检测数据的自动评定、自动报表与成果展示集成一体可视化功能,通过信息化手段实现了现场质量精细化管控目的;6.为克服传统隧道质量抽样检测的缺陷,提出了基于三维激光扫描的隧道快速、全息质量检测,并研发了隧道三维激光点云数据处理分析系统PCAS,实现了隧道超欠挖分析、喷砼方量预测、初支厚度检测、初支净空检测、初支平整度检测、衬砌钢筋检测、衬砌灌注方量预测、衬砌厚度检测、衬砌净空检测、衬砌平整度检测、衬砌表观缺陷检测等11项功能,为隧道质量精细化管控提供了很好的技术手段;7.通过数理统计,分析隧道工期与成本的相关影响因素,并基于其随机分布,构建蒙特卡洛模拟的精细化定量计算模型,一改传统按确定参数计算工期与成本的不准确性。其次,通过程序化2种模型(经验化模型+精细化模型),研发了隧道建设辅助决策系统RDAT,实现隧道工期与成本自动在线分析,辅助动态优化调整隧道施工组织设计,保证隧道正常有序开展,避免工期延误及赶工期的无序状态;8.为打通系统间的“信息孤岛”,借助信息化技术重塑隧道业务流程,通过整合隧道相关系统,构建隧道全寿命周期统一数据库及统一用户模块、功能权限分级,创新性地研发了隧道信息化平台TIM,实现了隧道建设信息“采集—传输—分析—决策—反馈”的高效网络化流转与利用,从而建立了“地质快速前馈—动态即时设计—监测即时反馈—多目标评估与调控”的信息化管理机制,实现隧道安全、质量、工期、成本目标过程的信息化、精细化、量化管控。并以青岛地铁为例,在青岛地铁三号、二号线海量实测数据挖掘的基础上,对青岛地铁的安全、环境、工期、成本等多目标风险进行统计与回归分析,建立了青岛地铁建设中的典型风险辨识模式及风险预测模型,可优化、指导在建及新建工程。更多还原

【Abstract】 Railway tunnels will develop towards "higher density,longer length,larger cross-section",leading to greater construction risk,which puts forward higher requirements for tunnel construction management.Compared with other non-underground projects,the probability of tunnel safety accidents,quality accidents,project delay,cost out of control is higher,and the consequences are more serious.If meticulous management and control of tunnel targets in the whole process can be achieved based on information technology,the risk of tunnel construction above will be reduced effectively,so this research is of great value.Through on-site investigation and consultation,and on the basis of continuously concreting user’s needs,the construction of tunnel information platform and the multi-objective fine management of tunnel based on data driven are deeply studied,obtaining the following conclusions and results:1.Tunnel geology must be grasped timely and accurately in tunnel management.Therefore,in view of the extensive management and low data utilization rate of tunnel geology prediction system at this stage,geology prediction business process was reshaped and a new system of tunnel geology information system TGIS was developed based on data driven.The geological information in the geological survey and design stage,the geological information of advanced geology prediction in the construction stage and the geological information revealed in tunnel face were integrated efficiently in full view.The interface was reserved for the operation and maintenance stage of tunnel.The functions of prediction optimization and dynamic adjustment based on geological post-evaluation of tunnel face,real-time warning of forecast implementation based on construction progress,real-time grading and early warning pushed by geological anomalies,the assistant disposal of early warning based on machine learning,etc,were achieved,to reach the goal of fine management of tunnel geology information and provide strong support for multi-objective effective control of tunnel safety,duration,cost and even environment.2.The geological information revealed in tunnel face was most accurate,intuitive and valuable.In order to overcome the problems of just a mere formality in traditional manual sketch work and the low accuracy in image recognition,on the basic of multi-image stereo reconstruction technology or 3D laser scanning technology,a algorithm about 3D geological reconstruction of tunnel face point cloud was built and solved,to 3D reproduce tunnel excavation geology truly and extract structure surface automatically.In this,tunnel face geological information system FGIS was developed,to basically realize the functions of automatic 3D geological reconstruction,automatic parameter extraction of structure surface,automatic classification of surrounding rock,automatic report and display of 3D results.3.The safety and economy of tunnel support can be accurately and quantitatively assessed by monitoring data.Therefore,monitoring process was remolded and tunnel monitoring and measurement system TMMS was developed based on data driven,to support both traditional measurement methods(required items and selected items)and new measurement methods(automatic monitoring + holographic deformation monitoring).The functions of real-time adjustment and update of monitoring scheme based on tunnel geology,real-time reminder and delay statistic of monitoring point layout and implementation,real-time early warning and dynamic adjustment of support parameters based on deformation rate or deformation acceleration,optimization of support parameters based on experience data analogy,were realized.4.Aiming at the problems existing in traditional tunnel monitoring and measurement,holographic deformation monitoring based on 3D laser scanning technology was proposed,without target burying,to display overall and local deformation in full view.The traditional monitoring boundedness of limited sections and limited measure points was overcomed,to 3D reappeare tunnel real construction site and avoid the possibility of data fraud completely.According to this,a tunnel support evaluation system based on holographic deformation monitoring was constructed,that was,deformation acceleration was responsible for safety and accumulated deformation value was responsible for limits and economy.The unified quantitative evaluation of safety and economy was achieved,to avoid the possibility of false safety and no economic evaluation with limited data in traditional monitoring.5.Aiming at the problems of many detection parameters,high detection frequency and difficult data management,quality inspection business process was reshaped and tunnel quality inspection system TQIS was developed based on data driven.The functions of automatic evaluation,automatic report and results integrately display of material detection,initial support quality detection and lining quality detection in tunnel construction process,were basically realized,to achieve the purpose of fine quality control on site by means of information technology.6.In order to overcome the sampling inspection shortcomings of traditional tunnel quality,a rapid and holographic inspection method based on 3D laser scanning was proposed,and tunnel 3D laser point cloud data processing and analysis system PCAS was developed.The 11 functions of tunnel over-excavation and under-excavation analysis,shotcrete volume prediction,initial support thickness detection,initial support clearance detection,initial support flatness detection,lining reinforcement detection,lining grouting volume prediction,lining thickness detection,lining clearance detection,lining smoothness detection,lining apparent defects detection,were realized,to provide a good technical means for the fine control of tunnel quality.7.Factors related to tunnel construction period and cost were analysed by mathematical statistics,and a precise quantitative calculation model of Monte Carlo simulation was established based on its random distribution,to change the traditional calculation inaccuracy of tunnel construction period and cost according to the determined parameters.Secondly,risk decision aids for tunnelling RDAT was developed by programming the empirical model and fine model.The functions of online automatic analysis of tunnel construction period and cost,and dynamic optimization and adjustment auxiliary of tunnel construction organization design were realized,to ensure the normal and orderly of tunel construction and avoid project delay and the disordered state of catching up the process.8.In order to break through the "information islands" among the systems,tunnel business process was remodeled by information technology.Through the integration of tunnel related systems,the unified database of tunnel life cycle,unified user modules and functional authority classification were constructed,and tunnel information modelling TIM was developed innovatively,to realize the efficient network transfer and utilization of tunnel construction information "collection—transmission—analysis—decision—feedback".Thus,the information management mechanism of "geological fast feedforward — dynamic real-time design — monitoring real-time feedback — multi-objective evaluation and regulation" was established,to realize informatization,refinement,quantitative process control of tunnel targets of safety,quality,duration,cost,and so on.Taking Qingdao Metro as an example,on the basis of data mining of the massive measured data of Qingdao Metro Line 3 and Line 2,the multi-objective risks of Qingdao Metro,such as safety,environment,duration and cost were statistical and regression analysis,to establish a typical risk identification model and risk prediction model in the construction of Qingdao Metro,which can optimize and guide the uncompleted and new construction projects.更多还原