【作者】 陈颖；

【导师】 周建春； 宋晓凯；

【作者基本信息】 华南理工大学 ， 土木工程（专业学位）， 2024， 硕士

【摘要】 隧道工程作为土木与交通基础设施的重要组成部分,其施工安全与效率至关重要,其中隧道掘进机(TBM)凭借其高效和安全的特点成为隧道机械化施工的重要装备。然而,面对复杂多变的地质条件,基于先进的信息技术开展TBM施工技术研究与开发,对于确保特长隧道施工安全与效率是至关重要的。本文以某特长引水隧洞工程为依托,针对特长隧道工程TBM智能建造的数字孪生技术及其应用开展了研究,在构建特长隧道TBM智能建造系统数字孪生技术框架的基础上,从TBM施工过程中滚刀磨损智能预测和超前地质智能预报两个维度进行了较深入的研究探讨,并结合依托工程开展了针对性的应用研究,从而实现了特长隧道TBM施工的智能建造。本文的主要研究内容和取得的成果如下:(1)在广泛调研和分析相关文献的基础上,构建了特长隧道TBM智能建造系统数字孪生技术框架,该框架旨在确保系统兼容性、数据实时交互及智能化服务应用目标,包括基准建模、泛在感知、实时预测数字孪生体和服务应用等模块,并基于深度学习算法研发了滚刀磨损超前智能预测和超前地质智能预报应用层。(2)基于长短期记忆神经网络算法(LSTM)研究了特长隧道TBM滚刀磨损智能预测数字孪生体构建方法,即先通过Ansys/LS-DYNA软件进行有限元模拟分析以获取深度学习样本,然后通过贝叶斯优化算法建立LSTM时序预测模型。通过模型性能评估验证了特长隧道TBM滚刀磨损智能预测方法的可靠性和有效性。(3)基于卷积-全连接神经网络算法(CNN-FCNN)研究了特长隧道TBM施工超前地质预报围岩力学参数智能预测数字孪生体构建方法,即先通过Ansys/LS-DYNA软件进行有限元正演模拟以获取深度学习样本,然后经数据预处理以及贝叶斯优化超参数后,实现对围岩力学参数的智能预测,为特长隧道TBM施工超前地质预报提供了新技术。(4)以某特长引水隧道TBM施工为工程依托,基于上述成果开展了数字孪生特长隧道TBM智能建造技术的应用研究,首先搭建了数字孪生特长隧道TBM智能建造系统平台,然后开发了TBM施工过程中滚刀磨损智能预测及超前地质预报智能预测软件系统,结合依托工程TBM施工开展了针对性的应用研究,实际工程实践验证了本文成果的可靠性,有效地提高了依托工程TBM施工安全性和效率。更多还原

【Abstract】 As an important part of civil and transportation infrastructure,tunnel engineering is very important for construction safety and efficiency,and tunnel boring machine(TBM)has become an important equipment for tunnel mechanized construction due to its high efficiency and safety.However,in the face of complex and changeable geological conditions,it is crucial to carry out the research and development of TBM construction technology based on advanced information technology to ensure the safety and efficiency of extra-long tunnel construction.Based on the construction of the digital twin technology framework of the TBM intelligent construction system of the extra-long tunnel,the two dimensions of intelligent prediction of disc cutter wear and advanced geological intelligent prediction in the TBM construction process are studied and discussed,and the targeted application research is carried out in combination with the supporting engineering,so as to realize the intelligent construction of TBM construction of extra-long tunnel.The main research contents and achievements of this paper are as follows:(1)On the basis of extensive research and analysis of relevant literature,a digital twin technology framework for TBM intelligent construction system for extra-long tunnels was constructed,which aims to ensure system compatibility,real-time data interaction and intelligent service application goals,including benchmark modeling,ubiquitous perception,real-time prediction of digital twin and service application modules,and based on deep learning algorithms,the application layer of advanced intelligent prediction of disc cutter wear and advanced geological intelligent prediction was developed.(2)Based on the long short-term memory neural network algorithm(LSTM),the construction method of intelligent prediction of TBM hob wear in extra-long tunnels was studied,that is,the finite element simulation analysis was carried out by Ansys/LS-DYNA software to obtain deep learning samples,and then the LSTM time series prediction model was established by Bayesian optimization algorithm.The reliability and effectiveness of the intelligent prediction method of TBM disc cutter wear in extra-long tunnels are verified by model performance evaluation.(3)Based on the convolutional-fully connected neural network algorithm(CNN-FCNN),a digital twin construction method for intelligent prediction of surrounding rock mechanical parameters for TBM construction of extra-long tunnels was studied,that is,the finite element forward simulation was carried out by Ansys/LS-DYNA software to obtain deep learning samples,and then the intelligent prediction of the mechanical parameters of surrounding rock was realized after data preprocessing and Bayesian optimization hyperparameters,which provided a new technology for the advanced geological prediction of TBM construction of extra-long tunnels.(4)Based on the TBM construction of an extra-long diversion tunnel,the application research of the intelligent construction technology of the digital twin extra-long tunnel was carried out based on the above results,firstly,the digital twin extra-long tunnel TBM intelligent construction system platform was built,and then the intelligent prediction software system of hob wear and advanced geological prediction in the TBM construction process was developed,and the targeted application research was carried out in combination with the TBM construction of the relying project,and the reliability of the results of this paper was verified by the actual engineering practice,and the safety and efficiency of the TBM construction of the relying project were effectively improved。更多还原