【作者】 王俊；

【导师】 彭立敏； 何丹；

【作者基本信息】 中南大学 ， 工程（专业学位）， 2024， 硕士

【摘要】 在城市地铁区间隧道的建造中,盾构法已被广泛采用,该方法具有施工质量优异、进度迅速、对环境影响小、施工安全性高等优点,但盾构滚刀磨损问题一直是困扰着工程界的难题。本文依托长沙市轨道交通1号线北延一期工程中盾构穿越微风化和中风化板岩典型地层时滚刀磨损问题,采用现场测试、理论分析、工程实践等方法,研究了滚刀磨损特征和主要影响因素,构建了基于神经网络理论机器学习的盾构机滚刀磨损预测模型,在此基础上提出滚刀降低磨损改善技术并分析了效果。论文完成的主要工作有:(1)在对国内外盾构机刀具磨损研究现状进行调查,综述盾构机工作原理的基础上,分析了滚刀的磨损特点和种类,探讨了盘形滚刀磨损的影响因素和规律,提出了不同类型滚刀磨损量限制建议值。(2)基于工程实例中盾构穿越典型地层时的滚刀磨损换刀与磨损数据统计,并结合电镜扫描观测,分析了滚刀磨损失效形态,研究了磨损机理。结果表明:滚刀磨损的失效主要表现为偏磨、崩角或刀圈断裂、以及磨损超限等形态。盾构机盘型滚刀刀圈磨损的微观机理主要为磨粒磨损和点蚀磨损。(3)采用BP神经网络和粒子群优化(PSO)算法的理论与技术,构建了基于特征抽取与选择的滚刀磨损预测模型,通过在依托工程中的应用,验证和比较了BP神经网络和PSO算法模型在预测盾构机滚刀磨损程度的可行性、有效性和优越性。(4)在深入分析造成盾构机滚刀磨损的原因、减少滚刀磨损优化技术和效果的基础上,研究了滚刀磨损率与掘进效率之间的关系;结合工程实例的地层特点和神经网络预测模型得出数据,提出了具有针对性的、合理的滚刀优化措施,评价了其优化的工程实践效果。图51幅,表20个,参考文献118篇。更多还原

【Abstract】 In the construction of urban subway tunnel,the shield method has been widely used.This method has the advantages of excellent construction quality,rapid progress,small impact on the environment and high construction safety,but the wear problem of shield hob has always been a difficult problem in the engineering field.Based on the hob wear problem of shield tunneling in the first phase of line 1,this thesis,and construct the wear prediction model of shield machine based on machine learning theory of neural network,and improve the technique of reducing wear and improving the effect.The main tasks completed in the paper are:(1)Based on the investigation of the research status of shield machine tool wear at home and abroad and summarizing the working principle of shield machine,the wear characteristics and types of hob are analyzed,the influencing factors and rules of disc hob wear are discussed,and the wear limit of different types of hob is proposed.(2)Based on the statistics of hob wear and wear data when the shield crosses the typical formation in the engineering example,combined with the electron microscope scanning observation,the wear failure form is analyzed and the wear mechanism is studied.The results show that the failure of roller wear is mainly wear,angle or ring fracture,and wear limit.The microscopic mechanism of blade ring wear is mainly grinding wear and pitting wear.(3)Using BP neural network and particle swarm optimization(PSO)algorithm theory and technology,build the hob wear prediction model based on feature extraction and selection,through the application in relying on engineering,verify and compare the BP neural network and PSO algorithm model in predicting the degree of shield machine wear feasibility,effectiveness and superiority.(4)On the basis of the thorough analysis of the optimization technology and effect,the relationship between the wear rate and the tunneling efficiency is studied;combined with the formation characteristics of the engineering examples and the neural network prediction model,the targeted and reasonable hob optimization measures are proposed,and the optimized engineering practice effect is evaluated.Figure Figure 51,table 20,and 118 references.更多还原