基于模糊神经网络的公路隧道洞口段施工阶段风险评估

发布时间：2018-06-03 23:20

本文选题：公路隧道 + 洞口段施工　；参考：《西安工业大学》2017年硕士论文

【摘要】：公路隧道洞口段是隧道施工的重要部位,洞口埋深较浅,一般围岩风化严重,加上降雨的影响,很容易发生边仰坡失稳、掌子面坍塌等事故,严重影响施工进度和工程安全。因此,在公路隧道施工中一直将洞口段作为隧道施工的关键点。对洞口段施工阶段进行安全风险评估,能够及时发现风险,采取有效措施降低风险发生概率和减少风险造成的后果,保障施工安全。本文基于洞口段为浅埋、围岩级别为V级的公路隧道,以模糊神经网络法为基本方法,建立了公路隧道洞口段施工阶段风险概率估计模型,并应用于宝汉高速某隧道洞口段施工阶段风险评估,研究成果如下:1)针对传统的风险评估方法人为主观因素大,具有很大的不确定性和局限性提出了将模糊数学理论与神经网络理论相结合的模糊神经网络法。针对公路隧道洞口段施工阶段存在的风险及风险因素,通过工程调研和研究国内外最新成果基础上,针对公路隧道洞口段施工阶段风险及风险因素采用专家调查法进行识别,并运用层次分析构建了公路隧道洞口段施工阶段风险评估指标体系。2)在应用中,针对模糊神经网络的学习样本,在公路隧道洞口段施工阶段风险评估指标体系基础上,利用模糊层次综合评判法结合相关资料建立了样本数据。3)运用MATLAB软件构建以模糊神经网络法为基本方法的公路隧道洞口段施工阶段风险概率估计模型,确定风险发生的概率等级;采用后果当量法建立风险后果估计模型,确定风险发生的后果等级,运用R= P × C法进行风险综合等级评价。4)将风险概率估计模型、后果估计模型应用到宝汉高速某隧道洞口段施工阶段进行风险估计,并根据公路隧道洞口段施工阶段风险等级标准进行综合评价,提出有效的风险处理措施,保证隧道安全施工。5)应用有限元软件MIDAS/GTS对宝汉高速某隧道洞口段施工阶段进行数值模拟分析,结合本文对其进行风险评估的结果和提出的风险处理措施,建立新的数值模型。将原模型、采取风险处理措施后优化模型及现场实际监控量测结果进行对比分析,验证基于模糊神经网路的公路隧道洞口段施工阶段风险概率估计模型的准确性和风险处理措施的可行性。
[Abstract]:The entrance section of highway tunnel is an important part of tunnel construction. Because of the shallow buried depth, serious weathering of surrounding rock and the influence of rainfall, accidents such as slope instability and palm collapse are easy to occur, which seriously affect the construction progress and engineering safety. Therefore, the entrance section is always regarded as the key point in the highway tunnel construction. The safety risk assessment in the construction stage of Dongkou section can find out the risk in time, take effective measures to reduce the probability of risk occurrence and reduce the consequences of risk, so as to ensure the safety of construction. In this paper, based on the highway tunnel with shallow burying and V grade surrounding rock, the risk probability estimation model of the tunnel entrance is established by using the fuzzy neural network method as the basic method. And applied to the risk assessment of a tunnel entrance section in Baohan Expressway. The research results are as follows: (1) aiming at the traditional risk assessment method, the subjective factors are large. A fuzzy neural network method which combines fuzzy mathematics theory with neural network theory is proposed with great uncertainty and limitation. In view of the risk and risk factors existing in the construction stage of highway tunnel entrance section, based on the engineering investigation and research on the latest achievements at home and abroad, the risk and risk factors in the construction stage of highway tunnel entrance section are identified by expert investigation method. The risk assessment index system of highway tunnel entrance section is constructed by using AHP. In the application, the risk assessment index system is based on the study sample of fuzzy neural network and the risk assessment index system in the construction stage of highway tunnel entrance section. This paper establishes the sample data. 3) using MATLAB software to construct the risk probability estimation model of highway tunnel entrance construction stage based on fuzzy neural network method. The risk probability estimation model is established by using the consequence equivalent method, the consequence grade of the risk occurrence is determined, and the risk comprehensive grade is evaluated by using the R= P 脳 C method. The result estimation model is applied to the risk estimation in the construction stage of a tunnel entrance section of Bao-Han Expressway, and the comprehensive evaluation is carried out according to the risk grade standard in the construction stage of the highway tunnel entrance section, and the effective risk treatment measures are put forward. The finite element software MIDAS/GTS is used to simulate and analyze the construction stage of a tunnel entrance of Bao-Han Expressway, and a new numerical model is established based on the results of risk assessment and the risk treatment measures proposed in this paper. Comparing and analyzing the original model, the optimized model after taking the risk treatment measures and the actual monitoring and measuring results on the spot. The accuracy of risk probability estimation model based on fuzzy neural network and the feasibility of risk treatment are verified.
【学位授予单位】：西安工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U455.4

【参考文献】