摘要
由于运营隧道的隐蔽性和风险因素的不确定性、复杂性增加了运营隧道防水系统可靠性研究的难度。根据历史数据和专家经验建立了贝叶斯网络拓扑结构并确定根节点发生故障的先验概率值,将Leaky Noisy OR模型引入到贝叶斯网络中,大大降低对条件概率值的需求量,同时提高模型的准确性。并应用Ge NIe Ver 2.0软件对模型进行推导,分别计算出隧道防水系统的失效概率和各风险因子的后验概率,根据后验概率的大小可以为隧道防水系统运营维护提供理论指导,提高工作效率。通过对各风险因子进行敏感性分析,找到系统中存在的敏感性因子,指导在隧道施工中降低敏感因子的发生概率以提高隧道防水系统的可靠性。
The concealment of the operating tunnel and the uncertain and complex risk factors both increase the difficulty in the reliability research of operational tunnel waterproof system. Firstly this paper established the bayesian network topology and identified the root node prior probability of the failure of values based on the history data and expert experience. Secondly the Leaky Noisy OR model is introduced into the bayesian network,which greatly reduced the demand of the conditional probability value,at the same time improved the accuracy of the models. Then this paper applied Ge NIe Ver 2.0 software to model deduction,calculated the tunnel waterproof system failure probability and posterior probability of each risk factor respectively. The size of the posterior probability can provide theoretical guidance for tunnel waterproofing system operation maintenance,and improve the work efficiency. Finally through the sensitivities analysis of each risk factor,this paper found out the sensitivity of the system factor,provided guidance to reduce the occurring probability of sensitive factors to improve the reliability of tunnel waterproofing system in the tunnel construction.
出处
《工程管理学报》
2015年第6期69-73,共5页
Journal of Engineering Management
关键词
运营隧道
防水系统
贝叶斯网络
残余变量
可靠性
operating tunnel
waterproof system
bayesian networks
residual variables
reliability
