基于贝叶斯网络的住宅防水风险研究

Housing Waterproof Risk Based on Bayesian Network

针对目前住宅小区经常出现漏水事故而没有成熟的方法找到其根源的问题,基于贝叶斯网络理论对住宅防水问题进行了研究。经过分析漏水事故的发生机理,结合相关领域专家的施工经验,确定主要的致险因素。在此基础上建立住宅防水风险贝叶斯网络结构,通过贝叶斯网络正向推理和反向推理,分别得出住宅漏水的概率和各个根节点的后验概率,经过排序得到关键致险因素。结合故障树理论,将各个底事件的重要度排序与贝叶斯网络得出的关键致险因素取交集,得到最终的关键致险因素。应用该方法可以对住宅漏水的问题进行有效的施工控制,指导性强。

In view of the current residential problem of water leakage accident,there is no mature method to find its origin. Based on Bayesian network theory, the problem of housing waterproof is studied. Through mechanism analysis of water leakage accident,combing with experience of related experts, main risk factors are determined. Bayesian network of housing waterproof risk is established. Through the forward and backward reasoning of Bayesian network, the probability of residential water leakage and each posterior probability of root node can be obtained. At last, the key risk factors can be determined. Using the fault tree theory, the same assembly from Bayesian Network can be obtained. Finally the key risk factors are gotten. The method can be applied to residential water leakage problem.