论文部分内容阅读
在火灾报警系统中火灾概率分析存在不确定性因素问题,为此文章提出用贝叶斯网络对火灾概率进行分析。首先通过分析火灾燃烧原理,得到火灾概率与燃烧过程产生的物化特征之间的内在逻辑关系;在定义火灾燃烧特征参量作为贝叶斯网络节点变量的基础上,创建了基于Netica的火灾报警系统贝叶斯网络模型。通过概率推理和对节点的证据敏感性分析,验证了利用贝叶斯网络模型对火灾发生概率进行分析是可行的、有效的。
In the fire alarm system, there is uncertainty in the fire probability analysis. Therefore, this paper proposes to analyze the probability of fire using Bayesian network. Firstly, by analyzing the theory of fire combustion, the inherent logical relationship between the probability of fire and the physicochemical characteristics of the combustion process is obtained. On the basis of defining the parameters of fire combustion as the node variables of Bayesian network, a fire alarm system based on Netica Yeats network model. Probabilistic reasoning and evidence sensitivity analysis of nodes proved that it is feasible and effective to use Bayesian network model to analyze the probability of fire occurrence.