基于Noisy-or Gate的贝叶斯网络机械伤害风险评估

Mechanical Injury Risk Assessment Based on Noisy-or Gate Bayesian Network

为改进故障树转化成贝叶斯网络时在数值转化中或门转化的缺陷,以某机械加工企业的机械伤害事故为例,使用Noisy-or Gate模型确定故障树中或门因素转化成贝叶斯网络节点的条件概率,降低或门中因0和1概率产生的误差,并与传统故障树转化成贝叶斯网络的模型进行对比分析,最后根据结果分析得出关键影响因素,为机械伤害中故障类型影响分析提供新的思路。

In order to improve the value transformation or gate transformation when fault tree is transformed into Bayesian network.Taking the mechanical injury accident of a machining enterprise as an example,the conditional probability of transforming the OR gate factor in the fault tree into the Bayesian network node is determined by using the Noisy-or gate model,so as to reduce the error caused by the probability of 0 and 1 in the OR gate.Finally,the key factors of failure are analyzed by comparing with the traditional model that transforms fault tree to Bayesian network.It provides a new way to analyze the influence of fault types in mechanical injury.