摘要
针对传统可靠性分析方法在评估多态复杂系统时的局限性,提出了一种基于贝叶斯网络的多态系统可靠性分析方法;该方法利用贝叶斯网络的多态变量节点来描述故障模式的多态性,用条件概率表及有向边来表达节点之间的不确定因果关系;通过对先验概率和后验概率的双向推理计算,得到节点与节点之间的相互影响关系,从而找到系统的薄弱环节,为提高系统可靠性提供依据;该方法在电池生产线系统的可靠性分析中得到了验证,有效地提高了电池生产线系统的可靠性。
Aiming at the limitations of traditional reliability analysis methods in the evaluation of multi-state complex system,a reliability analysis method for multi-state system based on Bayesian networks is proposed.In this model the multi-state variable nodes of Bayesian networks are applied to represent multi-states of failure modes,the conditional probability tables and directed arcs are used to represent uncertain cause–effect relationships among the variables.Through bidirectional inference calculation of the priori and posteriori probability,the interrelationship between notes is obtained in order to identify the system weakness,and provide basis to improve system reliability.The proposed method is verified in reliability analysis of the cell production line,and the system reliability is effectively improved.
作者
翟胜
田硕
陈倩倩
Zhai Sheng;Tian Shuo;Chen Qianqian(School of Electrical Engineering,Dalian Institute of Science and Technology,Dalian 116052,China)
出处
《计算机测量与控制》
2020年第9期262-266,共5页
Computer Measurement &Control
关键词
贝叶斯网络
多态
可靠性分析
复杂系统
Bayesian networks
multi-state
reliability analysis
complex system