摘要
给出一种基于符号有向图相容通路信度动态融合的故障诊断方法。建立基于趋势分析的SDG模型,当系统发生故障或异常时,根据节点之间的数据及相容规则计算相容通路的信度,并将其转化为相容通路证据。基于递归的条件化线性组合更新规则,利用当前证据更新历史证据,得到更新后的全局相容通路证据,全局证据中赋值最大的相容通路即为故障传播的路径。相对于基于趋势分析的SDG故障诊断方法,新方法提高了系统早期故障诊断的可靠性。文中通过三级液位控制系统的故障仿真实验对其进行验证,说明其可行性和有效性。
This paper presents the fault diagnosis method using dynamic fusion of SDG consistent path belief.A SDG model of system is firstly constructed based on trend analysis.When the faults or abnormal nodes are found out,the beliefs of consistent paths can be obtained by the data of these nodes and compatibility rule, and the consistent path evidence is calculated by normalizing the beliefs.Based on recursive conditional linear updating rule of evidence, the incoming consistent path evidence at t step is used to update the historical consistent path evidence so as to obtain the updated consistent path evidence at t step.The path having the biggest mass of belief is the fault propagation path.Compared with qualitative trend analysis based SDG fault diagnosis,the new method can improve the reliability of early fault diagnosis.This paper shows the effectiveness and feasibility of the proposed method by fault simulation experiment on three-level control system.
出处
《杭州电子科技大学学报（自然科学版）》
2015年第5期23-26,共4页
Journal of Hangzhou Dianzi University：Natural Sciences
基金
国家自然科学基金资助项目(61374123
61433001
61034006)
浙江省科学公益技术研究计划资助项目(2012C21025)
关键词
故障诊断
证据更新
信息融合
符号有向图
相容通路
fault diagnosis
evidence update
information fusion
signed directed graph
consistent path
