基于FPN的柔性焊装系统故障诊断方法研究

An FPN-based fault diagnosis method of flexible welding assembly system

文章结合模糊推理和产生式规则的信息融合方法,以模糊产生式规则中的置信度表示不确定的故障传播关系,用隶属度对无法精确度量的故障部位状态模糊化,在此基础上通过建立基于模糊Petri网的柔性焊装系统故障诊断模型,实现传感器监测信息、元器件和零部件状态、操作人员状态信息及其信息间不确定关联关系的有效融合;在考虑事前故障诊断经济性原则和事后故障诊断判明时效性原则的基础上,充分利用Petri网并行推理的优点,通过一种形式化推理算法,实时得到诊断结果;通过实例验证了该方法的可行性。

Firstly, based on the information fusion method of fuzzy reasoning and production rules, uncertain fault propagation relations are expressed by the confidence level of fuzzy production rules, and the fault posi- tion state that can not be measured accurately is fuzzed by the membership degree. Then, by creating a fault diagnosis model based on fuzzy Petri net（FPN）, sensors information, components and parts status informa- tion, operators＇ status information, and fuzzy relationships information between the failure characteristics and the failure mode are fused effectively. Finally, considering the economic principles in preventive fault diagnosis and the timeliness principle in diagnosis after the fault, and taking full advantage of Petri nets＇ parallel rea- soning, real-time fault diagnosis is fulfilled by a formal reasoning algorithm. The feasibility of the method is verified by an example.