故障诊断专家系统中的模糊推理方法  

Fuzzy Reasoning Method in Expert System for Fault Diagnosis

作　　者：梅胜敏[1] 许洪昌[1] 张幼桢 

机构地区：[1]南京航空航天大学机械工程系

出　　处：《南京航空航天大学学报》1995年第4期519-524,共6页Journal of Nanjing University of Aeronautics & Astronautics

基　　金：航空科学基金

摘　　要：分析了在构造机械加工故障诊断专家系统时领域知识的不确定性。认为在故障诊断领域，专家的思维逻辑以及大部分领域知识反映的是事物的模糊性而不是随机性，由此提出了基于模糊集理论的模糊诊断方法。利用隶属度和模糊综合评判的概念和方法有效地处理了领域知识描述和诊断推理中的不确定性问题。指出了在模糊推理模型中，模糊诊断权矩阵是专家诊断知识和诊断经验的集中体现，并且给出了确定初始诊断权矩阵的方法。The uncertain problem of the domain knowledge in building machining operation expert system for fault diagnosis is analysed. It was thought that diagnosing logic and most domain knowledge reflect the fuzziness but not the randomness in the fault diagnosis,then the fuzzy diagnosis method is proposed based on fuzzy set theory. The concept and the method of membership and fuzzy comprehensive judgement are used to deal with the field knowledge description and the uncertain problem in diagnosing inference. It was pointed out that the fuzzy diagnosis weight matrix in the fuzzy diagnosis model was the concentrative reflection of the diagnosis knowledge and experiences of the human experts.At last,the algorithm to originaly determine diagnosis weight matrix is given.

关 键 词：专家系统 故障诊断 模糊集论 权矩阵 金属切削 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置] TG506[自动化与计算机技术—控制科学与工程]