论文部分内容阅读
采集汽车发动机润滑系统中摩擦副经历摩擦学行为的信息是诊断发动机摩擦系统故障的有效手段。油液分析从摩擦学系统的润滑剂和磨损物两方面获得摩擦副的润滑和磨损状态的信息。但常规的油液分析信息处理方法上的不足影响了油液分析技术在汽车行业的实际应用。开发基于油液分析的故障诊断专家系统,无疑会改善油液分析技术的诊断准确性和诊断成本。本文着重针对该诊断专家系统知识库的建立,提出了带有自学习功能的基于多元回归统计分析的规则获取方式和基于“规则架─规则体”的知识表示模式,并用PRO-LOG语言给出了汽油机换油指标的“规则架─规则体”表示模式的一个实例。
It is an effective way to diagnose the engine friction system failure by collecting the information that the friction pair experienced the tribological behavior in the automobile engine lubricating system. Oil Analysis Obtains information on the lubrication and wear status of the friction pair from both tribological systems’ lubricants and abrasives. However, the conventional method of fluid analysis and information processing has hindered the practical application of the oil analysis technology in the automobile industry. The development of a fault diagnosis expert system based on oil analysis will undoubtedly improve the diagnostic accuracy and diagnostic cost of the oil analysis technique. This paper focuses on the establishment of the knowledge base of the expert system for diagnosis, and proposes a method of obtaining rules based on multivariate regression statistical analysis with self-learning function and a knowledge representation model based on “rule frame-rule body” The gasoline engine oil indicators “rule frame ─ rule body” that an example of the model.