基于贝叶斯网络的摩擦学系统状态辨识的知识获取  

Knowledge Acquisition of Tribological System Condition Identification Using Bayesian Network

作　　者：袁成清[1,2] 郭智威[1,2] 王志芳[3] 叶枫桦[1,2] 

机构地区：[1]武汉理工大学能源与动力工程学院可靠性工程研究所,湖北武汉430063 [2]武汉理工大学船舶动力工程技术交通行业重点实验室,湖北武汉430063 [3]中国船级社武汉规范研究所,湖北武汉430022

出　　处：《润滑与密封》2009年第12期1-4,共4页Lubrication Engineering

基　　金：国家自然科学基金项目(50705070);教育部博士点新教师项目(20070497029)

摘　　要：研究了基于贝叶斯网络的摩擦学系统状态辨识知识的获取方法。针对摩擦磨损试验数据一般比较齐全且数据量相对比较大的特点,将贝叶斯网络用于摩擦学系统状态辨识的知识获取,通过计算建立属性之间的依赖关系,获得各种依赖关系下的条件概率表,说明依赖关系的强弱,从而获得对机器摩擦学系统状态辨识有指导意义的客观知识,以利于机器摩擦磨损状态辨识的智能化。将贝叶斯网络用于滑动轴承摩擦磨损试验数据的知识获取,获取该摩擦学系统状态辨识的定性及定量知识。Knowledge acquisition method of tribologieal system condition identification using Bayesian network was pro- posed. For the much complete testing data of friction and wear, Bayesian network was used in tribological system condition identification for knowledge acquisition. The objective knowledge which has guiding significance for tribological system con- dition identification of machines was acquired through computing the dependence relation among different attributes. This is contributive to the intelligent condition identification of machine tribological system. The developed method based on Bayesian network was used in the knowledge acquisition of sliding bear to acquire the qualitative and quantitative knowl- edge of tribological system condition identification.

关 键 词：贝叶斯网络 摩擦学系统 状态辨识 知识获取 

分 类 号：TH117.1[机械工程—机械设计及理论]