基于含不可观变迁Petri网的码垛机器人零件加工系统故障检测方法  

Fault Detection for Robot Processing System Based on Petri Net with Unobservable Transitions

作　　者：刘久富[1] 陈柯[1] 梁娟娟[1] 叶文华[1] 王志胜[1] 

机构地区：[1]南京航空航天大学,南京210016

出　　处：《中国机械工程》2013年第5期589-593,共5页China Mechanical Engineering

基　　金：国家自然科学基金资助项目(60674100);南京航空航天大学基本科研业务费专项科研项目(NS2010069)

摘　　要：研究了基于含不可观变迁Petri网的码垛机器人零件加工系统故障检测问题。结合机器系统运行过程中的故障特征,对Petri网在机器系统故障诊断中的应用进行了研究,提出了包含不可观变迁Petri网的基本可达树和诊断分析函数的概念及基本可达树故障检测算法。以码垛机器人零件加工系统为例,建立了Petri网系统模型,应用基本可达树故障检测算法进行检测,故障检测结果符合码垛机器人零件加工系统实际运行情况,验证了该方法的有效性。This paper investigated the fault detection problem of the Petri net with unobservable transitions. Based on the fault characteristics for machine systems in the operation process as well as basic theory of Petri net,this paper made a study of the applications of Petri net in the fault detection of the robot processing system and presented the concept and algorithm of basis reachability tree and diagnosis analysis function in the Petri net with unobservable transitions. An algorithm of the diag- nostic state was developed by using the basis reachability tree. The system model of part processing system for the motion of stacking robot was built with Petri net, and the diagnosis results of the diagnostic states were analysed. The diagnosis results are consistent with the actual operation of parts processing system and prove the effectiveness of this method.

关 键 词：故障检测 PETRI网 不可观变迁 可达树 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]