摘要
针对永磁同步电机(PMSM)系统故障间关联性强、异构信息复杂繁多以及传统故障诊断方法无法追溯耦合故障发生的真正原因等问题,提出了基于本体追踪的故障诊断方法。以PMSM驱动系统开路故障为主要研究对象,首先,对PMSM故障诊断领域本体模型进行形式化定义及标准化构建,统一集成PMSM系统中密集的异构信息;其次,在改进归一化误差电流数据分析法能够精准快速地定位故障开关管的基础上,将定量分析所得故障特征及识别结果映射到故障领域本体模型中,扩展了故障诊断范围;最后,经语义Web(SWRL)规则库的预设以及基于规则的推理,融合本体技术进行故障诱因、故障类型等信息的追溯及诊断。经实验验证,该方法提高了各子系统间信息的利用率,能够合理地追溯故障原因,为大型复杂系统的故障诊断提供了一个可共享、易更新的标准化方法。
A fault diagnosis method based on ontology tracking is proposed for the problems of permanent magnet synchronous machine(PMSM) system strong fault correlation, complex heterogeneous information, and traditional fault diagnostic methods cannot trace the true cause of the coupling fault. Taking the open-circuit fault of PMSM drive system as the main research object, firstly,define and standardize the ontology model in PMSM fault diagnosis field, and integrate the dense heterogeneous information in the PMSM system;secondly, on the basis of improving the normalized fault current data analysis method to accurately and quickly locate the fault switch pipe, the fault characteristics and identification results obtained from quantitative analysis are map to the fault field ontology model, extending the scope of fault diagnosis;finally, through the preset of semantic web rule language(SWRL) rule library and rule-based reasoning, integrate the ontology technology to trace and diagnose the fault cause, fault type and other information. Experiments verify that the proposed method improves the utilization of information among subsystems and can reasonably trace the cause of failure,providing a shared and easy to update standardized method for fault diagnosis of large complex systems.
作者
张潇丹
段斌
刘昌杰
陈月平
Zhang Xiaodan;Duan Bin;Liu Changjie;Chen Yueping(School of Automation and Electronic Information Xiangtan University,Xiangtan 411105 China)
出处
《电工技术学报》
EI
CSCD
北大核心
2022年第13期3341-3349,3360,共10页
Transactions of China Electrotechnical Society
基金
湖南省自然科学基金资助项目(2020JJ6034)。
关键词
改进归一化误差电流
本体
开路故障
故障诊断
驱动系统
Improved normalized error current
ontology
open circuit fault
fault diagnosis
drive system
