Modelica扩展建模的故障诊断技术研究  

Failure Diagnosis based on Modelica Extended Model

作　　者：邱晓红[1,2] 龚姚腾[1] 邱晓辉[2] 

机构地区：[1]江西理工大学软件学院,南昌330013 [2]江西农业大学高教研究所,南昌330045

出　　处：《科技通报》2011年第5期641-646,共6页Bulletin of Science and Technology

基　　金：国家自然科学基金(60674054);江西省自然科学基金(2010GZS0090)

摘　　要：故障诊断技术是提高故障检测和隔离能力,提高任务可靠性的重要手段,基于模型的故障诊断分析方法得到了广泛应用,但难以用于实时检测诊断,难以计算虚警率等设计指标。本文提出扩展多领域物理系统建模语言Modelica,将多种故障模态嵌入模型中,利用基于假设的真值维护系统(ATMS)方法分析了测试集的生成算法。理论上分析了计算虚警率的可行方案,体现了新模型的优势。Failure Diagnosis is an important means to improve mission reliability, fault detection and isolation rate. Model-based fault diagnosis methods has been widely used, but it is difficult for real-time detection diagnosis and to calculate design specifications such as the false alarm rate. In this paper, it is proposed to extend the multi-domain physical system modeling language Modelica to describe a variety of failure modes model in the simulation model. And assumption-based truth maintenance system （ATMS） method is used to analyze the generation of test sets. In theory the false alarm rate calculation method is given. This has shown the advantages of the new model.

关 键 词：测试性 故障诊断 基于模型 基于假设的真值维护系统 

分 类 号：TP206[自动化与计算机技术—检测技术与自动化装置] TP202[自动化与计算机技术—控制科学与工程]