混杂系统元件故障的半定性辨识  

Fault identification in hybrid systems based on semi-qualitative analyses

作　　者：彭昭[1] 王文辉[1] 周东华[1] 

机构地区：[1]清华大学自动化系,北京100084

出　　处：《清华大学学报（自然科学版）》2005年第1期111-114,118,共5页Journal of Tsinghua University(Science and Technology)

基　　金：国家自然科学基金资助项目(60025307);教育部博士点基金项目;国家"九七三"基础研究基金项目(2002CB312200)

摘　　要：混杂系统用于描述包括连续状态和离散状态的复杂动态系统。为了研究混杂系统的故障辨识问题,提出了一种将模糊集和模糊熵理论相结合的新的半定性方法。该方法同时利用定量信息和定性知识,以模糊数描述系统特征参数,从而构造一组模糊故障诊断规则,利用模糊熵计算加权系数,从而辨识出混杂系统的故障。对一混杂三容水箱的元件故障进行了仿真研究。结果验证了该方法能够很好地对系统中的故障进行半定性辨识,并表明有应用价值。Hybrid systems have been used to describe complex dynamic systems that involve both continuous and discrete states. The complexity of hybrid systems and the increasing demand for quality, safety, and reliability have made the fault diagnosis of hybrid systems a crucial research topic. This paper describes an integrated fuzzy sets and fuzzy entropy approach to identify faults in hybrid systems. Based on both quantitative information and qualitative knowledge, this semi-qualitative approach uses fuzzy set describing the system character parameter, sequentially constructs a set of fuzzy fault diagnosis rules, computes the weighting coefficient with fuzzy entropy, thereby identifies the faults in hybrid systems. Simulation results for a fault detection and diagnostic system for a hybrid three-tank-system demonstrate the effectiveness of the approach and show that this semi-qualitative method can be used in fault identification of hybrid systems.

关 键 词：故障辨识 混杂系统 半定性 模糊集 模糊熵 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置]