连续系统故障可诊断性评价方法综述

Overview of fault diagnosability evaluation methods for continuous systems

故障可诊断性评价是故障诊断的基础,能够为诊断算法的开发和传感器的优化配置提供依据与指导.在梳理可诊断性概念、评价流程与功能的基础上,从定量模型、定性模型以及数据驱动这3个方面对连续系统可诊断性评价方法进行分类综述,归纳和评析了各方法的原理与特点,并比较了定性评价与量化评价结果的优缺点.最后探讨了故障可诊断性评价方法未来可能的发展方向.

Fault diagnosability is the fundamental for fault diagnosis. The knowledge of achievable diagnosability performance can provide the theoretical guidance for the development of diagnostic algorithms and the optimization of sensor placement. The concept and evaluation process of diagnosability are firstly introduced. Then existing fault diagnosability evaluation methods for the continuous systems are categorized into three types, i.e., methods based on the quantitative model, the qualitative model, and the statistical data. The characteristics of each method are summarized, and comparison of qualitative and quantitative evaluation is presented. Finally, several potential research tendencies are discussed.