基于MMD的故障可诊断性定量评价方法  

Quantitative evaluation approach of fault diagnosability based on maximum mean discrepancy

作　　者：秦玉峰 史贤俊 QIN Yu-feng;SHI Xian-jun(College of Coastal Defense Force,Naval Aviation University,Yantai 264001,China)

机构地区：[1]海军航空大学岸防兵学院,山东烟台264001

出　　处：《控制与决策》2023年第10期2925-2933,共9页Control and Decision

摘　　要：提出一种基于最大均值差异(maximum mean discrepancy,MMD)的故障可诊断性定量评价方法.该方法无需构建任何系统模型,通过度量不同故障模式下测量数据之间的距离定量评价故障可诊断性,适用于结构复杂、不易于建模且能够获取测量数据的复杂系统.首先,将测量数据通过特征核映射到可再生核希尔伯特空间(reproducing kernel Hilbert space,RKHS)中,以MMD作为多元分布距离度量指标,将故障可诊断性定量评价问题转换为多元分布在RKHS中的距离度量问题;然后,利用数学推导分析测量噪声强度对故障可诊断性评价结果的影响;最后,通过仿真实例验证所提出方法的有效性.This paper proposes a method of quantitative evaluation of fault diagnosability based on maximum mean discrepancy(MMD).The method evaluates the fault diagnosability quantitatively by measuring the distance between measurement data under different fault conditions without building any system model.It is suitable for the system with complex structures that are difficult to build models and obtain measurement data.Firstly,the measurement data is mapped to the reproducing kernel Hilbert space(RKHS)through the characteristic kernel.The MMD is taken as the distance measure of multivariate distributions,and the fault diagnosability quantitative evaluation is transformed into the distance measurement of multivariate distributions in the RKHS.Then,the influence of measurement noise intensity on the result of fault diagnosability evaluation is analyzed by mathematical derivation.Finally,the validity of the proposed method is verified by simulation analysis.

关 键 词：故障可诊断性 定量评价 最大均值差异 可再生核希尔伯特空间 多元分布 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]