引用本文:马广富,高升,郭延宁.一类伴有部分解耦干扰的非线性系统故障诊断[J].控制理论与应用,2024,41(2):240~248.[点击复制]
MA Guang-fu,GAO Sheng,GUO Yan-ning.Fault diagnosis design for nonlinear systems corrupted by partially decoupled disturbances[J].Control Theory and Technology,2024,41(2):240~248.[点击复制]
一类伴有部分解耦干扰的非线性系统故障诊断
Fault diagnosis design for nonlinear systems corrupted by partially decoupled disturbances
摘要点击 4032  全文点击 350  投稿时间：2022-04-18  修订日期：2023-11-24
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  10.7641/CTA.2023.20288
  2024,41(2):240-248
中文关键词  故障检测  故障估计  H?指标  L∞指标  有限频域  Lipschitz非线性系统
英文关键词  fault detection  fault estimation  H? index  L∞ index  finite-frequency domain  Lipschitz nonlinear systems
基金项目  中国科学院战略性先导科技专项(A类)子课题项目(XDA14030303), 国家自然科学基金项目(61973100, 61876050)
作者单位E-mail
马广富 哈尔滨工业大学  
高升 哈尔滨工业大学  
郭延宁* 哈尔滨工业大学 guoyn@hit.edu.cn 
中文摘要
      针对Lipschitz非线性系统执行器故障检测和传感器故障估计问题, 本文提出了一种基于H?/L∞未知输入观测器的有限频域故障诊断策略. 首先, 将系统处理成包含传感器故障的增广系统. 然后, 将该系统的未知输入干扰分为可解耦与不可解耦两部分. 针对可解耦部分, 利用观测器匹配条件将其从估计误差中消除. 针对不可解耦部分,设计L∞指标抑制其对残差的影响并结合有限频域H?指标提高执行器故障检测灵敏度. 接着, 给出观测器存在的充分条件并将其转化为受LMIs约束的线性优化问题, 实现了执行器故障的鲁棒检测及传感器故障的鲁棒估计. 最后, 结合仿真算例验证了所提方法的正确性与有效性.
英文摘要
      : A finite-frequency domain fault diagnosis scheme based on the H?/L∞ unknown input observer (UIO) is proposed for Lipschitz nonlinear system actuator fault detection and sensor fault estimation. First, an augmented system is developed by constructing an augmented state composed of system states and sensor faults. Then, the unknown input disturbances are divided into decoupled disturbances and the disturbances that cannot be decoupled. For the decoupled disturbances, the observer matching condition is used to eliminate them from the estimation error. The L∞ index is designed to attenuate the influence of disturbances that cannot be decoupled and the finite-frequency domain H? index is used to improve the actuator fault detection sensitivity. Moreover, sufficient conditions for the proposed observer are derived and converted into a linear optimization problem constrained by linear matrix inequalities (LMIs). Robust detection of actuator faults and robust estimation of sensor faults are achieved. Finally, the correctness and effectiveness of the developed observer are validated through simulation examples.