Sensor fault detection based on fuzzy singularly perturbed model. Issue 8 (11th June 2022)

• Record Type: Journal Article
• Authors: He, Chunxiao; Chen, Jinxiang; Li, Xisheng
• Is Part Of: International journal of systems science. Volume 53:Issue 8(2022)
• Page Start: 1690
• Page End: 1705
• Publication Date: 2022-06-11
• Subjects: Fuzzy singularly perturbed model -- sensor fault detector -- nonlinear singularly perturbed systems -- LMIs -- H∞; System analysis -- Periodicals; 003.3
• DOI: 10.1080/00207721.2021.2020368
• Languages: English
• ISSNs: 0020-7721

Abstract

Abstract : Fuzzy singularly perturbed modelling and sensor fault detection for nonlinear singularly perturbed systems (NSPSs) are investigated. A continuous-time fuzzy singularly perturbed model with superimposed faults on state variables and external disturbances are established to describe NSPSs with sensor faults. Based on the model, a novelty H ∞ fuzzy fault detector (FFD) is designed, and the sufficient conditions for H ∞ performance are derived. The above approaches are applied to a CE150 helicopter with abrupt and intermittent faults on sensors. The innovations of this paper are as follows: (i) The nonlinear and slow–fast time-scales are described by a united continuous-time fuzzy singularly perturbed model, and based-it a FFD is designed, which can avoid the incomplete decoupling problem caused by decomposition methods; (ii) The sensor faults are expressed by the superimposed faults on state variables when the continuous-time fuzzy singularly perturbed model is constructed, which is less conservative than existing methods; (iii) In the FFD design, the first-order derivative model of the estimate function for sensor faults is established, and it is combined with a fuzzy observer, which can improve the detection accuracy of the FFD; (iv) H ∞ control technology is introduced into the FFD design process to suppress external disturbances.