A data-driven fault-tolerant control design of linear multivariable systems with performance optimization. (September 2017)

• Record Type: Journal Article
• Authors: Li, Zhe; Yang, Guang-Hong
• Is Part Of: ISA transactions. Volume 70(2017:Sep.)
• Page Start: 200
• Page End: 208
• Publication Date: 2017-09
• Subjects: Residual generator -- Fault-tolerant control (FTC) -- Data-driven -- Adaptive scheme.; Engineering instruments -- Periodicals; Engineering instruments; Periodicals; Electronic journals; 629.805
• DOI: 10.1016/j.isatra.2017.07.021
• Languages: English
• ISSNs: 0019-0578

Abstract

Abstract: In this paper, an integrated data-driven fault-tolerant control (FTC) design scheme is proposed under the configuration of the Youla parameterization for multiple-input multiple-output (MIMO) systems. With unknown system model parameters, the canonical form identification technique is first applied to design the residual observer in fault-free case. In faulty case, with online tuning of the Youla parameters based on the system data via the gradient-based algorithm, the fault influence is attenuated with system performance optimization. In addition, to improve the robustness of the residual generator to a class of system deviations, a novel adaptive scheme is proposed for the residual generator to prevent its over-activation. Simulation results of a two-tank flow system demonstrate the optimized performance and effect of the proposed FTC scheme. Highlights: The canonical form identification technique is introduced to design the data-driven residual generator. The Youla parameters are tuned iteratively based on the online system data for fault-olerant control. A novel adaptive law is developed for the proposed residual generator to avoid the over-activation of the FTC mechanism.