A new method for identifying redundant sensors based on completely consistent fault isolability properties. (March 2023)

• Record Type: Journal Article
• Authors: Chen, Changxiong; Chen, Liping; Cheng, Xianjuan; Ding, Jianwan
• Is Part Of: European journal of control. Volume 70(2023)
• Publication Date: 2023-03
• Subjects: Model-based fault diagnosis -- Structural analysis -- Fault isolability -- Sensor placement -- Potential faults; Control theory -- Periodicals; Automatic control -- Periodicals; Automatic control -- Mathematics -- Periodicals; Electronic journals; 629.805
• DOI: 10.1016/j.ejcon.2022.100766
• Languages: English
• ISSNs: 0947-3580

Abstract

Highlights: Fault isolability properties of the related systems are completely consistent. The same ability to isolate both the known faults and potential faults. Removing the sensor does not change the fault isolability properties. Abstract: This paper presents a new method to disclose redundant sensors from the viewpoint of the completely consistent fault isolability properties. If a sensor is removed from a diagnosis system, and the updated system also can isolate the same known faults and the potential faults, then the removed sensor is a redundant sensor based on completely consistent fault isolability properties. Through structural model decomposition, if a set of specific variables derived from the updated system is the same as the one obtained from the diagnosis system, then the fault isolability properties of the two related systems are the same for the known faults and the potential faults. Based on this criterion, the corresponding algorithm is provided and all of these redundant sensors can be extracted. It is useful for fault-tolerant control design and optimal sensor placement. Finally, the effectiveness and feasibility of the proposed method are validated by the simulations of a four-tank system.