Intermittent fault manifestability of discrete event systems

• Published in: Soft computing (Berlin, Germany) Vol. 27; no. 11; pp. 6999 - 7009
• Main Authors: Liang, Ye, Lai, Aiwen, El-Meligy, Mohammed A., Sharaf, Mohamed
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.06.2023
• Subjects: Algebraic; Analytical Methods in Soft Computing; Artificial Intelligence; Computational Intelligence; Control; Engineering; Foundation; Mathematical Logic and Foundations; Mechatronics; Robotics; Manifestability; Intermittent fault diagnosis; Diagnosability; Discrete event system
• ISSN: 1432-7643; 1433-7479
• DOI: 10.1007/s00500-023-08030-1

The diagnosability problem has been extensively studied over the past two decades for discrete event systems associated with intermittent faults. Diagnosability is a property that requires a large number of sensors for fault detection, which leads to expensive system deployment. To alleviate the above drawbacks, we propose a new property called intermittent fault manifestability, which is a weaker observation requirement for diagnosing the occurrence of intermittent faults. In order to verify the property, we consider fault events and their corresponding reset events and introduce two properties of f -manifestability and r -manifestability, respectively. Based on an intermittent fault pattern automaton, we first propose a diagnoser structure to track the occurrence of intermittent faults and prove that the verification of intermittent fault diagnosability can be transformed into the problem of detecting an intermittent fault language. Then, we present a pair verifier constructed by the strict composition of a fault diagnoser and a normal diagnoser, which are the subparts of the diagnoser with respect to the fault events and the normal events, respectively. Finally, we introduce necessary and sufficient conditions for verifying intermittent fault manifestability.