Sufficient Structural Conditions for Diagnosability and Heuristic Diagnoser Design in Timed Continuous Petri Nets

This work is concerned with a structural characterization of the diagnosability property in Timed Continuous Petri Nets (TCPNs) systems under infinite server semantics. Regarding this problem, three novel results are presented. The first one is the introduction of structural sufficient conditions for diagnosability in TCPNs, which are based on the concepts of relative degree, system distinguishability, and parameter identification. To this aim, the concept of a simple directed path from other works is extended to diagnosable directed paths . These new paths include attribution-places, pre-join-places and post-join-places , which allows to deal with multiple non-concurrent tokens-leak faults, and enlarging the class of systems that can be analyzed. Based on these structural conditions, a novel methodology to place a reduced number of sensors that guarantees the net diagnosability is proposed. Finally, a diagnoser based on a modified Differential Evolution algorithm is presented, which introduces individual searching sets in orthogonal spaces to diagnose (detect, locate, and identify) faults when an error is detected, avoiding the use of a bank of diagnosers of other approaches. The effectiveness and applicability of the main results are illustrated through an illustrative example.