Validation of Industrial Automation Systems Using a Timed Model of System Requirements

Abstract

Validation of industrial automation systems is the process of checking that commissioner requirements are successfully implemented. Formal approaches are needed when the considered system is critical. The method presented in this article relies on a model-based approach that exploits the compactness and graphical representation of time-interpreted Petri nets, which adds input–output interpretation to transitions/places and embeds time information. These nets are here used with multiple-server semantic to allow effective modeling of typical automation system requirements. The key idea of the system validation approach is to compare the observed behavior of the automation system with the expected behavior, as generated by updating online the model of system requirements using a state estimation algorithm. Also, an off-line procedure is provided to evaluate the evolutions admitted by the model but not observed. Both procedures yield useful data to the validation engineer, allowing to speed up the validation process. Technological issues due to the synchronous nature of controllers and the implications of their programming are considered.