The general objective of the PERSI team is to guarantee the operational availability and sustainability of industrial system equipment. This integrated problem requires in a first phase to detect and isolate the faults before failure (progressive faults and cataleptic faults) then in a second step, to predict (anticipate failures) and finally, in a third step, to estimate the remaining useful life or manage operation in degraded mode. Based on literature review, the three phases are treated mainly independently., the three phases are treated independently.
The general theme that characterizes the PERSI team consists of considering the different aspects of sustainable systems design in a simultaneous and consistent way, namely modeling (functional and behavioral), diagnosis, prognosis of energy systems together with their computational implementation.
The tools used are based on graphical formalisms and multiphysics (Event Driven Hybrid Bond Graph, functional graphs) Artificial Intelligence and formal modeling tools. The main application domain concerns renewable energies (green hydrogen, multisource systems, fuel cells, etc.), transport and production systems.
The areas of interest mainly cover the following topics:
• Knowledge and data (particle filters and learning) and hybrid models based for diagnosis and prognosis.
• Multiphysics, Hamiltonian, hybrid models and structural analysis for integrated design of supervision systems.
• Optimal operating modes management of the of hybrid systems.
• AI for health status estimation and robust decision making.
• Software prototyping and formal techniques for automated diagnosis model builder (AMESIM, Modelica, FDIpad ….).
Belkacem Ould Bouamama
Diagnostic et supervision d'une station de rechargement à base d'énergies renouvelables pour une flotte de véhicules partagés fonctionnant à 'hydrogène et l'électricité
LFT Bond Graph et ensemble Machine Learning pour le diagnostic et le pronostic des systèmes non linéaires. Applications aux systèmes de production d'hydrogène vert
Optimal design of a multi-energy system applied to a commercial building - COSMAC
Jumeau numérique à base d'apprentissage profond pour convertisseurs de puissance