Nonlinear observer-based fault diagnosis in photovoltaic systems integrated with voltage source converters using an extended Kalman filter
Résumé
With the rapid penetration of renewable generation, the stability and operability of photovoltaic (PV) systems increasingly depend on reliable fault diagnosis mechanisms.This work develops a nonlinear observerbased diagnostic framework for PV systems interfaced through a voltage source converter (VSC).The approach formulates a nonlinear state-space model of the PV converter system and deploys an Extended Kalman Filter (EKF) for real-time state estimation and sensor fault detection, targeting voltage and current measurements under varying operating conditions.The contribution does not rest on the EKF formulation itself, but on its integration within a unified estimation diagnosis control structure.The estimator generates residuals that enable rapid fault detection while simultaneously supporting measurement correction to maintain control performance.This coupling ensures continuity of maximum power point tracking (MPPT) even in the presence of sensor degradation.Simulation results demonstrate a reduction of estimation error by up to 99.9% for both voltage and current states, with fault detection occurring within milliseconds.The MPPT efficiency remains above 95% under faulty conditions, while output power exhibits improved stability compared to non-diagnosed configurations.The inclusion of the VSC further enhances dynamic response and robustness to environmental variability.These results confirm that the proposed observer-based architecture strengthens diagnostic reliability while preserving system performance in real time.
Citer ce document
Accès au document
Texte intégral en lecture en ligne, réservé aux abonnés SPHAERO et aux membres de l'institution. Se connecter
Voir l'article sur le site de la revueStatistiques
Consultations : 1
Téléchargements : 0