%0 Conference Proceedings %T Data-driven approach for SOH estimation and alarms generation for complex on-grid energy storage systems %+ Département des Technologies Solaires (DTS) %+ Institut National de L'Energie Solaire (INES) %A Karoui, Fathia %A Ha, D-L. %A Delaplagne, T. %A Bouaziz, M-F. %A Vinit, L. %A Montaru, M. %< avec comité de lecture %@ 3-936338-60-4 %( Proceedings of the 36th European Photovoltaic Solar Energy Conference and Exhibition %B EU PVSEC 2019 - 36th European Photovoltaic Solar Energy Conference and Exhibition %C Marseille (FRANCE), France %8 2019-09-09 %D 2019 %K State of health %K Lithium-Ion battery %K data-driven %Z Engineering Sciences [physics]Conference papers %X On-grid energy storage systems are used increasingly worldwide in order to optimize the use of intermittent energies such as photovoltaics. An accurate estimation of the state of health of the batteries is necessary to optimise their lifetime and reduce their Levelized Cost of Storage. The misuse of these batteries leads to frequent failures. A robust analysis of the misuse events with automatic alarms generation will also be valuable to send warnings to the system end-user before failure. This paper presents advanced analysis of some of these systems based on new approaches of data-driven diagnosis and prognosis. Several PV-Storage systems have been monitored for four years and an original diagnostic and prognostic tool is developed for the analysis of the performance and defaults of such systems. This generic approach allowed to have a feedback on the performance of grid connected PVstorage systems with two storage technologies (Li-ion and NiNaCl2). Recently, the efficiency of this kind of systems have been analysed including the performance of batteries and power conversion systems. The analysis doesn’t include the state of health evolution and the alarms generation. %G English %2 https://cea.hal.science/cea-04057307/document %2 https://cea.hal.science/cea-04057307/file/5CO.14.5_EUPVSEC_presentation_FKAROUI.pdf %L cea-04057307 %U https://cea.hal.science/cea-04057307 %~ CEA %~ UNIV-SAVOIE %~ UGA %~ CNRS %~ DSV %~ DRT %~ CEA-DRF %~ LITEN %~ CEA-GRE %~ USMB-COMUE %~ INES