The main aim of PVgnosis is to create an ICT Platform and associated tools needed for delivering advanced diagnosis, predictive maintenance and intelligent visual inspection on installed PV plants. In-depth, real-time information will be provided to the Plant Operator in terms of individual panel operation, shading and thermal strain from sensory equipment on-site via heterogeneous sources, such as sensory equipment on PV panels and BoS components, drone-thermal cameras, plant’s historical data and cloud-based services. Image processing & machine learning-based pattern recognition techniques will process the obtained real-time and historical data and train an automated decision making mechanism; delivering recommendations to optimize plant operation and maintenance (predictive maintenance, needs for replacement, cleaning, most cost-effective operation management, etc.). Through the advanced data processing, early signs of problems will be revealed, such as solar cell potential induced degradation or PID, which can lead to power losses of up to 50% and may be difficult to spot at very early stages within traditional operations and maintenance schedules. On the inverter-side, a Fault Detection and Accommodation scheme will be introduced, for detecting and isolating faults that affect the PV inverter. It will then be embedded into an actual inverter-controller and feed-in relevant information to the central Platform as required (though the development of proper APIs). Additionally, the Inverter functionalities will be enriched with new control modes and capabilities (such as novel fault ride through support and phase balancing services) for enabling the actual ancillary services’ delivery from the PV plant infrastructure. The associated business models will be established by the project as well. As such, PVgnosis is expected to deliver a holistic, end-to-end O&M ICT solution for established and new PV plants worldwide.
