There can be a number of reasons for performance losses in solar power plants. Soiled modules, shadow impact, electronic failures, inverter faults or faulty tracking can all result in significant energy losses – or even shut down production completely.

Working together with battery storage business Fluence Energy, the Smart Services and Maintenance Research Group at the ZHAW School of Engineering have developed an AI-based software module that can diagnose energy losses and pinpoint the faulty component. The program can also determine when maintenance of photovoltaic systems will pay off, i.e., when costs from energy loss exceed the costs incurred by repairing the fault. The algorithms were trained using existing data of solar power plants, the laws of physics and technical expertise.

More power, lower costs

“The software can tell operators whether to call for a technician or have a gardener cut back a tree, or whether production losses are simply a result of bad weather,” says project manager Lilach Goren Huber. The new application is already market-ready and was integrated into Fluence Energy’s Nispera platform. According to Lilach Goren Huber, it is particularly useful for large-scale solar power plants. “The software module can identify faults at an earlier stage and more accurately, and thus contribute to reducing energy losses and ensuring a more cost-efficient maintenance of solar power plants.”