Aerospec recently inspected a 100MW solar farm in the U.S. which consists of approximately 450,000 modules. With its proprietary image processing technology and detection algorithm, Aerospec’s inspection report provided immediate actionable items to replace faulty panels, Li notes.
Aerospec determined that the solar farm had 3,500 faulty PV modules, worth approximately $1 million in replacement cost, which will be covered within the warranty period. If left unnoticed, these faulty PV modules would have resulted in close to a $350,000 annual revenue loss.
Using simple GPS and altitude information from a UAV has proven ineffective because thermal imaging cannot be adequately performed at a 90-degree angle. To overcome this challenge, engineers at Aerospec developed proprietary image-processing software that uses metadata with advanced computer vision and machine learning algorithms to precisely pinpoint faulty panels.
In the future, historical inspection data will be analyzed through cross-sectional and time-series comparison to identify performance hurdles and provide predictive analytics, Li says.
According to the Electric Power Research Institute (EPRI), drones can perform a variety of tasks – including visual imaging (of modules, wiring, and other plant components), infrared thermography, and vegetation monitoring – that have the potential to update largely manual conventional processes and more efficiently identify and diagnose PV system performance issues.