Feasibility study on the automatic detection of faults in photovoltaic plants with unmanned aerial vehicles using low resolution thermographic cameras

Abstract

Due to the growing trend in renewable energy production, an ever increasing number of photovoltaic (PV) systems are being installed worldwide. To ensure their efficient operation, they must be inspected and maintained regularly. Since traditional inspection methods such as electrical tests are time-consuming and costly, new methods are being developed. A promising inspection method is the so-called inspection by using unmanned aerial vehicles (UAVs) with dual camera systems (RGB and infrared). The inspection of PV modules using thermographic cameras has proven to be a suitable method for assessing the performance and condition of individual PV modules. The use of this technology offers a faster, autonomous and cost-effective inspection of large photovoltaic plants. Although there is a trend towards higher resolutions and more expensive camera systems, this thesis aims to conduct a feasibility study for the automatic inspection of PV systems using lowcost, commercially available drones with lower camera resolution. The advantages are a high availability, easy handling, a flexible applicability and a huge price advantage compared to the expensive industrial drones. The significantly lower camera resolution is a challenge in the detectability of defects by means of data accuracy. The aim of this study is thus to assess whether reliable statements about the quality and state of individual PV modules can be made employing thermography with low resolution camera systems. For this purpose, thermographic images are generated during the inspection process and subsequently verified by an automated algorithm. For conducting this thesis the following is available: - Dual camera system consumer drone (RGB and infrared) - Photovoltaic test environment for data generation - QFly postprocessing algorithm