5BV.2.34 DRONE-BASED ASSESSMENT OF CLEANING EFFECTS ON PV INSTALLATIONS Manuel Lanz, Eva Schuepbach, Urs Muntwyler Bern University of Applied Sciences BFH, Department of Engineering and Information Technology Institute for Energy and Mobility Research, Photovoltaic Laboratory (PV LAB) Jlcoweg 1, CH-3400 Burgdorf, Switzerland Tel. +4134 426 68 37, Fax +4134 426 68 13, urs.muntwyler@bfh.ch ABSTRACT: Monitoring the performance of PV installations can be very labour intensive. The efficiency can significantly be increased by using a thermal imaging drone system. The Photovoltaic Laboratory (PV LAB) at Bern University of Applied Sciences BFH in Burgdorf, Switzerland, hence developed and built its own thermal imaging drone. Drone-based inspection of several Swiss PV installations operated by the PV LAB was conducted and degradation (k G ) factors were calculated. Comparison of the k G -factors evidenced that cleaning the PV module surfaces may increase the energy yield production up to 9%. The drone-based thermal images offered a real advantage for the detection of “hot spots” in PV modules and associated differences in solar cell temperatures, indicating power decreases. However, the quantification of extra energy yields gained from cleaning homogeneous environmental pollution on PV module surfaces needs to be based on measurements. Keywords: Monitoring, Performance, PV System, Soiling, Thermography 1 INTRODUCTION Reductions of electrical output performance of PV modules due to environmental contamination (e.g., dust) can be significant [1,2,3]. Irregular and inhomogeneous contamination of PV modules may generate a mismatch of the cells in the substring. The affected cells may heat up and cause thermal stress, which can result in potential defects. As the inspection of such defects, as well as the overall monitoring of the performance and quality of (especially large) PV installations are time-consuming, unmanned aerial drone vehicles (UAV) can provide an added value. An infrared multicopter drone was hence developed and built at the Photovoltaic Laboratory (PV LAB) at Bern University of Applied Sciences BFH in Burgdorf-Switzerland. Using this drone, several PV installations in Switzerland were inspected with regard to performance quality. The inspected PV installations are part of the monitoring network operated by the PV LAB at BFH since the 1990es. At some specific PV installations, the energy yield output was quantified before and after cleaning the PV module surfaces using installed temperature and irradiation sensors as input data, and by measuring transformers at the PV system. 2 THE IR-MULTICOPTER DRONE In order to facilitate the maintenance, the drone system was assembled from commercially available components. The measuring device is a thermal imaging camera with a resolution of 382 x 288 pixels. Combined with a 62°x49° wide-angle optics, the viewing angle is 3.14 milliradian, i.e., the smallest measurable spot describes a square with 31.4 mm edge length in 10m distance. A fully radiometric video can be recorded, as reflections are easier recognizable on the video than on thermal images. Reflections are also more distinguishable from the thermal abnormalities in the video. Hence, a live video stream is sent to the pilot to check the image. As the pilot can switch between two video signals for easier flight navigation, rapid overflight of a PV installation is possible. A digital compact camera is mounted for capturing comparison shots. More technical and operational details of the IR-multicopter drone system can be found in [4]. Figure 1 shows the drone ready for take-off. Figure 1: The IR-multicopter drone assembled by the Photovoltaic Laboratory (PV LAB) at Bern University of Applied Sciences BFH in Burgdorf, Switzerland (model S1000 octocopter with camera-system). Photo: BFH-TI. Figure 2 shows a video capture of the IR camera mounted on the drone. At the left hand side in Figure 2, a Siemens M55 module of the PV LABs own PV installation displays a thermal abnormality (“hot spot”). Figure 2: IR video capture, magnified (left) and close-up of marked module (right). Photo: BFH-TI. The right hand side in Figure 2 shows a close-up of the same module and reveals that the module has a bad solder joint (at the top of the module) and a short circuit in a cell (bottom right). The deficient module was