[Translate to Englisch:]

The Environmental Informatics Technical Committee of the German Informatics Society (Gesellschaft für Informatik e.V.) awarded the Environmental Informatics Student Prizes at the 35th International Conference on Environmental Informatics and Communication Technologies (EnviroInfo 2021). The first prize went to Philipp Akharath, Jaqueline Altkrüger and Harkiran Sahota from Ulm University of Technology.

As part of the project module in the master's degree program "Intelligent Systems" at THU, the students investigated the extent to which machine learning algorithms are suitable for detecting operational faults in photovoltaic systems at an early stage. A comparison with other methods showed that anomaly detection using the "Isolation Forest" method could reliably identify defects in photovoltaic systems.

Photovoltaics play a central role in transforming the global energy system from fossil fuels to an emission-free energy supply. However, especially for smaller rooftop systems, it has been shown that technical faults can occur due to a lack of remote monitoring of the energy yield and a lack of a quality assurance concept, which often remain undetected for a long time. This leads to a reduction of the yield of these photovoltaic systems, which is why early detection of malfunctions is important.

The publication was written under Prof. Dr. Volker Herbort (Ulm University of Applied Sciences) and Prof. Dr. Henrik te Heesen (Trier University of Applied Sciences, Environmental Campus Birkenfeld).