Search documents

Browse topics

Document details

Effect of Availability and Quality of Data on the Detection of Defects Utilizing Artificial Neural Networks in PV System’s Monitoring Data
D. Daßler, S. Malik, R. Gottschalg, M. Ebert
Monitoring, PV System, Machine Learning, Data Quality, Artificial Neural Networks
PV Systems Engineering, Integrated/Applied PV
Subtopic: Operation, Performance and Maintenance of PV Systems
Event: 8th World Conference on Photovoltaic Energy Conversion
Session: 4DO.1.5
1074 - 1079
ISBN: 3-936338-86-8
Paper DOI: 10.4229/WCPEC-82022-4DO.1.5
0,00 EUR
Document(s): paper


For analyzing high volumes of operational data from grid-connected PV systems, meanwhile, algorithms help to detect exceptional deviations from normal behavior. Machine learning methods in particular are used to overcome complex relationships within the data. One category are artificial neural networks, which are utilized to map the specific system performance as a function of environmental parameters. This paper investigates the influence of the data used for training on the detectability of yield deviations. Based on four different training scenarios the effect of missing data, the selection of irradiance sensor data, the data resolution as well as the monitoring resolution are analyzed in detail. Recommendation in respect to spatial irradiance sensor coverage and data aggregation will be given. This paper will explain the approach of defect detection, the data modification as well the impact of training quality on the detectability.