Solar Panel Fault Detection: Applying Convolutional Neural Network for Advanced Fault Detection in Solar-Hydrogen System at University

초록

This study uses the Samcheok Campus of Kangwon National University as a case study inside a solar-hydrogen system to provide a novel method of advanced defect detection in solar panels. Although they are essential to renewable energy sources, solar panels can malfunction for a number of reasons. The inaccuracy and inefficiency of traditional defect detection approaches have led to the investigation of more advanced methods, such as Convolutional Neural Networks (CNNs). The study entails compiling and preprocessing an extensive dataset that covers a range of solar panel fault scenarios. Using this dataset, a CNN architecture is painstakingly created and trained, utilizing methods like data augmentation and normalization to increase model resilience. Partitioning the data into training and validation sets, putting optimization tactics into practice to reduce loss, and assessing the model using metrics like data accuracy and loss, training and validation Loss, and training and validation accuracy. are all part of the training process. Specific implementation elements are covered, such as software configuration, system requirements, and difficulties experienced during training. The results demonstrate how well and consistently the CNN model can identify different kinds of solar panel faults. Faults that have been identified are examined closely, and their implications for the operation of the solar-hydrogen system are discussed. The report also discusses its shortcomings and makes recommendations for further research directions. This research, which uses the Samcheok Campus of Kangwon National University as a case study, advances fault detection algorithms for solar panels and provides insights into the campus's solar-hydrogen system, which may improve its efficiency and dependability. © 2024 IEEE.

키워드