Document Type : Research Article

Authors

• Nadira Madjoudj ¹
• Hichem Hafdaoui ²
• Nasreddine Belhaouas ¹
• Fateh Mehareb ¹
• Houria Assem ¹

¹ Centre de Développement des Energies Renouvelables - CDER

² Centre de Développement des Energies Renouvelables -CDER

Abstract

In this paper, a Support Vector Machine (SVM) classifier is employed to analyze the degradation of photovoltaic (PV) modules in Algeria after five months of operation under moderate and humid climate conditions. The PV module examined has a nominal maximum power of 270 W. Using a comprehensive electrical and environmental dataset, the SVM model effectively classified the performance states of the module. The analysis of irradiance evolution over time shows that the peak power delivered during the day reaches approximately 249 W when solar irradiance ranges from 950 W/m² to 1050 W/m², representing about 92% of the module’s nominal power, during peak irradiation hours, the module operated under cloudy conditions for nearly 30% of the time, resulting in noticeable power fluctuations and contributing to degradation effects. The SVM-based classification enabled the creation of heatmaps that intuitively highlight degradation patterns, offering a clearer and more interpretable diagnostic tool compared to traditional analytical methods. The results demonstrate that the proposed methodology is effective for detecting degradation in individual PV modules and scalable to PV power plants, thereby supporting improved monitoring, maintenance, and performance optimization in similar climatic environments.

Keywords

• Degradation
• Maximum Power
• Monitoring
• PV modules
• SVM Classifier