Identification of Partial Shading Patterns in Photovoltaic Arrays Using RF Algorithm with Integrated Data

This research presents a machine learning-based methodology for detecting and analyzing partial shading scenarios in photovoltaic (PV) systems. Utilizing an experimental setup of a 3x2 PV array configuration alongside a MATLAB/Simulink model, a comprehensive dataset was developed. To mitigate class imbalance, the dataset was enhanced using the Synthetic Minority Over-sampling Technique (SMOTE). An optimized Random Forest (RF) classifier was then evaluated against three alternative models: K-nearest neighbor (KNN), decision tree (DT), and gradient boosting classifier. The RF classifier outperformed the others, achieving high accuracy in binary and multiclass classifications. This approach enhances the performance of monitoring and energy optimization in PV systems, contributing to advancements in renewable energy applications.