Machine Learning Based Fault Detection in Photovoltaic System

This project focuses on improving the efficient and reliable operation of photovoltaic (PV) systems by detecting faults such as partial shading and dirt accumulation, which reduce power output. A machine learning-based framework is proposed to classify faults using electrical and environmental data. Supervised algorithms including Logistic Regression, Decision Tree, Random Forest, Support Vector Machine (SVM), and Artificial Neural Network (ANN) were applied. The dataset contains features like voltage, current, ambient temperature, and irradiance under normal, shading, and dirt conditions. Data preprocessing techniques such as normalization, label encoding, and data splitting were performed. Model performance was evaluated using accuracy, precision, recall, F1-score, and confusion matrix. Results show that ANN achieved the highest accuracy (above 98%). However, model performance decreased when tested on different PV systems, indicating the importance of system-specific training.