Comparative Analysis of Fault Detection Algorithms for Photovoltaic System

Solar energy is a well-known renewable energy source, and in recent years, PV systems have played a crucial role in converting solar energy to electricity. Quick action is needed to resolve any operational problems with PV systems if energy conversion efficiency is to be ensured. Therefore, a reliable method for quickly identifying and fixing problems is essential. This study focuses on fault analysis, including shading anomalies and various fault types, in PV arrays. To address this, an Artificial Neural Network (ANN) approach is employed, leveraging its capacity to discern complex patterns. The investigation aims to scrutinize non-linear I-V curve patterns in solar panels, using a range of fault scenarios. The research employs diverse machine learning models, including Decision Trees, Naïve Bayes Classifier, MLP Neural Network, Random Forest, and LSTM, to identify and categorize fault types. The study follows a systematic approach, encompassing data collection, preprocessing, simulation, model development, and performance evaluation. Validation against real PV plant data enhances the reliability of the outcomes. Through this investigation, the study seeks to contribute to the field by advancing fault analysis techniques for improved PV system performance and longevity.