AI-Enabled Fault Detection and Optimization in Photovoltaic Systems using Digital Twin Model

Photovoltaic (PV) technology is a cornerstone of sustainable energy, yet its performance is often degraded by faults such as microcracking and potential-induced degradation (PID). These faults lead to internal resistance variations, affecting power output and system efficiency. Early detection is essential to mitigate long-term performance losses. This paper proposes a machine learning-based approach to fault detection in PV panels, utilizing convolutional neural networks (CNNs) for microcrack identification through image analysis and recurrent neural networks (RNNs) for PID detection based on internal resistance variations over time. By capturing these resistance fluctuations, the proposed method enhances diagnostic accuracy and enables predictive maintenance. Experimental validation with real-world datasets demonstrates the effectiveness of this approach in improving fault detection precision, ultimately contributing to the reliability and longevity of PV systems.