Development of a convolutional neural network model for solar panel fault detection with preventive maintenance reporting options

As the global demand for renewable energy continually increases, operational efficiency and reliability of photovoltaic (PV) systems have become increasingly crucial. Solar panels or photovoltaic modules are susceptible to various anomalies such as cracks, electrical damage, and hotspots, which can significantly reduce their energy output if not detected early. Convention al or manual inspections are inadequate and prone to human error, especially in large-scale solar farms. This paper introduces the development of a Convolutional Neural Network Model for Solar Panel Fault Detection with a Preventive Maintenance Reporting System, which uses machine learning, thermal imaging, and cloud-based data processing to identify real-time faults. The system architecture integrates high-resolution visual cameras and thermal cameras with a Raspberry Pi4 microcontroller, the core of the system is a deep learning model based on a lightweight Convolutional Neural Network (CNN) analysis captured images to identify defects such as cracks, hotspots, and electrical damages. This data is processed in the cloud, where the anomaly detection model classifies faults with an overall accuracy of 95% with user-friendly web dashboard for monitoring and reporting, preventive maintenance measures. The results show that the CNN model significantly enhances inspection accuracy, reduces operational downtime, and lowers maintenance costs by identifying issues early.