AI-driven IoT system for diagnosing and rectifying issues with solar panels

Abstract

Solar photovoltaic (PV) technology has come a long way in recent years because to its many advantages, such as being easy to maintain, reliable, renewable, and having little impact on the environment. However, various photovoltaic flaws may nonetheless manifest, leading to degradation, diminished production impact, or even a spike at various levels. This is conditional on the exterior working conditions and the usual weather changes that might damage production, delivery, or setup. Consequently, measuring the efficiency of PVS electricity generation is crucial. The "Internet of Things" (IoT) refers to a set of innovative technologies that are being rigorously studied to improve fault detection including predictive analysis in environmental monitoring including solar power plant operations. This study's results propose a machine learning-based strategy for assessing power data and anticipating problems to facilitate solar power plant management.The provided data provided by solar power facilities include both the plant's electricity output and meteorological information. The data undergoes initial pre-processing before being used for training with AI. The trained model was capable of identifying both significant and minor mistakes, along with anomalies in the provided data. Traditionally, these identifications need an increased volume of effort for detection and maintenance actions. The study's results indicate that the suggested model yielded 8.9 Mean Squares Errors, 2.98 and Root Mean Squares Errors  with the present technique.  The figures were obtained by model testing.