Application of Data Science and AI for Health Monitoring and Economic Assessment of Photovoltaic Systems

Abstract

Energy is crucial for societal advancement, and the shift to renewable sources like photovoltaic (PV) systems mitigates the negative impacts of fossil fuels. This study applies machine learning and artificial intelligence, specifically image segmentation and object detection, to PV module health monitoring. We use state-of-the-art segmentation techniques to isolate solar panels and object detection to identify anomalies within the modules. Three publicly available datasets are employed to validate our approach, covering semantic segmentation and object detection tasks. We compare our model's performance against an existing benchmark, highlighting its strengths and areas for improvement. Additionally, we perform a basic economic assessment of implementing the proposed AI models in a real-world PV project, considering costs and potential savings.

Date of Award	Apr 2023
Original language	American English
Supervisor	Ameena Al Sumaiti (Supervisor)