TINYML FOR SOLAR PANELS: BRINGING EDGE COMPUTING APPLICATIONS TO SOLAR ENERGY SYSTEMS

Abstract

This work comprehends the knowledge provided by previous work in the field of deep learning for solar energy systems. Furthermore, it recognises the need for edge computing applications to replace the existing state of the art. When we talk about solar energy systems, one of the utilities involved in giving large amounts of energy is solar farms. However, the size of these farms makes them remote and inaccessible. Thus causing a demand for an in-house device for monitoring and other data analytical applications. The applications of edge computing for Solar Power Systems that will be explored for this work are Identification and Estimation of Power Loss of Solar Panels, Classification of Soiling on Solar Panels and Solar Irradiance Forecasting through Sky Images.