Cloud images capturing system for solar power level prediction

Abstract

Solar energy has received increasing attention as one of the potential renewable energy sources for power generation in recent past. Introduction of Net Metering and the increment in provision for renewables encouraged the usage of Solar PV systems in Sri Lanka. However, the intermittent nature of solar energy has become one of the barriers for solar energy based power to be integrated to the national power grids. Due to unpredictability solar energy based power plants are non-dispatchable and can cause network instability. With an efficient and reasonably accurate predictable model, a better system balance can be achieved. Shadowing on solar PV modules results in reduction of power produced. Cloud coverage blocking the sun can be identified as the major contributor in shadowing. Identifying and tracking the clouds can be used to finally predict the solar PV output. This paper presents a methodology to obtain cloud image data and an algorithm to process the images which can be used to predict the relationship between the cloud movements and the solar PV output.