Solar irradiance and pv forecasting model based on satellite observations

Abstract

The aim of the project is to make short term forecasting model for India, using INSAT 3D satellite data. We made probabilistic model which gives solar irradiance forecasting for every thirty minutes and one hour. Model is made using Gaussian process regression (GPR). Clearness index is used as an input of model for better results and uncertainty reduction. We estimated the power generation using forecasted solar irradiance data and validate it with ground data and discussed parameters effecting on ground data which creates huge difference between satellite estimated power and ground data.