Prediction of solar irradiation in Bangladesh using artificial neural network (ANN) and data mapping using GIS technology

Abstract

In this paper, an artificial neural network (ANN) model is used to predict the monthly solar energy potential in Bangladesh. Used data are taken from NASA database for the past 22 years average from 1983 to 2005 and eight divisional cities are considered in this study. A multi-layered feed forward ANN model of four layers with eight independent input variables i.e. average temperature, sunshine duration, wind speed, precipitation, humidity, elevation, cloud coverage and atmospheric pressure to predict the monthly solar irradiation. Data from six cities are used for training and the remaining two cities were considered for testing and validation. A solar irradiation map is developed by data mapping using GIS technology. From the illustrations, the predicted data show good agreement with the observed data. This indicates that, this model can be used to predict solar irradiation of Bangladesh and to provide sufficient information about the feasibility of solar powered projects.