Prediction of Power Production from a Single Axis Photovoltaic System by Artificial Neural Networks

Abstract

The power production from a photovoltaic module depends on environmental factors such as radiation, air temperature, wind speed and direction and relative humidity. This study aimed to predict some atmospheric indicators which may affect upon the power production from a single axis (East-West) tracking PV module by using Artificial Neural Network architecture. The novelty side of this research is to use a single axis tracking PV system as first in the literature to predict effective factors of the power production. Besides, unlike previous study, many atmospheric indicators were used to estimate the power production. In this study, the performance of Artificial Neural Network is measured with Mean Square Error, Root Mean Square Error, Mean Absolute Error, Relative Absolute Error, Root Relative Square Error and correlation coefficient. In the established Artificial Neural Network model for sunny and cloudy days, the correlation coefficients are computed as 0.995 and 0.990, respectively. The models of the study will project for future estimation of power production from a single axis photovoltaic module.