Photovoltaic power forecasting based on Elman Neural Network software engineering method

Abstract

Solar energy has the property of alternating, fluctuation and periodicity, and it has severe impact on large scale photovoltaic (PV) grid-connected generation. This turn power utilities contrary to use PV power since the forecasting and overall assessment of the grid becomes very difficult. To develop a reliable algorithm that can minimize the errors associated with forecasting the nearby future PV power generation is particularly helpful for efficiently integration into the grid. PV power prediction can play a significant role in undertaking these challenges. This paper presents 3 days ahead power output forecasting of a PV system using a Theoretical Solar radiation and Elman Neural Network (ENN) software engineering technique by including the relations of PV power with solar radiation, temperature, humidity, and wind speed data. In the proposed method, the ENN is applied to have a significant effect on random PV power time-series data, and tackle the nonlinear fluctuations in a better approach.