Solar energy foecasting using modified polynomial neural network

Abstract

In this paper, solar energy forecasting using Polynomial Neural Network (PNN) has been presented. The PNN is a flexible neural architecture whose structure is developed through different learning process. In PNN, the number of layers are not fixed. Node of the PNN are highly flexible and each node mapping among each other by realizing some polynomial type of equations such as linear, quadratic, and cubic type. Accurate forecasting of solar energy causes an efficient operations of the solar based power plants. An accurate forecasting can only make the solar energy a prominent alternative of conventional fossil fuel based power sources. In this study, the accuracy of forecasting using PNN has been observed by dividing the entire observation into training and test data. Two sets of data are used to validate the model accuracy. � 2020 IEEE.