An Open loop time series ANN model for forecasting solar insolation for standalone PV applications

Abstract

A standalone DC Microgrid comprising PV as a distributed generator has gained popularity as it gives a promising solution for pollution control and supplies increasing DC loads. The intermittent nature of PV gives rise to challenges in energy management. Hence a system that aids in making appropriate decisions in energy management is essential. In this regard, a system that forecasts solar insolation accurately is imperative to guarantee uninterrupted energy supply to the critical loads. The existing closed loop Artificial Neural Network model developed for predicting solar insolation is costly and complex. Hence, the authors propose an open loop time series Artificial Neural Network model that is simple and economical with comparable accuracy. Bayesian Regularization algorithm is recommended. The model's performance is validated by measuring the Root Mean Square Error and coefficient of Regression.