Deep Neural Networks for Predicting Solar Radiation at Hail Region, Saudi Arabia

Sahbi Boubaker,Omar Kahouli,Adel Mellit,Ayoub Lefza,Mohamed Benghanem,Lioua Kolsi

doi:10.1109/access.2021.3062205

Abstract

Forecasted global horizontal irradiation (GHI) can help for designing, sizing and performances analysis of photovoltaic (PV) systems including water PV pumping systems used for irrigation applications. In this paper, various deep neural networks (DNN) models for one day-ahead prediction of GHI at Hail city (Saudi Arabia) are developed and investigated. The considered DNN models include long-short-term memory (LSTM), bidirectional-LSTM (BiLSTM), gated recurrent unit (GRU), bidirectional-GRU (Bi-GRU), one-dimensional convolutional neural network (CNN1D) and other hybrid configurations such as CNN-LSTM and CNN-BiLSTM. A dataset of daily GHI recordings collected during January 1, 2000 to June 30, 2020 from National Aeronautics and Space Administration (NASA) at an arid location (Hail, Saudi Arabia) is used to develop and compare the above DNN-based models. The parameters affecting the accuracy of the models have been also deeply analyzed. Only historical values of daily GHI have been used to build the DNN-based models whereas additional weather parameters such as air temperature, wind speed, wind direction, atmospheric pressure and relative humidity are not considered in this work. Keras library and Python language have been used to develop and compare the GHI forecasting models. The evaluation metrics such as correlation coefficient ( $r$ ), Mean Absolute Percent Error (MAPE), Mean Absolute Error (MAE), cumulative distribution function (CDF) and standard deviation ( $\sigma$ ) are opted to evaluate the performance of the prediction models. The obtained results showed that the DNN models have provided globally good performances with a maximum reached value of $r=96$ %, for daily GHI forecasting.

Highlights

Accurate forecasting of solar irradiation including global, direct, diffuse and normal is a crucial task for designing optimal solar energy systems [1].During the last three decades, numerous studies used machine learning (ML) techniques such as artificial neural networks (ANNs) and support vector machine (SVM) for forecasting global solar irradiation (GSR) [2]
Different deep neural networks (DNN) models have been evaluated for the prediction of one-day-ahead of global horizontal irradiation (GHI) in Hail region, Saudi Arabia
To show the performance of the studied DNN-based models, cumulative distribution function (CDF) is firstly plotted in figure 7

Summary

Introduction

During the last three decades, numerous studies used machine learning (ML) techniques such as artificial neural networks (ANNs) and support vector machine (SVM) for forecasting global solar irradiation (GSR) [2]. The classical methods including stochastic approaches and shallow ANNs, recurrent neural networks (RNNs) have limited capability in. With the availability of a huge amount of collected data across the globe and advances in computing technologies, researchers working in this area are more and more attracted towards deep learning (DL) techniques for developing prediction models. The DL based prediction models are playing an important role in numerous areas such as optimal scheduling of merchant-owned energy storage systems [5] and optimization of networked distributed energy resources [6].

Methods

Results

Conclusion