Data-Based Solar Radiation Forecasting with Pre-Processing Using Variational Mode Decomposition

• Published in: International Conference on Control, Decision and Information Technologies (Online) pp. 2061 - 2066
• Main Authors: El Bakali, Saida, Ouadi, Hamid, Giri, Fouad, Gheouany, Saad, El-Bakkouri, J.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 03.07.2023
• Subjects: Buildings; Data analysis; Data preprocessing; Feature extraction; Prediction algorithms; Predictive models; Production
• ISSN: 2576-3555
• DOI: 10.1109/CoDIT58514.2023.10284151

This paper presents a hybrid method for accurately predicting Global Horizontal Irradiance (GHI) over the following 24 hours to forecast energy production from a photo-voltaic system in a positive energy building. The input data is preprocessed using the Variational Mode Decomposition (VMD) method to extract wide-bandwidth features and decompose them into smooth modes focused on specific frequency ranges. The Salp Swarm Algorithm (SSA) is utilized to identify the optimal VMD parameters for accurate extraction. The data analysis is employed to identify the most critical modes of input features. The model's efficiency is further enhanced by performing a residual preprocessing step between the observed solar radiance data and the decomposed modes. The Stacking technique (ST) is employed to predict the 24-hour GHI modes and the residual, which are summed to reconstruct the final signal. The proposed method's performance is evaluated using the Normalized Root Mean Square Error (NRMSE) and Normalized Mean Absolute Error (NMAE) metrics on three years of available data (2019-2022) in Rabat, and compared with the model based on raw data. The results show that the proposed method achieved promising results with an NRMSE of 1.35% and NMAE of 0.82% on a cloudy day.