A comprehensive study and performance analysis of deep neural network-based approaches in wind time-series forecasting

• Main Authors: Rahman M.M., Shakeri M., Khatun F., Tiong S.K., Alkahtani A.A., Samsudin N.A., Amin N., Pasupuleti J., Hasan M.K.
• Other Authors: 58831072700
• Format: Article
• Published: Springer Science and Business Media Deutschland GmbH 2024
• Subjects: NARX neural network; Recurrent neural network; Renewable energy; Time-series forecasting; Wind-speed prediction; Curve fitting; Deep neural networks; Errors; Gas emissions; Global warming; Greenhouse gases; Mean square error; Recurrent neural networks; Time series; Time series analysis; Weather forecasting; Wind speed; Forecasting: applications; Network-based approach; Neural network model; Performance; Prediction modelling; Renewable energies; Time series forecasting; Wind speed prediction; Wind time series; Neural network models

The increasing energy demand and expansion of power plants are provoking the effects of greenhouse gas emissions and global warming. To mitigate these issues, renewable energies (like solar, wind, and hydropower) are blessings for modern energy sectors. The study focuses on wind-speed prediction in energy forecasting applications. This paper is a comprehensive review of deep neural network based approaches, like the �nonlinear autoregressive exogenous inputs (NARX)�, �nonlinear input-output (NIO)� and �nonlinear autoregressive (NAR)� neural network models, in time-series forecasting applications. This study proposed NARX based prediction models in wind-speed forecasting for short-term scheme. The meteorological parameters related to wind time-series have been analyzed, and used for evaluating the performance of the proposed models. The experiments revealed the best performance of the prediction models in terms of �mean square error (MSE)�, �correlation-coefficient (R2)�, �auto-correlation�, �error-histogram�, and �input-error cross-correlation�. Comparing with the other neural network models, like �recurrent neural network (RNN)� and �curve fitting neural network (CFNN)� models, the NARX-based prediction model achieved better performance in regard to �auto-correlation�, �error-histogram�, �input-error cross-correlation�, and training time. The results also showed that the RNN and CFNN models performed better prediction accuracy with R2 and MSE values. While this performance index is slightly higher, it is negligible in forecasting applications and concluded that the proposed NARX-based model achieved the better prediction accuracy in terms of other performance evaluation measures. � 2022, The Author(s), under exclusive licence to Springer Nature Switzerland AG.