Anupong Banjongkan, Nittaya Kerdprasop and Kittisak Kerdprasop

 

School of Computer Engineering, Suranaree University of Technology, Nakhon Ratchasima, 30000, Thailand

---

Abstract
Renewable energy has been a hot topic recently, especially wind power which has grown considerably in the past decade. The forecast of wind speed in advance is important information for wind power plant management. In this paper, a high-efficiency time series model for forecasting wind speed day-ahead is proposed, developed from the nonlinear hybrid model called Adaptive Neuro-Fuzzy Inference System (ANFIS). It brings together the advantages of fuzzy and neural network learning. In addition, a comparative study was done with Autoregressive Integrated Moving Average (ARIMA) and other nonlinear time series models including Artificial Neural Network (ANN) and Long Short-Term Memory (LSTM) models. The realistic data from the meteorological data of Chaiyaphum province, Thailand were used in this research. The dataset was split into learning and testing data in the ratio of 75% and 25%, respectively. The result shows that the forecasting performance of the ANFIS model was comparable to the ARIMA model. Both models achieve high accuracy than other neural network models. The proposed model achieves high efficiency at 22.89 MAPE and 0.41 of R2. Interestingly, the ANFIS model has a learning time faster than ANN and LSTM models by at least 100 times.

Keywords
ANFIS, fuzzy, neural network, renewable energy, time series

---

Copyright © 2023 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.