Wind Speed Prediction Using Wavelet Decomposition Based on Lorenz Disturbance Model

ABSTRACT

As a kind of renewable energy, wind energy is getting more and more attention with its advantage of rich, clean and environmentally sustainable. Stochastic volatility is the inherent property of wind energy and the essential factor hindering the development of wind power prediction research. In this paper, by considering the actual movement rule of wind energy described by Lorenz system, the Lorenz system and the wavelet decomposition were combined to improve the wind speed prediction models of BP, RBF, and Elman neural network for the first time, and a wind speed prediction model with the Lorenz system based on wavelet decomposition was established. The model has been compared with three traditional numerical prediction methods, proving that Lorenz disturbance model can get better prediction accuracy and can grasp the actual movement of wind more accurately. The research of this paper can make up for the neglect of the atmospheric system in the field of wind speed prediction, which is helpful to the large-scale development and utilization of wind energy resources.

KEYWORDS:

• Atmospheric disturbance
• Lorenz system
• Neural network
• Prediction
• Wavelet decomposition
• Wind speed

ACKNOWLEDGEMENTS

The authors thank Distinguished Prof Shiban K Koul and the anonymous referees for the thoughtful and constructive suggestions that led to a considerable improvement of the paper.

Additional information

Funding

This research was supported partly by the National Natural Science Foundation of China [51637005], the Fundamental Research Funds for the Central Universities [2017MS166] and the Natural Science Foundation of Hebei Province [G2016502009].

Notes on contributors

Yagang Zhang

Yagang Zhang received his B.S. degree in the Department of Mathematics from Yunnan University, Kunming, Yunnan Province, China in 2001 and the M.S. degree in Center for Nonlinear Complex Systems from Yunnan University, Kunming, Yunnan Province, China, in 2004. He received the Ph.D. degree from North China Electric Power University, Beijing, in 2011. His current research topics include nonlinear complex system theory, complex networks and relay protection of power system. Corresponding author. Email: [email protected]

Chenhong Zhang

Chenhong Zhang received her B.S. degree in the Department of Mathematics and Applied Mathematics from Handan University, Handan, Hebei Province, China, in 2016 and the M.S. degree in Applied Mathematics from North China Electric Power University, Baoding, Hebei Province, China. She contributes to wind speed data preprocessing and nonlinear system prediction. Email: [email protected]

Shuang Gao

Shuang Gao received her B.S. degree in the Department of Mathematics and Information Science from Xinyang normal University, Xinyang, Henan Province, China, in 2016. Currently, she is studying for her master's degree in Applied Statistics at North China Electric Power University, Baoding, Hebei Province, China. Her current research topics include the application of neural network in the field of nonlinear prediction, promotion of clean energy in power system and the nonlinear dynamic system. Email: [email protected]

Penghui Wang

Penghui Wang received her B.S. degree in the Department of Mathematics and Applied Mathematics from Baoding University, Baoding, Hebei Province, China, in 2015 and the M.S. degree in Applied Mathematics from North China Electric Power University, Baoding, Hebei Province, China, in April 2018. Her research interests include nonlinear system prediction and application of clean energy. Email: [email protected]

Fenglin Xie

Fenglin Xie is studying for her B.S. degree in the Department of Electrical Engineering and Automation at North China Electric Power University, Baoding, Hebei Province, China. Her current research topics include the prediction theory on power system and the Matlab programming language. Email: [email protected]

Penglai Cheng

Penglai Cheng is studying for his B.S. degree in the Department of Control and Computer Engineering at North China Electric Power University, Baoding, Hebei Province, China. He will study further for master's degree in Department of Control Science and Engineering at Zhejiang University, Hangzhou, Zhejiang Province. His current research topics include the machine learning and data mining. Email: [email protected]

Shuang Lei

Shuang Lei received his B.S. degree in the Department of Logistics Engineering from Chang'an University, Xi'an, Shanxi Province, China, in 2015 and the M.S. degree in Applied Statistics from North China Electric Power University, Baoding, Hebei Province, China, Baoding, Hebei Province, China, in April 2018. His research interests include prediction method based on error correction and application of big data. Email: [email protected]