Offshore Wind Power Prediction Based on Variational Mode Decomposition and Long Short Term Memory Networks
- Resource Type
- Conference
- Authors
- Yang, Shuangmian; Zheng, Xiaoxia; Li, Meina; Wei, Yanbin
- Source
- 2021 6th International Conference on Power and Renewable Energy (ICPRE) Power and Renewable Energy (ICPRE), 2021 6th International Conference on. :1034-1039 Sep, 2021
- Subject
- Power, Energy and Industry Applications
Renewable energy sources
Uncertainty
Fluctuations
Wind power generation
Predictive models
Prediction algorithms
Feature extraction
offshore wind power prediction
long short term memory networks
extreme learning machine
variational modal decomposition
seagull optimization algorithm
- Language
- ISSN
- 2768-0525
A method of offshore wind power prediction based on VMD and LSTM is proposed to improve the accuracy of offshore wind power prediction. For reducing the volatility of offshore wind power data to extract the salient features of the power signal, a variational modal decomposition algorithm optimized by the seagull algorithm is used to decompose the power series into several eigenmodal components and a residual component that may contain uncertainties affecting the power prediction; the decomposed several eigenmodal components are built into a prediction model based on the long and short-term memory network, and the residual component is built into a prediction model based on the extreme learning machine. The final offshore wind power is obtained by superimposing and integrating the prediction results of each component. It is shown that the combined model can predict both the slow fluctuation of offshore wind power and the drastic change of power in the climbing section compared with the single model and BP neural network.