Wind turbine short-term power forecasting method based on hybrid probabilistic neural network

Jiewen Deng, Zhao Xiao, Qiancheng Zhao, Jun Zhan, Jie Tao, Minghua Liu and Dongran Song

Energy, 2024, vol. 313, issue C

Abstract: Predicting wind power is crucial for wind farm operations and power system stability. Most existing prediction methods use cabin wind speed as the input variable, but but few of them correct the wind speed data and consider the correlation between input data. This paper proposes a hybrid probabilistic neural network model for short-term wind power probabilistic prediction, which primarily consists of two deep neural network models connected in series. The first model corrects SCADA wind speed using an LSTM neural network with mechanism information. The second model uses a self-attention mechanism to strengthen the correlation among input time series and constructed a probabilistic prediction model named SA-DeepAR. Using real wind farm data to verify results shows the corrected wind speed improves power prediction accuracy by 44 %, and the prediction accuracy of the SA-DeepAR model improved by about 15 % in RMSE and MAE compared to the DeepAR model, and by about 6 % in R2. In terms of probability prediction, the SA-DeepAR model can still maintain an average prediction interval coverage probability of 95 % at a 40 % confidence level. The proposed model can predict short-term wind power generation effectively and offer reliable data for decision-making.

Keywords: Wind turbine; Short-term power prediction; Hybrid probabilistic neural network; Wind speed correction (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (1)

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544224038209
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:313:y:2024:i:c:s0360544224038209

DOI: 10.1016/j.energy.2024.134042

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().