Attention-LSTM Based on Nuclear Principal Component Analysis Ultra-Short-Term Forecast of Wind Power

HUANG Yafeng; HAO Jianfeng; SUN Boyan; HUANG Jiuhong; DONG Shuo; SUN Jiaqing#br#<div>	#br#</div>

doi:10.19718/j.issn.1005-2992.2025-01-0095-09

English Version Reader LoginJournal Login

您当前的位置：

首页 >

文章列表页 >

Attention-LSTM Based on Nuclear Principal Component Analysis Ultra-Short-Term Forecast of Wind Power

更新时间：2025-10-16

- Attention-LSTM Based on Nuclear Principal Component Analysis Ultra-Short-Term Forecast of Wind Power
- Journal of Northeast Electric Power University Vol. 45, Issue 1, (2025)
- 作者机构：
  
  东北电力大学电气工程学院,吉林,吉林,132012
- 作者简介：
- 基金信息：
- DOI：10.19718/j.issn.1005-2992.2025-01-0095-09
  CLC：
- Published Online：28 February 2025，
  
  Published：2025
- 稿件说明：
移动端阅览
HUANG Yafeng, HAO Jianfeng, SUN Boyan, et al. Attention-LSTM Based on Nuclear Principal Component Analysis Ultra-Short-Term Forecast of Wind Power[J]. Journal of northeast electric power university, 2025, 45(1).
DOI：

HUANG Yafeng, HAO Jianfeng, SUN Boyan, et al. Attention-LSTM Based on Nuclear Principal Component Analysis Ultra-Short-Term Forecast of Wind Power[J]. Journal of northeast electric power university, 2025, 45(1). DOI： 10.19718/j.issn.1005-2992.2025-01-0095-09.

摘要

针对含气象信息的多维风电功率预测数据集并不能获得更高的预测精度问题，文中提出一种基于核主成分分析结合注意力（Attention）机制的长短期记忆神经网络（Long Short--Term Memory

LSTM）的风电功率超短期预测方法。该方法将风电场数据集所有特征进行核主成分分析降维，并引入Attention机制，采用LSTM将更多的注意力分配给对预测时刻功率影响较大的特征元素，以达到更加准确的预测结果。文中采用真实的风电场历史数据进行预测，实验结果表明：所提出的预测模型能够有效实现风电功率的超短期预测，相比传统的反向传播神经网络（Back Propagation Neural Network

BPNN）和LSTM网络具有更高的预测精度。

Abstract

The accuracy of wind power prediction has a significant impact on the operation of the power system.This paper proposes a wind power ultra short term prediction method based on kernel principal component analysis combined with attention mechanism using short-term and short-term memory neural networks.This method performs kernel principal component analysis to reduce the dimensionality of all features in the wind farm dataset.And the Attention mechanism is introduced

using LSTM to allocate more attention to feature elements that have a significant impact on the power at the prediction time

in order to achieve more accurate prediction results.The experimental results show that the proposed prediction model can effectively achieve ultra short term prediction of wind power

and has higher prediction accuracy compared to traditional BP neural networks and LSTM networks.

关键词

Keywords

references

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

No data

Related Author

No data

Related Institution

No data

AI问答

⁰