基于自适应分解与频域增强 iTransformer 的短期风电功率预测

刘新宇; 王振; 王思磊; 杨淞焜; 兰宇翔; 王赛赛

doi:10.19783/j.cnki.pspc.260045

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于自适应分解与频域增强 iTransformer 的短期风电功率预测

应用研究 | 更新时间：2026-07-27

- 基于自适应分解与频域增强 iTransformer 的短期风电功率预测
- Short-term wind power forecasting based on adaptive decomposition and frequency-domain enhanced iTransformer
- 电力系统保护与控制 2026年54卷第13期
- 作者机构：
  
  华北水利水电大学电气工程学院,河南,郑州,450045
- 作者简介：
- 基金信息：
- DOI：10.19783/j.cnki.pspc.260045
  中图分类号：
- 纸质出版：2026
- 稿件说明：
移动端阅览
刘新宇, 王振, 王思磊, 等. 基于自适应分解与频域增强 iTransformer 的短期风电功率预测[J]. 电力系统保护与控制, 2026,54(13).

LIU Xinyu, WANG Zhen, WANG Silei, et al. Short-term wind power forecasting based on adaptive decomposition and frequency-domain enhanced iTransformer[J]. Power System Protection and Control, 2026, 54(13).
刘新宇, 王振, 王思磊, 等. 基于自适应分解与频域增强 iTransformer 的短期风电功率预测[J]. 电力系统保护与控制, 2026,54(13). DOI： 10.19783/j.cnki.pspc.260045.

LIU Xinyu, WANG Zhen, WANG Silei, et al. Short-term wind power forecasting based on adaptive decomposition and frequency-domain enhanced iTransformer[J]. Power System Protection and Control, 2026, 54(13). DOI： 10.19783/j.cnki.pspc.260045.

摘要

为提高风电功率预测在复杂波动、多变量和强周期条件下的准确性，提出一种融合自适应信号分解与频域增强iTransformer的预测方法。首先，利用霜冰优化算法(rime ice optimization algorithm

RIME)对变分模态分解(variational mode decomposition

VMD)的模态数与惩罚因子进行自适应寻优，对原始风电功率序列进行分解。随后引入样本熵(sample entropy

SE)对分解得到的模态分量进行重构，以降低输入维度并抑制噪声干扰。在预测阶段，构建基于iTransformer的iFANformer模型，通过以傅里叶分析网络(Fourier analysis networks

FAN)替代前馈结构，引入可学习频率特征以增强对序列周期特征的刻画能力。最后，基于实际风电场运行数据开展对比实验与消融分析。结果表明，所提方法在预测精度上优于对比模型，验证了自适应分解与频域增强策略的有效性。

Abstract

To improve wind power forecasting accuracy under complex fluctuation

multivariable coupling

and strong periodicity conditions

this paper proposes a forecasting method integrating adaptive signal decomposition and a frequency-domain enhanced iTransformer. First

the rime ice optimization algorithm (RIME) is employed to adaptively optimize mode number and penalty factor of variational mode decomposition (VMD)

which is then used to decompose the original wind power series. Subsequently

the decomposed modes are then reconstructed using sample entropy

thereby reducing the input dimension and suppressing noise. In the forecasting stage

an iTransformer-based iFANformer model is developed. By replacing the conventional feed-forward network with Fourier analysis networks (FAN)

learnable frequency features are incorporated to strengthen the model's capability of capturing periodic characteristics in time series data. Finally

comparative experiment and ablation analysis are carried out based on real wind farm data. Results show that the proposed method outperforms baseline models in forecasting accuracy

validating the effectiveness of the adaptive decomposition and frequency-enhancement strategies.

关键词

Keywords

references

ZHAO Siying, SHAO Changzheng, DING Jinfeng, et al. Unreliability tracing of power systems for identifying the most critical risk factors considering mixed uncertainties in wind power output[J]. Protection and Control of Modern Power Systems, 2024, 9(5): 96-111.

ZHANG J, PENG C, YANG Y, et al. Hybrid model for short-term wind power prediction by the SAO and a parallel architecture of LSTM and GRU[J]. Electric Power Systems Research, 2026, 253.

DHAKA P, SREEJETH M, TRIPATHI M. Wind power forecasting using multivariate signal decomposition and stacked GRU ensembles with error correction[J]. Future Generation Computer Systems, 2026, 175.

IBRAHEEMI A Z, JANABI A S. Sustainable energy: advancing wind power forecasting with grey wolf optimization and GRU models[J]. Results in Engineering, 2024, 24.

FANG M, ZHANG F, ZHU D, et al. The influence of variational mode decomposition on LSTM prediction accuracy-a case study with wind turbine power signals[J]. Renewable Energy, 2025, 245.

HU M, ZHENG G, SU Z, et al. Short-term wind power prediction based on improved variational modal decomposition, least absolute shrinkage and selection operator, and BiGRU networks[J]. Energy, 2024, 303.

HANG S, DONG Z, ASGHAR A H, et al. RIME: a physics-based optimization[J]. Neurocomputing, 2023, 532: 183-214.

LIU W, BAI Y, YUE X, et al. A wind speed forecasting model based on rime optimization based VMD and multi-headed self-attention-LSTM[J]. Energy, 2024, 294.

JIN H, ZHANG C, ZHANG H. Hybrid noise suppression of transformer partial discharge signals using SVD and RIME-VMD[J]. Measurement, 2026, 259.

LIU Y, HU T, ZHANG H, et al. iTransformer: inverted transformers are effective for time series forecasting[R]. arXiv preprint arXiv:2310.06625, 2023.

DONG Y, LI G, TAO Y, et al. Fan: Fourier analysis networks[R]. arXiv preprint arXiv:2410.02675, 2024.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于混沌进化优化算法优化 VMD 的小波去噪与支持向量机的漏磁信号早期故障识别

一种混合噪声的局部放电信号去噪方法

基于SAO-VMD-S的双端柔性直流输电故障测距方案

高比例分布式电源接入下基于变分模态分解的励磁涌流辨识

考虑电流保护定值不变的分布式电源最大可接入容量评估方法