基于PXGBoost-GAE-GMM-GRU多风力机功率预测

彭驿尧; 关新宇; 李强仁; 李春华; 雷艾虎; 贺得珺

doi:10.19912/j.0254-0096.tynxb.2025-0107

PDF(1562 KB)

太阳能学报 ›› 2026, Vol. 47 ›› Issue (6) : 334-343. DOI: 10.19912/j.0254-0096.tynxb.2025-0107

基于PXGBoost-GAE-GMM-GRU多风力机功率预测

彭驿尧¹, 关新宇², 李强仁², 李春华², 雷艾虎², 贺得珺²

作者信息 +

MULTI-WIND POWER FORECASTING BASED ON PSO-OPTIMIZED XGBoost-GAE-GMM-GRU MODEL

Peng Yirao¹, Guan Xinyu², Li Qiangren², Li Chunhua², Lei Aihu², He Dejun²

Author information +

文章历史 +

摘要

提出一种基于高斯图卷积的循环神经网络预测多风力机功率。首先,利用粒子群算法（PSO）优化的极致梯度提升树（XGBoost）算法构成特征筛选网络选取重要特征;其次,利用图自编码器和余弦相关度融合的方法构建网络图有效挖掘站点间潜在相关性,描述多风力机站点特征的空间相关性;最后采用高斯图卷积神经网络深度提取网络图中的内在关系,由循环神经网络整合其时空相关性实现多风力机功率预测。在真实风电场数据上开展实验,将预测结果与其他模型对比。实验结果表明,模型在MSE、RMSE、MAE和MAPE指标上分别达7.7851 MW、2.7902 MW、2.6851 MW和0.2283,较传统方法降低60%~80%,尤其在风速波动场景下误差降低40%,所提模型能显著提升多风力机功率的预测性能。

Abstract

Wind power forecasting is a crucial topic in the field of wind energy generation. With the increasing penetration of renewable energy in power systems, wind energy, as a rapidly developing renewable resource, necessitates accurate short-term forecasting for the energy industry. In this paper, a Gaussian Mixture Model-gated Recurrent Unit （GMM-GRU） based on Gaussian graph convolution is proposed. Firstly, the XGBoost algorithm optimized by Particle Swarm Optimization （PSO） is utilized to construct a feature selection network for identifying important features. Secondly, using the graph auto-encoder and cosine correlation fusion method, a network graph to effectively capture the potential long-distance correlations among sites and accurately describe the spatial correlation of multi-site features are built. Finally, the Gaussian Mixture Model （GMM） is employed to deeply extract the intrinsic relationships within the network graph, and the Gated Recurrent Unit （GRU） integrates its spatio-temporal correlations to address the power forecasting problem. The experiments on real wind farm data are carried out. The comparisons with other models demonstrate that the proposed model significantly enhances wind power forecasting performance.

导出引用

彭驿尧, 关新宇, 李强仁, 李春华, 雷艾虎, 贺得珺. 基于PXGBoost-GAE-GMM-GRU多风力机功率预测[J]. 太阳能学报. 2026, 47(6): 334-343 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0107

Peng Yirao, Guan Xinyu, Li Qiangren, Li Chunhua, Lei Aihu, He Dejun. MULTI-WIND POWER FORECASTING BASED ON PSO-OPTIMIZED XGBoost-GAE-GMM-GRU MODEL[J]. Acta Energiae Solaris Sinica. 2026, 47(6): 334-343 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0107

中图分类号： TM715

参考文献

[1] 刘旺, 邢晓帆, 王戎. 全球陆地风能发电潜力的历史变化及未来预测[J]. 北京大学学报(自然科学版), 2023, 59(5): 833-842.
LIU W, XING X F, WANG R.Historical trend and future projections of global on-shore wind power generation potential[J]. Acta scientiarum naturalium universitatis pekinensis, 2023, 59(5): 833-842.
[2] 陈子含, 滕伟, 胥学峰, 等. 基于图卷积网络和风速差分拟合的中长期风功率预测[J]. 中国电力, 2023, 56(10): 96-105.
CHEN Z H, TENG W, XU X F, et al.Medium and long term wind power prediction based on graph convolutional network and wind velocity differential fitting[J]. Electric power, 2023, 56(10): 96-105.
[3] 晏吴宇歆, 张海波, 刘童蕙, 等. 基于损失函数改进和补丁时序Transformer网络的风功率超短期多步预测[J]. 太阳能学报, 2025, 46(6): 510-521.
YAN W Y Y, ZHANG H B, LIU T H, et al. Ultra-short-term multi-step prediction of wind power based on loss function improvement and patch timing transformer network[J]. Acta energiae solaris sinica, 2025, 46(6): 510-521.
[4] 曹俊波, 周任军, 邓学华, 等. 考虑优化ARIMA模型差分次数的风功率预测[J]. 电力系统及其自动化学报, 2019, 31(1): 105-111.
CAO J B, ZHOU R J, DENG X H, et al.Wind power forecast considering differential times of optimal ARIMA model[J]. Proceedings of the CSU-EPSA, 2019, 31(1): 105-111.
[5] 刘震, 刘丹, 王彦文, 等. 基于多元回归-ARIMA-卡尔曼组合模型的风电功率超短期预测[J]. 风能, 2018(6): 74-78.
LIU Z, LIU D, WANG Y W, et al.Ultra-short-term wind power forecasting based on multiple regression-ARIMA- Kalman combined model[J]. Wind energy, 2018(6): 74-78.
[6] 符杨, 郑紫宸, 时帅, 等. 考虑气象相似性与数值天气预报修正的海上风功率预测[J]. 电网技术, 2019, 43(4): 1253-1259.
FU Y, ZHENG Z C, SHI S, et al.Offshore wind power forecasting considering meteorological similarity and NWP correction[J]. Power system technology, 2019, 43(4): 1253-1259.
[7] 王帅党, 程鹏. 基于GWO优化BiLSTM模型的地区短期风功率的预测[J]. 黑龙江科学, 2023, 14(16): 28-31, 35.
WANG S D, CHENG P.Prediction of BiLSTM model based on GWO for regional short-term wind power prediction[J]. Heilongjiang science, 2023, 14(16): 28-31, 35.
[8] 李进友, 李媛, 黄露秋, 等. 基于XGBoost-GRNN算法的分段式风功率预测[J]. 计算机集成制造系统, 2025, 31(10): 3831-3845.
LI J Y, LI Y, HUANG L Q, et al.Piecewise wind power forecast based on XGBoost-GRNN algorithm[J]. Computer integrated manufacturing systems, 2025, 31(10): 3831-3845.
[9] 付波, 李昊, 权轶, 等. 基于SHAP重要性排序和时空双流的多风场超短期功率预测[J]. 重庆理工大学学报(自然科学), 2024, 38(5): 249-258.
FU B, LI H, QUAN Y, et al.Ultra-short-term power forecasting for multiple wind fields based on SHAP importance ranking and spatio-temporal dual flow[J]. Journal of Chongqing University of Technology (natural science), 2024, 38(5): 249-258.
[10] 赵泽妮, 云斯宁, 贾凌云, 等. 基于统计模型的短期风能预测方法研究进展[J]. 太阳能学报, 2022, 43(11): 224-234.
ZHAO Z N, YUN S N, JIA L Y, et al.Recent progress in short-term forecasting of wind energy based on statistical models[J]. Acta energiae solaris sinica, 2022, 43(11): 224-234.
[11] AHMED U, MUHAMMAD R, ABBAS S S, et al.Short-term wind power forecasting using integrated boosting approach[J]. Frontiers in energy research, 2024, 12: 1401978.
[12] 梁超, 刘永前, 周家慷, 等. 基于卷积循环神经网络的风电场内多点位风速预测方法[J]. 电网技术, 2021, 45(2): 534-541.
LIANG C, LIU Y Q, ZHOU J K, et al.Wind speed prediction at multi-locations based on combination of recurrent and convolutional neural networks[J]. Power system technology, 2021, 45(2): 534-541.
[13] 赵子琪, 杨斌, 张远广. 基于图自编码器和GRU网络的分层交通流预测模型[J]. 计算机科学, 2024, 51(S1): 680-685.
ZHAO Z Q, YANG B, ZHANG Y G.Hierarchical traffic flow prediction model based on graph autoencoder and GRU network[J]. Computer science, 2024, 51(S1): 680-685.
[14] KIPF T N, WELLING M. Variational graph auto-encoders[EB/OL].2016: arXiv: 1611.07308. https://arxiv.org/abs/1611.07308
[15] JI Shan, 姜巍, 景鑫. 基于余弦相似度和图卷积网络的电力负荷预测方法[J]. 浙江电力, 2025, 44(1): 68-75.
JI S, JIANG W, JING X.A power load forecasting method using cosine similarity and a graph convolutional network[J]. Zhejiang electric power, 2025, 44(1): 68-75.
[16] BISHOP C M.Pattern recognition and machine learning[M]. New York: Springer, 2006.
[17] 汪鸿翔, 柳培忠, 骆炎民, 等. 高斯核函数卷积神经网络跟踪算法[J]. 智能系统学报, 2018, 13(3): 388-394.
WANG H X, LIU P Z, LUO Y M, et al.Convolutional neutral network tracking algorithm accelerated by Gaussian kernel function[J]. CAAI transactions on intelligent systems, 2018, 13(3): 388-394.
[18] 杨丽, 吴雨茜, 王俊丽, 等. 循环神经网络研究综述[J]. 计算机应用, 2018, 38(S2): 1-6, 26.
YANG L, WU Y Q, WANG J L, et al.Research on recurrent neural network[J]. Journal of computer applications, 2018, 38(S2): 1-6, 26.
[19] 文洁. MSE与MAE对机器学习性能优化的作用比较[J]. 信息与电脑, 2018, 30(15): 42-43.
WEN J.Comparison of the effects of MSE and MAE on machine learning performance optimization[J]. Information & computer, 2018, 30(15): 42-43.
[20] 常振成, 游国栋, 肖梓跃, 等. 基于VMD-MBWO-KELM的短期风电功率预测[J]. 太阳能学报, 2025, 46(6): 623-631.
CHANG Z C, YOU G D, XIAO Z Y, et al.Short-term wind power forecasting based on VMD-MBWO-KELM[J]. Acta energiae solaris sinica, 2025, 46(6): 623-631.
[21] 陆继翔, 张琪培, 杨志宏, 等. 基于CNN-LSTM混合神经网络模型的短期负荷预测方法[J]. 电力系统自动化, 2019, 43(8): 131-137.
LU J X, ZHANG Q P, YANG Z H, et al.Short-term load forecasting method based on CNN-LSTM hybrid neural network model[J]. Automation of electric power systems, 2019, 43(8): 131-137.
[22] 刘会家, 滕杰, 冯铃, 等. 基于GAT-GRU的高渗透率分布式新能源接入的配电网无功优化[J]. 现代电力, 2025, 42(3): 531-541.
LIU H J, TENG J, FENG L, et al.GAT-GRU based reactive power optimization for distribution networks with high-penetration distributed renewable energy[J]. Modern electric power, 2025, 42(3): 531-541.
[23] 殷豪, 李奕甸, 谢智锋, 等. 混合图神经网络和门控循环网络的短期光伏功率预测[J]. 太阳能学报, 2024, 45(3): 523-532.
YIN H, LI Y D, XIE Z F, et al.Short-term photovoltaic power prediction method based on mixed graph neural network and gated recurrent unit network[J]. Acta energiae solaris sinica, 2024, 45(3): 523-532.