基于闭环聚类和多目标优化的风电短期功率预测方法

doi:10.16183/j.cnki.jsjtu.2024.079

上海交通大学学报 ›› 2026, Vol. 60 ›› Issue (2): 246-255.doi: 10.16183/j.cnki.jsjtu.2024.079

• 新型电力系统与综合能源 • 上一篇下一篇

基于闭环聚类和多目标优化的风电短期功率预测方法

郭琦¹^,², 闫军¹, 郝乾鹏¹, 韩东¹, 杨志豪¹, 闫馨月³, 张海鹏⁴, 李然³^,⁴^,⁵()

1 内蒙古电力(集团)有限责任公司,呼和浩特 010020
2 天津大学未来技术学院,天津 300072
3 上海交通大学国家电投智慧能源创新学院,上海 200240
4 上海交通大学电力传输与功率变换控制教育部重点实验室,上海 200240
5 上海非碳基能源转换与利用研究院,上海 200240

收稿日期:2024-03-11 修回日期:2024-05-16 接受日期:2024-06-11 出版日期:2026-02-28 发布日期:2026-03-06
通讯作者: 李然 E-mail:rl272@sjtu.edu.cn.
作者简介:郭琦(1981—),正高级工程师,从事新型电力系统调控体系、电力系统运行、电力市场建设、电力系统数字化转型及新能源消纳等方向研究.
基金资助:
内蒙古自治区科学技术厅揭榜挂帅项目(2022JBGS0044)

Short-Term Wind Power Prediction Method Based on Closed-Loop Clustering and Multi-Objective Optimization

GUO Qi¹^,², YAN Jun¹, HAO Qianpeng¹, HAN Dong¹, YANG Zhihao¹, YAN Xinyue³, ZHANG Haipeng⁴, LI Ran³^,⁴^,⁵()

1 Inner Mongolia Power (Group) Co., Ltd., Hohhot 010020, China
2 School of FutureTechnology, Tianjin University, Tianjin 300072, China
3 College of Smart Energy, Shanghai Jiao Tong University, Shanghai 200240, China
4 Key Laboratory of Control of Power Transmission and Conversion of the Ministry of Education, Shanghai Jiao Tong University, Shanghai 200240, China
5 Shanghai Non-Carbon Energy Conversion and Utilization Institute, Shanghai 200240, China

Received:2024-03-11 Revised:2024-05-16 Accepted:2024-06-11 Online:2026-02-28 Published:2026-03-06
Contact: LI Ran E-mail:rl272@sjtu.edu.cn.

摘要/Abstract

摘要：

在区域风电短期功率组合预测领域,常用深度学习预测模型虽然能够充分学习各单一模型的预测特征,但样本数据规模较小时,深度学习模型容易对训练数据分布产生依赖,从而产生过拟合现象;此外,虽然聚类方法被用来提升区域级预测的精度,但现有方法的聚类目标通常为不同类中风电场的差异最小化,而未考虑与预测目标的一致性.为解决以上问题,提出一种基于闭环聚类和多目标优化的风电短期功率预测方法.首先,通过闭环聚类算法将风电场分为多类;对于每类中的风电场数据,运用Bootstrap方法从原始数据集中随机有放回地抽取 $N$ 组训练子集;然后,利用 $N$ 组子集数据分别训练卷积神经网络模型;最后,采用多目标优化算法对 $N$ 个卷积神经网络的预测结果进行集成.基于中国内蒙古地区的实际风电场数据开展算例验证,所提方法在均方根误差方面相比长短期记忆网络模型降低33.81%,相比基于卷积神经网络的组合预测模型降低24.08%,相比基于K-means聚类方法的预测降低14.05%.

关键词: 新能源预测, 聚类, 多目标优化, 深度学习

Abstract:

In the field of regional short-term combined forecasting of wind power, although deep learning methods can effectively learn the predictive features of each individual model, they tend to rely heavily on the training data distribution, which results in overfitting when the sample data size is small. Additionally, although clustering methods are used to improve the accuracy of regional-level forecasting, existing methods typically aim to minimize the dissimilarity among wind farms, such as geographic dissimilarity, without considering consistency with the forecasting objectives. To address these issues, this paper proposes a short-term wind power prediction method based on closed-loop clustering and multi-objective optimization. Initially, wind farms are divided into multiple clusters by the closed-loop clustering method. For each cluster, the Bootstrap method is utilized to randomly extract N training subsets with replacement from the original dataset. Subsequently, N convolutional neural networks are trained independently using these subsets. Finally, multi-objective optimization is employed to integrate the prediction results from the N convolutional neural networks. Case studies utilizing wind farm data from Inner Mongolia Autonomous Region, China, demonstrate that the proposed method reduces root mean square error by 33.81% compared with the long short-term memory model, by 24.08% compared with the convolutional neural network-based combined forecasting model, and by 14.05% compared to predictions based on the K-means clustering method.

Key words: new energy forecasting, clustering, multi-objective optimization, deep learning

中图分类号:

TM614

郭琦, 闫军, 郝乾鹏, 韩东, 杨志豪, 闫馨月, 张海鹏, 李然. 基于闭环聚类和多目标优化的风电短期功率预测方法[J]. 上海交通大学学报, 2026, 60(2): 246-255.

GUO Qi, YAN Jun, HAO Qianpeng, HAN Dong, YANG Zhihao, YAN Xinyue, ZHANG Haipeng, LI Ran. Short-Term Wind Power Prediction Method Based on Closed-Loop Clustering and Multi-Objective Optimization[J]. Journal of Shanghai Jiao Tong University, 2026, 60(2): 246-255.

图/表 8

图1

图2

图3

图4

表1

图5

表2

表3

参考文献 30

[1]	梁以恒, 杨冬梅, 刘刚, 等. 基于功率预测精度提升和市场交易的平抑新能源出力波动策略[J]. 上海交通大学学报, 2025, 59(2): 221-229.
	LIANG Yiheng, YANG Dongmei, LIU Gang, et al. A strategy for smoothing power fluctuations of new energy based on improved power prediction accuracy and market transaction[J]. Journal of Shanghai Jiao Tong University, 2025, 59(2): 221-229.
[2]	罗逸夫, 胡秦然, 钱涛, 等. 计及多工况对机组寿命损耗影响的机组组合优化模型[J]. 上海交通大学学报, 2025, 59(6): 768-779.
	LUO Yifu, HU Qinran, QIAN Tao, et al. Unit commitment optimization model considering impact of multiple operating conditions on unit life loss[J]. Journal of Shanghai Jiao Tong University, 2025, 59(6): 768-779.
[3]	殷高文, 沈非凡, 黄晟, 等. 面向电力市场多时间尺度场景应用的风储系统调度方法[J]. 上海交通大学学报, 2024, 58(9): 1410-1419.
	YIN Gaowen, SHEN Feifan, HUANG Sheng, et al. Dispatching method of combined wind-storage system for multi-time scale scenarios application in electricity markets[J]. Journal of Shanghai Jiao Tong University, 2024, 58(9): 1410-1419.
[4]	叶林, 路朋, 赵永宁, 等. 含风电电力系统有功功率模型预测控制方法综述[J]. 中国电机工程学报, 2021, 41(18): 6181-6198.
	YE Lin, LU Peng, ZHAO Yongning, et al. Review of model predictive control for power system with large-scale wind power grid-connected[J]. Proceedings of the CSEE, 2021, 41(18): 6181-6198.
[5]	苗长新, 王霞, 李昊, 等. 基于数值天气预报风速误差修正的风电功率日前预测[J]. 电网技术, 2022, 46(9): 3455-3464.
	MIAO Changxin, WANG Xia, LI Hao, et al. Day-ahead prediction of wind power based on NWP wind speed error correction[J]. Power System Technology, 2022, 46(9): 3455-3464.
[6]	YUNUS K, THIRINGER T, CHEN P Y. ARIMA-based frequency-decomposed modeling of wind speed time series[J]. IEEE Transactions on Power Systems, 2016, 31(4): 2546-2556. doi: 10.1109/TPWRS.59 URL
[7]	MA J Y, YANG M, HAN X S, et al. Ultra-short-term wind generation forecast based on multivariate empirical dynamic modeling[C]// 2017 IEEE Industry Applications Society Annual Meeting. Cincinnati, USA: IEEE, 2017: 1-8.
[8]	王于波, 郝玲, 徐飞, 等. 分布式光伏集群发电功率波动模式识别与超短期概率预测[J]. 上海交通大学学报, 2024, 58(9): 1334-1343.
	WANG Yubo, HAO Ling, XU Fei, et al. Pattern recognition and ultra-short-term probabilistic forecasting of power fluctuating in aggregated distributed photovoltaics clusters[J]. Journal of Shanghai Jiao Tong University, 2024, 58(9): 1334-1343.
[9]	李芬, 孙凌, 王亚维, 等. 基于CEEMDAN-GSA-LSTM和SVR的光伏功率短期区间预测[J]. 上海交通大学学报, 2024, 58(6): 806-818.
	LI Fen, SUN Ling, WANG Yawei, et al. Short-term interval forecasting of photovoltaic power based on CEEMDAN-GSA-LSTM and SVR[J]. Journal of Shanghai Jiao Tong University, 2024, 58(6): 806-818.
[10]	LI M L, YANG M, YU Y X, et al. Adaptive weighted combination approach for wind power forecast based on deep deterministic policy gradient method[J]. IEEE Transactions on Power Systems, 2024, 39(2): 3075-3087. doi: 10.1109/TPWRS.2023.3294839 URL
[11]	TATINATI S, WANG Y B, KHONG A W H. Hybrid method based on random convolution nodes for short-term wind speed forecasting[J]. IEEE Transactions on Industrial Informatics, 2022, 18(10): 7019-7029. doi: 10.1109/TII.2020.3043451 URL
[12]	DU P W. Ensemble machine learning-based wind forecasting to combine NWP output with data from weather station[J]. IEEE Transactions on Sustainable Energy, 2019, 10(4): 2133-2141. doi: 10.1109/TSTE.5165391 URL
[13]	BATES J M, GRANGER C W J. The combination of forecasts[J]. Journal of the Operational Research Society, 1969, 20(4): 451-468. doi: 10.1057/jors.1969.103 URL
[14]	王丽婕, 刘田梦, 王勃, 等. 基于奇异值分解与卡尔曼滤波修正多位置NWP的短期风电功率预测[J]. 太阳能学报, 2022, 43(12): 392-398.
	WANG Lijie, LIU Tianmeng, WANG Bo, et al. Short-term wind power prediction based on SVD and Kalman filter correction of multi-position NWP[J]. Acta Energiae Solaris Sinica, 2022, 43(12): 392-398.
[15]	丁婷婷, 杨明, 于一潇, 等. 基于误差修正的短期风电功率集成预测方法[J]. 高电压技术, 2022, 48(2): 488-496.
	DING Tingting, YANG Ming, YU Yixiao, et al. Short-term wind power integration prediction method based on error correction[J]. High Voltage Engineering, 2022, 48(2): 488-496.
[16]	刘大贵, 王维庆, 张慧娥, 等. 马尔科夫修正的组合模型在新疆风电中长期可用电量预测中的应用[J]. 电网技术, 2020, 44(9): 3290-3297.
	LIU Dagui, WANG Weiqing, ZHANG Hui’e, et al. Application of Markov modified combination model mid-long term available quantity of electricity forecasting in Xinjiang wind power[J]. Power System Technology, 2020, 44(9): 3290-3297.
[17]	WANG S, WANG J Z, LU H Y, et al. A novel combined model for wind speed prediction—Combination of linear model, shallow neural networks, and deep learning approaches[J]. Energy, 2021, 234: 121275. doi: 10.1016/j.energy.2021.121275 URL
[18]	CUI W K, WAN C, SONG Y H. Ensemble deep learning-based non-crossing quantile regression for nonparametric probabilistic forecasting of wind power generation[J]. IEEE Transactions on Power Systems, 2023, 38(4): 3163-3178.
[19]	WANG J J, LIU Y F. An attention mechanism based deep nonlinear ensemble paradigm of strengthened feature extraction method for wind power prediction[J]. Journal of Renewable & Sustainable Energy, 2023, 15(6): 063305.
[20]	吴杰, 丁明, 张晶晶. 基于云模型和K-means聚类的风电场储能容量优化配置方法[J]. 电力系统自动化, 2018, 42(24): 67-73.
	WU Jie, DING Ming, ZHANG Jingjing. Capacity configuration method of energy storage system for wind farm based on cloud model and K-means clustering[J]. Automation of Electric Power Systems, 2018, 42(24): 67-73.
[21]	WAN C, CUI W K, SONG Y H. Machine learning-based probabilistic forecasting of wind power generation: A combined bootstrap and cumulant method[J]. IEEE Transactions on Power Systems, 2024, 39(1): 1370-1383. doi: 10.1109/TPWRS.2023.3264821 URL
[22]	邵成成, 任孟极, 徐天元, 等. 基于分类-回归卷积神经网络的新能源电力系统可靠性评估方法[J]. 中国电机工程学报, 2024, 44(23): 9134-9145.
	SHAO Chengcheng, REN Mengji, XU Tianyuan, et al. Renewable power system reliability evaluation method based on classification-regression convolutional neural networks[J]. Proceedings of the CSEE, 2024, 44(23): 9134-9145.
[23]	曾金灿, 何耿生, 李姚旺, 等. 基于卷积神经网络与轻量级梯度提升树组合模型的电力行业短期以电折碳方法[J]. 上海交通大学学报, 2025, 59(6): 746-757.
	ZENG Jincan, HE Gengsheng, LI Yaowang, et al. A short-term carbon emission accounting method using electricity data based on convolutional neural networks and light gradient boosting machine combined model in power industry[J]. Journal of Shanghai Jiao Tong University, 2025, 59(6): 746-757.
[24]	陆文安, 朱清晓, 李兆伟, 等. 基于卷积神经网络的新型电力系统频率特性预测方法[J]. 上海交通大学学报, 2024, 58(10): 1500-1512.
	LU Wenan, ZHU Qingxiao, LI Zhaowei, et al. A prediction method of new power system frequency characteristics based on convolutional neural network[J]. Journal of Shanghai Jiao Tong University, 2024, 58(10): 1500-1512.
[25]	张良, 郑丽冬, 冷祥彪, 等. 基于灰狼算法的风-光-抽水蓄能联合系统多目标优化策略[J]. 上海交通大学学报, 2024, 58(10): 1554-1566.
	ZHANG Liang, ZHENG Lidong, LENG Xiangbiao, et al. Multi-objective optimization strategy for wind-photovoltaic-pumped storage combined system based on gray wolf algorithm[J]. Journal of Shanghai Jiao Tong University, 2024, 58(10): 1554-1566.
[26]	RIBEIRO M H D M, DA SILVA R G, MORENO S R, et al. Efficient bootstrap stacking ensemble learning model applied to wind power generation forecasting[J]. International Journal of Electrical Power & Energy Systems, 2022, 136: 107712. doi: 10.1016/j.ijepes.2021.107712 URL
[27]	WOO S, PARK J, PARK J, et al. Wind field-based short-term turbine response forecasting by stacked dilated convolutional LSTMs[J]. IEEE Transactions on Sustainable Energy, 2020, 11(4): 2294-2304. doi: 10.1109/TSTE.5165391 URL
[28]	XIA M, SHAO H D, MA X D, et al. A stacked GRU-RNN-based approach for predicting renewable energy and electricity load for smart grid operation[J]. IEEE Transactions on Industrial Informatics, 2021, 17(10): 7050-7059. doi: 10.1109/TII.2021.3056867 URL
[29]	王亚伦, 周涛, 陈中, 等. 基于堆叠式降噪自动编码器和深度神经网络的风电调频逐步惯性智能控制[J]. 上海交通大学学报, 2023, 57(11): 1477-1491.
	WANG Yalun, ZHOU Tao, CHEN Zhong, et al. Stepwise inertial intelligent control of wind power for frequency regulation based on stacked denoising autoencoder and deep neural network[J]. Journal of Shanghai Jiao Tong University, 2023, 57(11): 1477-1491.
[30]	LIU H, LI Y, DUAN Z, et al. A review on multi-objective optimization framework in wind energy forecasting techniques and applications[J]. Energy Conversion & Management, 2020, 224: 113324.

预测模型	MAE	RMSE	MAPE/ %	准确率/ %	合格率/ %
预测产品1	785.06	1002.08	17.49	95.89	98.82
预测产品2	1503.47	1736.69	45.15	92.87	97.20
SVR	1257.81	1494.41	38.90	93.87	97.03
MLP	952.01	1147.00	21.75	95.29	98.58
LSTM	759.44	963.75	18.76	96.04	100.00
本文模型	553.34	637.89	14.27	97.34	100.00

模型	MAE	RMSE	MAPE/ %	准确率/ %	合格率/ %
优化法	816.70	980.80	18.64	95.97	99.10
XGBoost法	757.71	911.64	27.27	96.26	100.00
CNN组合法	732.84	840.21	20.52	96.55	100.00
本文模型	553.34	637.89	14.27	97.34	100.00

聚类方法	MAE	RMSE	MAPE/ %	准确率/ %	合格率/ %
单体法	640.96	763.060	23.50	96.87	100.00
总体法	575.55	746.204	19.82	96.93	100.00
K-means	631.86	742.160	16.46	96.95	100.00
闭环聚类	553.34	637.890	14.27	97.34	100.00

基于闭环聚类和多目标优化的风电短期功率预测方法

Short-Term Wind Power Prediction Method Based on Closed-Loop Clustering and Multi-Objective Optimization

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 30

相关文章 15

编辑推荐

Metrics

本文评价

[1]	. 低渗油井节能生产参数多目标优化方法[J]. J Shanghai Jiaotong Univ Sci, 2026, 31(2): 486-498.
[2]	. 基于改进YOLOv5l的交通信号灯识别[J]. J Shanghai Jiaotong Univ Sci, 2026, 31(2): 319-333.
[3]	范宏, 魏心武, 贾庆山, 罗佳怡. 能源-交通融合下电-气-热多能系统协同优化调度方法[J]. 上海交通大学学报, 2026, 60(2): 211-223.
[4]	陈亮汶, 朱宇昕, 沈涛, 俞羿帆, 凌霄, 盛庆红. 基于红外尾迹匹配的舰船目标检测算法[J]. 空天防御, 2026, 9(1): 80-90.
[5]	刘琦, 贺轶斐, 顾铭, 陈梓浩, 李昀豪, 汪涛. 基于密度聚类的毫米波雷达目标点云杂点去除技术[J]. 空天防御, 2026, 9(1): 63-72.
[6]	罗志军, 王健瑞, 殷佳伟. 复杂战场环境下的任务驱动智能目标识别方法综述[J]. 空天防御, 2026, 9(1): 1-11.
[7]	刘佳惠, 王聪, 张宏立, 马萍, 李新凯, 董颖超. 代理模型辅助的复杂地形风电场微观选址多目标优化设计[J]. 上海交通大学学报, 2025, 59(9): 1315-1326.
[8]	沈赋, 杨志文, 徐潇源, 张微, 王哲, 曹旸. 基于统计聚类的综合能源系统基础单元模型[J]. 上海交通大学学报, 2025, 59(9): 1348-1358.
[9]	朱峰, 单超, 吴宁, 蔡奇新, 祝宇楠, 刘云鹏, 左强. 电力市场环境下电力用户电价特征提取和异常识别方法[J]. 上海交通大学学报, 2025, 59(7): 995-1006.
[10]	张理, 王宝, 贾健雄, 宋竹萌, 叶钰童, 余跃, 林嘉庆, 徐潇源. 微电网功率预测与调度端到端协同优化方法[J]. 上海交通大学学报, 2025, 59(6): 720-731.
[11]	荣光, 张业鑫, 唐朝, 陈金宝, 周奕玲, 王建园. 基于仿真数据驱动的无人飞行器故障诊断技术研究[J]. 空天防御, 2025, 8(6): 73-84.
[12]	. 基于三维卷积特征金字塔网络的高光谱卫星图像分类[J]. J Shanghai Jiaotong Univ Sci, 2025, 30(6): 1073-1084.
[13]	. 基于深度学习序列方法的多人姿态估计用来检测人体与关键点位置[J]. J Shanghai Jiaotong Univ Sci, 2025, 30(6): 1103-1113.
[14]	. 基于CEEMDAN 和 GRU的停车位预测[J]. J Shanghai Jiaotong Univ Sci, 2025, 30(5): 962-975.
[15]	夏伊琳, 刘刚, 鄢丛强, 蔡云泽. 基于深度学习的SAR图像舰船尾迹旋转框检测算法研究[J]. 空天防御, 2025, 8(5): 64-74.