基于数据驱动的风电场发电功率迁移预测方法研究

闫润珍; 苏蕊; 延亮

doi:doi:10.19912/j.0254-0096.tynxb.2024-1583

English Version Journal Login

您当前的位置：

首页 >

文章列表页 >

基于数据驱动的风电场发电功率迁移预测方法研究

更新时间：2026-02-06

- 基于数据驱动的风电场发电功率迁移预测方法研究
- Vol. 47, Issue 1, Pages: 567-574(2026)
- 作者机构：
- 作者简介：
- 基金信息：
- DOI：doi:10.19912/j.0254-0096.tynxb.2024-1583
  CLC： TM614
- Online First：05 February 2026，
  
  Published：05 February 2026
- 稿件说明：
移动端阅览
闫润珍, 苏蕊, 延亮. 基于数据驱动的风电场发电功率迁移预测方法研究[J]. 2026, 47(1): 567-574.
DOI：

闫润珍, 苏蕊, 延亮. 基于数据驱动的风电场发电功率迁移预测方法研究[J]. 2026, 47(1): 567-574. DOI： doi:10.19912/j.0254-0096.tynxb.2024-1583.

摘要

针对现阶段中国部分风电场历史运行数据较为稀缺的情况

基于数据驱动方式提出一种卷积神经网络（CNN）-门控循环单元（GRU）的风电场发电功率迁移预测模型。首先

基于CNN与GRU模型优势

构建CNN-GRU组合模型

以消除过拟合问题并减少训练周期。其次

利用K-均值特征聚类算法对风电场历史运行数据进行聚类

以少数典型场景反映大规模场景中特征

减少计算复杂度的同时提升训练精度。再次

进一步明确各迁移预测场景中的迁移条件

既能避免迁移过程中过拟合问题又可为源域与目标域间参数迁移提供决策依据。最后

对比不同迁移预测模型的性能

选择最佳迁移预测方式。一系列训练结果表明：经过修正后的CNN-GRU模型迁移预测结果精确度明显高于传统LSTM模型以及未修正的CNN-GRU模型预测结果;通过K-均值特征聚类算法可对参考风电场进行优化识别

并进一步提升CNN-GRU组合模型迁移预测结果精度。

Abstract

In response to the current scarcity of historical operating data for some wind farms in China

this paper proposes a data-driven wind farm power generation migration prediction model based on a convolutional neural network gated-recurrent unit（CNN-GRU） model. Firstly

based on the advantages of CNN and GRU models

a CNN-GRU combined model is constructed to eliminate overfitting problems and reduce training cycles. Secondly

the -means clustering algorithm is used to cluster the historical operation data of wind farms

reflecting the features of large-scale scenes with a few typical scenes

reducing computational complexity while improving training accuracy. Thirdly

the migration conditions in each migration prediction scenario are further clarified

which not only avoids overfitting problems during the migration process

but also provides decision-making for parameter migration between the source domain and the target domain. Finally

compare the performance of the different migration prediction models is and compared the best migration prediction method is selected. Based on the training results

it can be concluded that the transfer prediction accuracy of the modified CNN-GRU model is significantly higher than that of the traditional LSTM model and the uncorrected CNN-GRU model. The -means clustering algorithm can be used to optimize and identify reference wind farms

and to further improve the accuracy of transfer prediction results of the CNN-GRU combined model.

关键词

Keywords

references

孙义鸣,谭剑锋,周天熠. 叶片旋转方向对NREL Phase Ⅵ风力机功率特性的影响分析[J]. 南京工业大学学报(自然科学版), 2019, 41(6): 695-702.

唐清苇, 向月, 代佳琨, 等. 基于CNN-LSTM的风电场发电功率迁移预测方法[J]. 工程科学与技术, 2024, 56(2): 91-99.

张姗, 冬雷, 纪德洋, 等. 基于NWP相似性分析的超短期光伏发电功率预测[J]. 太阳能学报, 2022, 43(4): 142-147.

康逸群, 刘厦, 雷兢. 基于鲁棒稀疏宽度学习系统的短期风电功率预测[J]. 太阳能学报, 2024, 45(5): 32-43.

黄玲玲, 李锁, 符杨, 等. 基于风电机组状态的超短期海上风电功率预测[J]. 太阳能学报, 2022, 43(8): 391-398.

魏鹏飞, 樊小朝, 史瑞静, 等. 基于改进麻雀搜索算法优化支持向量机的短期光伏发电功率预测[J]. 热力发电, 2021, 50(12): 74-79.

WANG J.A Comparative study of common clustering algorithms in ais ship track clustering: a case study in the Bohai Sea[J]. Computer informatization and mechanical system, 2024, 7(2): 71-76.

薛阳, 王琳, 王舒, 等. 一种结合CNN和GRU网络的超短期风电预测模型[J]. 可再生能源, 2019, 37(3): 456-462.

YANG C, YUE P, GONG J Y, et al.Detecting road network errors from trajectory data with partial map matching and bidirectional recurrent neural network model[J]. International journal of geographical information science, 2024, 38(3): 478-502.

叶瑞丽, 郭志忠, 刘瑞叶, 等. 基于小波包分解和改进Elman神经网络的风电场风速和风电功率预测[J]. 电工技术学报, 2017, 32(21): 103-111.

陈金富, 朱乔木, 石东源, 等. 利用时空相关性的多位置多步风速预测模型[J]. 中国电机工程学报, 2019, 39(7): 2093-2106.

HOSSAIN M A, CHAKRABORTTY R K, ELSAWAH S, et al.Very short-term forecasting of wind power generation using hybrid deep learning model[J]. Journal of cleaner production, 2021, 296: 126564.

QING X Y, NIU Y G.Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM[J]. Energy, 2018, 148: 461-468.

刘长良, 赵陆阳, 王梓齐, 等. 基于时空注意力-SEQ2SEQ模型的多风电机组多步风速预测算法[J]. 太阳能学报, 2023, 44(8): 420-429.

姜建国, 杨效岩, 毕洪波. 基于VMD-FE-CNN-BiLSTM的短期光伏发电功率预测[J]. 太阳能学报, 2024, 45(7): 462-473.

吴凡曈, 杨俊华, 杨梦丽, 等. 基于卷积门控循环单元的波浪发电系统输出功率预测[J]. 太阳能学报, 2024, 45(8): 682-688.

李铁成, 任江波, 刘清泉, 等. 继电保护压板图像识别与模型聚类匹配[J]. 哈尔滨理工大学学报, 2021, 26(4): 70-77.

梁涛, 陈春宇, 谭建鑫, 等. 基于多方面特征提取和迁移学习的风速预测[J]. 太阳能学报, 2023, 44(4): 132-139.

张童彦, 廖清芬, 唐飞, 等. 基于气象资源插值与迁移学习的广域分布式光伏功率预测方法[J]. 中国电机工程学报, 2023, 43(20): 7929-7940.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

WIND SPEED PREDICTION BASED ON MULTIPLE FEATURE EXTRACTION AND TRANSFER LEARNING

RESEARCH ON BEARING FAULT DIAGNOSIS OF WIND TURBINES BASED ON TRANSFER LEARNING

DESIGN OF POWER GENERATION PREDICTION MODEL FOR WIND FARMS DURING SHORT-TERM HEAVY RAINFALL WEATHER

面向频率快速支撑的风电场自适应分层协调控制

风电场AVC子站电压无功灵敏度自主感知及控制参数自适应跟随方法研究

Related Author

Liang Tao

Chen Chunyu

Tan Jianxin

Jing Yanwei

An Wenjie

Chen Changzheng

Tian Miao

Su Xiaoming

Related Institution

Hebei Construction Energy Investment Co., Ltd.

College of Artificial Intelligence and Data Science, Hebei University of Technology

Ningbo Kunbo Measurement and Control Technology Co., Ltd.

Liaoning Vibration and Noise Control Engineering Research Center

School of Mechanical Engineering, Shenyang University of Technology

AI问答

⁰