基于卷积双向长短期记忆网络的风电机组传动系统疲劳载荷预测

doi:10.11930/j.issn.1004-9649.202409072

中国电力 ›› 2025, Vol. 58 ›› Issue (4): 90-97.DOI: 10.11930/j.issn.1004-9649.202409072

• 风电机组暂态运行控制与试验验证关键技术 • 上一篇下一篇

基于卷积双向长短期记忆网络的风电机组传动系统疲劳载荷预测

王晓东¹(), 李清¹(), 付德义², 刘颖明¹, 王若瑾¹

1. 沈阳工业大学电气工程学院，辽宁沈阳　110870
2. 新能源与储能运行控制国家重点实验室（中国电力科学研究院有限公司），北京　100192

收稿日期:2024-09-18 录用日期:2024-12-17 发布日期:2025-04-23 出版日期:2025-04-28
作者简介:
王晓东（1978），男，博士，教授，从事风电大数据处理与智能功率预测、大型风电机组优化设计与控制、大规模储能系统及其功率控制等研究，E-mail：13889296091@163.com
李清（2000），女，通信作者，硕士研究生，从事风电机组功率载荷协调优化研究，E-mail：2569351753@qq.com
基金资助:
国家电网有限公司科技项目（考虑安全约束的电网故障过程风电机组机电耦合机理及控制方法研究，4000-202355454A-3-2-ZN）。

Fatigue Load Prediction of Wind Turbine Drive Train Based on CNN-BiLSTM

WANG Xiaodong¹(), LI Qing¹(), FU Deyi², LIU Yingming¹, WANG Ruojin¹

1. School of Electrical Engineering, Shenyang University of Technology, Shenyang 110870, China
2. State Key Laboratory of Operation and Control of Renewable Energy & Storage Systems (China Electric Power Research Institute), Beijing 100192, China

Received:2024-09-18 Accepted:2024-12-17 Online:2025-04-23 Published:2025-04-28
Supported by:
This work is supported by Science and Technology Project of SGCC (Research on Electromechanical Coupling Mechanism and Control Method of Wind Turbine During Grid Fault Considering Security Constraints, No.4000-202355454A-3-2-ZN).

摘要/Abstract

摘要：

在役风电机组传动系统的疲劳载荷一般基于关键部位应力测量，通过雨流计数法计算进行量化，该过程耗时长、成本高。针对在役风电机组控制策略和参数优化中传统疲劳载荷量化模型偏差较大的问题，在风电机组状态数据的基础上提出了一种基于卷积双向长短期记忆神经网络（convolutional neural network-bidirectional long short-term memory，CNN-BiLSTM）的传动系统疲劳载荷预测模型。首先，以基于额定风速及以上工况OpenFAST的仿真数据构建疲劳载荷特征数据库，并进行训练和测试。然后，将模型的预测数据与实际数据进行对比，利用相关评价指标对模型的预测性能进行评估，验证了该模型的有效性。最后，通过与长短期记忆和深度神经网络两种模型的预测结果对比，证明了CNN-BiLSTM载荷预测模型能进一步提高风电机组传动系统载荷预测的准确度。

关键词: 疲劳载荷, 风电机组, LSTM, 载荷预测, CNN-BiLSTM

Abstract:

The fatigue loads of operational wind turbine drivetrain systems are typically quantified using the rainflow counting method based on stress measurements at critical components, a process that is time-consuming and costly. This paper addresses the significant deviations observed in traditional fatigue load quantification models employed for control strategies and parameter optimization in operational wind turbines. We propose a fatigue load prediction model for the drivetrain system based on a convolutional neural network-bidirectional long short-term memory (CNN-BiLSTM) architecture, utilizing state data from wind turbines. First, we construct a fatigue load feature database using simulation data from OpenFAST under rated wind speed conditions and above, which is subsequently used for training and testing the model. We then compare the model's predicted data with actual data, employing relevant evaluation metrics to assess the predictive performance of the model, thereby validating its effectiveness. Finally, by comparing the prediction results with those from long short-term memory and deep neural network models, we demonstrate that the CNN-BiLSTM load prediction model significantly enhances the accuracy of load predictions for wind turbine drivetrain systems.

Key words: fatigue load, wind turbine, LSTM, load prediction, CNN-BiLSTM

王晓东, 李清, 付德义, 刘颖明, 王若瑾. 基于卷积双向长短期记忆网络的风电机组传动系统疲劳载荷预测[J]. 中国电力, 2025, 58(4): 90-97.

WANG Xiaodong, LI Qing, FU Deyi, LIU Yingming, WANG Ruojin. Fatigue Load Prediction of Wind Turbine Drive Train Based on CNN-BiLSTM[J]. Electric Power, 2025, 58(4): 90-97.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202409072

https://www.electricpower.com.cn/CN/Y2025/V58/I4/90

图/表 9

图 1 风电机组传动链双质量块模型

Fig.1 Double mass block model of wind turbine drive chain

表 1 相关性分析结果

Table 1 Correlation analysis results

	std(${\omega _{\text{r}}}$)	std($P$)	std($\Delta P$)	std($\beta $)	${\bar \omega _{\text{r}}}$	$\bar \beta $	$\bar P$	${L_{{\text{DE}}}}$
std(${\omega _{\text{r}}}$)	1.0000	0.8816	0.8727	0.8096	–0.4991	0.7657	–0.0911	0.7274
std($P$)	0.8816	1.0000	0.9961	0.7976	–0.1952	–0.3286	0.0487	0.8165
std($\Delta P$)	0.8727	0.9961	1.0000	0.7592	0.5497	0.4404	0.1012	0.8120
std($\beta $)	0.8096	0.7976	0.7592	1.0000	0.6949	0.3449	0.4762	0.5557
${\bar \omega _{\text{r}}}$	–0.4991	–0.1952	0.5497	0.6949	1.0000	0.7918	0.3913	–0.0297
$\bar \beta $	0.7657	–0.3286	0.4404	0.3449	0.7918	1.0000	0.1811	–0.0898
$\bar P$	–0.0911	0.0487	0.1012	0.4762	0.3913	0.1811	1.0000	0.2118
${L_{{\text{DE}}}}$	0.7274	0.8165	0.8120	0.5557	–0.0297	–0.0898	0.2118	1.0000

图 2 基于CNN-BiLSTM的风电机组传动系统疲劳载荷预测模型流程

Fig.2 CNN-BiLSTM based fatigue load prediction model flow chart of wind turbine drive system

图 3 BiLSTM模型结构

Fig.3 BiLSTM model structure

表 2 相关工况设置

Table 2 Setting of relevant working conditions

参数	区间	间隔
平均风速/(m·s^–1)	12~25	0.5
湍流强度/%	10~24	0.2
风剪切系数	0.1~0.3	0.02

图 4 基于CNN-BiLSTM的风电机组传动系统疲劳载荷预测模型性能评价

Fig.4 Performance evaluation of fatigue load prediction model for wind turbine drive system based on CNN-BiLSTM

图 5 基于CNN-BiLSTM的风电机组传动系统疲劳载荷的预测结果与实际值的相关性

Fig.5 Correlation between fatigue load prediction results and actual values of wind turbine drive system based on CNN-BiLSTM

图 6 3种模型的风电机组传动系统疲劳载荷的预测结果对比

Fig.6 Comparison of fatigue load prediction results of three models of wind turbine drive system

表 3 3种风电机组传动系统疲劳载荷模型的评价指标对比

Table 3 Comparison of evaluation indexes of three fatigue load models of wind turbine drive systems

模型	R²	E_MA/(kN·m)	E_RMS/(kN·m)	E_MAP
CNN-BiLSTM	0.9801	36.7533	95.9051	0.0237
LSTM	0.9634	52.1705	151.0571	0.0339
DNN	0.9776	46.3469	110.0505	0.0302

参考文献 22

1	薛磊. 大型风电机组载荷监测与延寿评估技术研究[D]. 太原: 山西大学, 2020.
	XUE Lei. Research on load monitoring and life time extension assessment technology of large wind turbine[D]. Taiyuan: Shanxi University, 2020.
2	宾世杨, 张振, 唐俊杰, 等. 基于机器学习的风电机组机械传动系统故障诊断研究[J]. 机械与电子, 2024, 42 (1): 11- 15.
	BIN Shiyang, ZHANG Zhen, TANG Junjie, et al. Research on fault diagnosis of wind turbine mechanical transmission system based on machine learning[J]. Machinery & Electronics, 2024, 42 (1): 11- 15.
3	张磊, 王满康, 随权, 等. 考虑场间-场内风电机组疲劳载荷均衡的海上风氢系统灵活调度[J]. 电力系统保护与控制, 2024, 52 (21): 151- 160.
	ZHANG Lei, WANG Mankang, SUI Quan, et al. Flexible scheduling of offshore wind-hydrogen systems considering inter-and intra-field turbine fatigue load balancing[J]. Power System Protection and Control, 2024, 52 (21): 151- 160.
4	姚兴佳, 谢洪放, 朱江生, 等. 基于LMI的5MW海上风力发电机组载荷控制技术研究[J]. 可再生能源, 2016, 34 (1): 44- 48.
	YAO Xingjia, XIE Hongfang, ZHU Jiangsheng, et al. LMI-based load control for 5 MW offshore wind turbine[J]. Renewable Energy Resources, 2016, 34 (1): 44- 48.
5	LI Z P, YANG S Y, ZHANG F, et al. Analysis of load characteristics of wind turbine blade root bolts under loosened and fractured conditions[J]. Journal of Mechanical Science and Technology, 2024, 38 (4): 1731- 1741. DOI
6	ZHAO H R, WU Q W, HUANG S J, et al. Fatigue load sensitivity-based optimal active power dispatch for wind farms[J]. IEEE Transactions on Sustainable Energy, 2017, 8 (3): 1247- 1259. DOI
7	KNUDSEN T, BAK T, SVENSTRUP M. Survey of wind farm control: power and fatigue optimization[J]. Wind Energy, 2015, 18 (8): 1333- 1351. DOI
8	XIAOPENG Z, DING J, QIANG Z. The Application of Rain-Flow Counting Method in the Analysis of Load Spectrum[J]. Science & Technology Review, 2009, 27 (3): 67- 73.
9	WEBER F, WU H R, STARKE P. A new short-time procedure for fatigue life evaluation based on the linear damage accumulation by Palmgren–Miner[J]. International Journal of Fatigue, 2023, 172, 107653. DOI
10	MONTAVON G, SAMEK W, MÜLLER K R. Methods for interpreting and understanding deep neural networks[J]. Digital Signal Processing, 2018, 73, 1- 15. DOI
11	王伟, 王海云. 基于Adam优化BP神经网络的风机等效疲劳载荷预测[J]. 现代电子技术, 2023, 46 (17): 102- 106.
	WANG Wei, WANG Haiyun. Wind turbine equivalent fatigue load prediction based on Adam optimized BP neural network[J]. Modern Electronics Technique, 2023, 46 (17): 102- 106.
12	巫发明, 李晔, 杨从新, 等. 风电机组传动链动态载荷识别算法及验证[J]. 振动测试与诊断, 2022, 42(4): 664-670, 822.
	WU Faming, LI Ye, YANG Congxin, et al. Dynamic load identification algorithm and verification of wind turbine drive train[J]. Journal of Vibration, Measurement & Diagnosis, 2022, 42(4): 664-670, 822.
13	卢晓光, 李凤格. 基于数字孪生的风机实时载荷预估研究[J]. 机械与电子, 2021, 39 (6): 24- 28, 34.
	LU Xiaoguang, LI Fengge. Real time load prediction of wind turbine based on digital twin[J]. Machinery & Electronics, 2021, 39 (6): 24- 28, 34.
14	ZHANG B H, SOLTANI M, HU W H, et al. Optimized power dispatch in wind farms for power maximizing considering fatigue loads[J]. IEEE Transactions on Sustainable Energy, 2018, 9 (2): 862- 871. DOI
15	彭超. 风力发电机组传动链调谐惯量阻尼器研究[J]. 太阳能学报, 2018, 39 (7): 2036- 2043.
	PENG Chao. Research of tuned inertia damper for wind turbine drive train[J]. Acta Energiae Solaris Sinica, 2018, 39 (7): 2036- 2043.
16	刘颖明, 赵哲, 王晓东, 等. 考虑机组疲劳载荷的风电集群有功功率分配方法[J]. 太阳能学报, 2021, 42 (5): 430- 436.
	LIU Yingming, ZHAO Zhe, WANG Xiaodong, et al. Active power distribution method for wind farm cluster based on cluster analysis[J]. Acta Energiae Solaris Sinica, 2021, 42 (5): 430- 436.
17	张斌, 朱春生, 贾鈜崴, 等. 数据驱动的风电机组传动系统疲劳载荷计算与主动抑制研究[J]. 智慧电力, 2023, 51 (1): 24- 30.
	ZHANG Bin, ZHU Chunsheng, JIA Hongwei, et al. Fatigue load calculation and active suppression of wind turbine transmission system based on data-driven method[J]. Smart Power, 2023, 51 (1): 24- 30.
18	Wind turbines - Part 13: Measurement of mechanical loads: IEC 61400-13[S]. 2015.
19	臧海祥, 赵勇凯, 张越, 等. 基于低风速功率修正和损失函数改进的超短期风电功率预测[J]. 电力系统自动化, 2024, 48 (7): 248- 257.
	ZANG Haixiang, ZHAO Yongkai, ZHANG Yue, et al. Ultra-short-term wind power prediction based on power correction under low wind speed and improved loss function[J]. Automation of Electric Power Systems, 2024, 48 (7): 248- 257.
20	朱乔木, 李弘毅, 王子琪, 等. 基于长短期记忆网络的风电场发电功率超短期预测[J]. 电网技术, 2017, 41 (12): 3797- 3802.
	ZHU Qiaomu, LI Hongyi, WANG Ziqi, et al. Short-term wind power forecasting based on LSTM[J]. Power System Technology, 2017, 41 (12): 3797- 3802.
21	CHAI Z K. BiLSTM short-term wind power prediction based on attention mechanism[C]//2023 IEEE 3rd International Conference on Electronic Technology, Communication and Information (ICETCI). Changchun, China. IEEE, 2023: 1341–1346.
22	SHEN Y Z, TANG B P, LI B, et al. Remaining useful life prediction of rolling bearing based on multi-head attention embedded Bi-LSTM network[J]. Measurement, 2022, 202, 111803. DOI

[1]	季湛洋, 胡阳, 孔令行, 宋子秋, 邓丹, 刘吉臻. 考虑多领域耦合特性的风电机组一次调频动态建模与仿真[J]. 中国电力, 2025, 58(4): 56-67.
[2]	李丹, 秦世耀, 李少林, 贺敬. 基于混沌粒子群的双馈风电机组LVRT实测建模及暂态参数辨识[J]. 中国电力, 2024, 57(8): 75-84.
[3]	赵晶晶, 杜明, 刘帅, 李梓博, 马闻鹤. 基于模型预测控制的双馈风电机组调频与转子转速恢复策略[J]. 中国电力, 2023, 56(6): 11-17.
[4]	李博浩, 郭昆丽, 吕家君, 蔡维正, 刘璐豪, 刘凤仪, 郝翊帆. 弱电网下改进LADRC抑制直驱风机次同步振荡研究[J]. 中国电力, 2023, 56(4): 56-67.
[5]	李冰, 白云山, 赵宽, 郭聪彬, 翟永杰. 基于HSCA-YOLOv7的风电机组叶片表面缺陷检测算法[J]. 中国电力, 2023, 56(10): 43-52.
[6]	马海飞, 滕伟, 彭迪康, 柳亦兵, 靳涛. 基于DRS与改进Autogram的风电齿轮箱复合故障特征提取[J]. 中国电力, 2023, 56(10): 71-79.
[7]	王磊, 柳亦兵, 滕伟, 黄心伟, 刘剑韬. 风电机组叶片无损检测技术研究与进展[J]. 中国电力, 2023, 56(10): 80-95.
[8]	吴宇辉, 张扬帆, 高峰, 王玙, 王耀函, 杨伟新, 张鸿. 基于工作模态分析的风电机组叶片裂纹损伤在线监测研究[J]. 中国电力, 2023, 56(10): 106-114.
[9]	王宇航, 王文玲, 周绪红, 王康, 罗崯滔. 海上风电机组格构式浮式基础结构优化及响应分析[J]. 中国电力, 2022, 55(5): 21-31.
[10]	贾睿, 杨国华, 郑豪丰, 张鸿皓, 柳萱, 郁航. 基于自适应权重的CNN-LSTM&GRU组合风电功率预测方法[J]. 中国电力, 2022, 55(5): 47-56,110.
[11]	滕伟, 黄乙珂, 吴仕明, 柳亦兵. 基于XGBoost与LSTM的风力发电机绕组温度预测[J]. 中国电力, 2021, 54(6): 95-103.
[12]	庄家懿, 杨国华, 郑豪丰, 张鸿皓. 基于多模型融合的CNN-LSTM-XGBoost短期电力负荷预测方法[J]. 中国电力, 2021, 54(5): 46-55.
[13]	杨蕾, 王智超, 周鑫, 李胜男, 和鹏, 向川, 张杰, 王德林. 大规模双馈风电机组并网频率稳定控制策略[J]. 中国电力, 2021, 54(5): 186-194.
[14]	唐英杰, 张哲任, 徐政. 基于二极管不控整流单元的远海风电低频交流送出方案[J]. 中国电力, 2020, 53(7): 44-54,168.
[15]	张兴, 孙艳霞, 李丽娜, 高彩云, 马俊华, 高丽萍, 刘可可, 何国庆, 李光辉. 风电机组电磁暂态建模及验证[J]. 中国电力, 2020, 53(7): 106-112.

基于卷积双向长短期记忆网络的风电机组传动系统疲劳载荷预测

Fatigue Load Prediction of Wind Turbine Drive Train Based on CNN-BiLSTM

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 22

相关文章 15

编辑推荐

Metrics

模态框（Modal）标题

基于卷积双向长短期记忆网络的风电机组传动系统疲劳载荷预测

Fatigue Load Prediction of Wind Turbine Drive Train Based on CNN-BiLSTM

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 22

相关文章 15

编辑推荐

Metrics