该文提出一种能够考虑尾流效应的风力机疲劳载荷代理模型。首先,根据工程经验选取合理的模型输入参数样本。其次,采用TurbSim生成相应的入流风场文件,并利用FAST.Farm软件开展双风力机载荷仿真,得到尾流区和非尾流区的风力机疲劳载荷数据库。最后,采用支持向量回归和反向传播神经网络分别建立尾流区和非尾流区的风力机疲劳载荷代理模型。结果表明,相较于传统风力机疲劳载荷代理模型,该文提出的包含风力机间距输入参数的风力机疲劳载荷代理模型能够考虑风力机间的尾流效应,此建模方式更加合理。此外,综合考虑归一化均方根误差（NRMSE）和回归系数（R2）两项模型评价指标,支持向量回归的建模效果优于反向传播神经网络,且模型在叶片和塔顶部位的预测效果普遍好于塔基。
Abstract
A wind turbine fatigue load surrogate model that can account for wake effects is proposed in this paper. First, reasonable samples of input parameters for the model are selected based on engineering experience. Then, TurbSim is used to generate corresponding inflow wind field files, and the FAST. Farm software is used to carry out simulations of dual wind turbine loads to obtain a database of fatigue loads in wake and non-wake regions. Finally, support vector regression （SVR） and backpropagation neural networks （BPNN） are used to establish wind turbine fatigue load surrogate models for wake and non-wake regions, respectively. The results show that compared with traditional wind turbine fatigue load surrogate models, the proposed model, which includes wind turbine spacing as an input parameter, can consider wake effects between turbines, making the modeling approach more reasonable. Furthermore, considering both R² and NRMSE evaluation metrics, the modeling performance of SVM is superior to that of BPNN, with the models generally predicting loads on blades and the tower-top more accurately than at the tower base.
关键词
风力机 /
尾流效应 /
疲劳载荷 /
代理模型 /
FAST.Farm /
支持向量回归 /
反向传播神经网络
Key words
wind turbines /
wake effects /
fatigue load /
surrogate model /
FAST.farm /
support vector regression /
backpropagation neural networks
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 国家能源局. 2024年可再生能源并网运行情况[EB/OL]. https://www.nea.gov.cn/20250221/e10f363cabe3458aaf78ba4558970054/c.html.
National Energy Administration.Operation of renewable energy grid connection in2024 [EB/OL].https://www.nea.gov.cn/20250221/e10f363cabe3458aaf78ba4558970054/c.html.
[2] KAMAL M, RAHMAN M M.Advances in fatigue life modeling: a review[J]. Renewable and sustainable energy reviews, 2018, 82: 940-949.
[3] Wind turbine generator systems-Part 1: safety requirement: IEC 61400-1 Ed. 2.0 en: 1999 systems-Part 1: safety requirement: IEC 61400-1 Ed. 2.0 en: 1999[S]. International Electrotechnical Commission, 1999.
[4] 刘伟杰, 黄国庆, 彭留留, 等. 考虑风力机疲劳载荷限制的风电场布局优化研究[J]. 太阳能学报, 2025, 46(6):548-555.
LIU W J, HUANG G Q, PENG L L, et al.Research on wind farm layout optimization considering fatigue loads constraint of wind turbines[J]. Acta energiae solaris sinica, 2025, 46(6): 548-555.
[5] DIMITROV N, KELLY M C, VIGNAROLI A, et al.From wind to loads: wind turbine site-specific load estimation with surrogate models trained on high-fidelity load databases[J]. Wind energy science, 2018, 3(2): 767-790.
[6] SCHRÖDER L, KRASIMIROV DIMITROV N, VERELST D R, et al. Wind turbine site-specific load estimation using artificial neural networks calibrated by means of high-fidelity load simulations[J]. Journal of physics: conference series, 2018, 1037(6): 062027.
[7] GASPARIS G, LIO W H, MENG F Z.Surrogate models for wind turbine electrical power and fatigue loads in wind farm[J]. Energies, 2020, 13(23): 6360.
[8] HAGHI R, CRAWFORD C.Data-driven surrogate model for wind turbine damage equivalent load[J]. Wind energy science, 2024, 9(11): 2039-2062.
[9] 牟哲岳, 孙勇, 王瑞良, 等. 基于实测数据和机器学习的风电机组载荷预测模型[J]. 太阳能学报, 2023, 44(10): 414-419.
MOU Z Y, SUN Y, WANG R L, et al.Prediction model for wind turbine loads based on experimental data and machine learning[J]. Acta energiae solaris sinica, 2023, 44(10): 414-419.
[10] 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 W, WANG H Y.Wind turbine equivalent fatigue load prediction based on Adam optimized BP neural network[J]. Modern electronics technique, 2023, 46(17): 102-106.
[11] 黄国庆, 刘伟杰, 王彬滨, 等. 基于深度神经网络的风力机疲劳载荷代理模型研究[J]. 太阳能学报, 2025, 46(4): 398-405.
HUANG G Q, LIU W J, WANG B B, et al.Research on surrogate models for fatigue loads prediction of wind turbines based on deep neural network[J]. Acta energiae solaris sinica, 2025, 46(4): 398-405.
[12] JONKMAN B J,BUHL M L.TurbSim user’s guide[R]. National Renewable Energy Laboratory, Technical Report Golden, Colorado, 2006.
[13] JONKMAN J M,SHALER K.FAST.farm user’s guide and theory manual[M]. Golden, CO, USA: National Renewable Energy Laboratory, 2021.
[14] JONKMAN J, BUTTERFIELD S, MUSIAL W, et al.Definition of a 5-MW reference wind turbine for offshore system development[J]. Contract, 2009(February): 1-75.
[15] HAYMAN G J, BUHL J M.Mlife users guide for version 1.00[J]. National renewable energy laboratory, 2012,74(75): 112.
[16] FRANDSEN S, BARTHELMIE R, PRYOR S, et al.Analytical modelling of wind speed deficit in large offshore wind farms[J]. Wind energy, 2006, 9(1/2): 39-53.
[17] AWAD M, KHANNA R.Support vector regression[M]//Efficient Learning Machines. Berkeley, CA: Apress, 2015: 67-80.
[18] HECHT R.Theory of the backpropagation neural network[M]. Neural networks for perception. Academic press, 1992, 1: 65-93.
基金
国家自然科学基金面上项目（52378480）; 重庆市自然科学基金创新发展联合基金（CSTB2024NSCQ-LZX0010）; 学科创新引智基地项目（B18062）; 中央高校基本科研业务费项目（2022CDJQY-009）
PDF(2166 KB)
