Machine learning-based design and optimization method of critical parts of wind turbine blades

doi:10.19936/j.cnki.2096-8000.20260128.017

Abstract

Abstract: An analysis of the structural performance of the spar cap of wind turbine blades was conducted, and a reverse structural optimization method for key components of wind turbine blades based on a machine learning model was developed, combining the use of FOCUS and the Python programming language. Taking a 1.5 MW wind turbine blade design as an example, the finite element analysis (FEA) model of the blade was established. The thickness of the spar cap was selected as the design variable, and the peak strain of the spar cap served as the optimization objective. A machine learning model was developed to reflect the underlying mapping relationship between the spar cap layup parameters, strain, and mass. Based on this machine learning model, a self-learning cyclic optimization method was developed. This method enables the rapid iteration of key parts of the same blade type under different wind fields and load conditions. The optimized spar cap improves performance by about 11.44% while maintaining the same cost. Due to its high portability, this method is expected to become an effective tool for the design and optimization of key parts of wind turbine blades.

Key words: wind turbine blades, spar cap strain, FEM, machine learning, reverse optimization, composites

CLC Number:

TB332

LIU Junbang, LIU Qing, LIN Qiyang, ZHANG Wenhua, HUANG Xuanqing. Machine learning-based design and optimization method of critical parts of wind turbine blades[J]. COMPOSITES SCIENCE AND ENGINEERING, 2026, 0(1): 124-132.

References

[1] REBECCA WILLIAMS, ZHAO F. Global offshore wind report 2024[R]. Belgium: GWEC, 2024.
[2] ELLOUMI W, BAKLOUTI N, ABRAHAM A, et al. The multi-objective hybridization of particle swarm optimization and fuzzy ant colony optimization[J]. Journal of Intelligent & Fuzzy Systems, 2014, 27(1): 515-525.
[3] SUN K, WU H, WANG H, et al. Hybrid ant colony and particle swarm algorithm for solving TSP [J]. Computer Engineering and Applications, 2012, 48(34): 60-63.
[4] WANG M Y, WANG X, GUO D, et al. A level set method for structural topology optimization[J]. Computer Methods in Applied Mechanics and Engineering, 2003, 192(1/2): 227-246.
[5] KUZNETSOV I, KOZHIN V, NOVOKSHENOV A, et al. Optimization approach for the core structure of a wind turbine blade[J]. Structural and Multidisciplinary Optimization, 2024, 67(7): 108.
[6] CHEN J, WANG Q, SHEN W Z, et al. Structural optimization study of composite wind turbine blade[J]. Materials Design, 2013, 46: 247-255.
[7] 沈鑫成, 孙后环, 吴旭龙. 基于粒子群优化算法的风机叶片铺层厚度优化与分析[J]. 轻工学报, 2019, 34(6): 103-108.
[8] GENEID A A, ATIA M R, ELSABBAGH A, et al. Structural design optimization of laminated composites for vertical axis wind turbine blades: a single objective approach[J]. Wind Energy Science Discussions, 2024(2): 5194.
[9] SAMUEL A L. Some studies in machine learning using the game of checkers[J]. IBM Journal of research, 1959, 3(3): 210-229.
[10] CHEN L-C, PAPANDREOU G, KOKKINOS I, et al. Deeplab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFS[C]//IEEE Transactions on Pattern Analysis and Machine Intelligence. IEEE, 2017: 834-848.
[11] DONAHUE J, ANNE HENDRICKS L, GUADARRAMA S, et al. Long-term recurrent convolutional networks for visual recognition and description[C]//Proceedings of The IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 2015: 2625-2634.
[12] LE Q V. Building high-level features using large scale unsupervised learning[C]//2013 IEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2013: 8595-8598.
[13] DENG L, LI J, HUANG J-T, et al. Recent advances in deep learning for speech research at Microsoft[C]//2013 IEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2013: 8604-8608.
[14] KARPATHY A, LI F-F. Deep visual-semantic alignments for generating image descriptions[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 2015: 3128-3137.
[15] JORDI Z J, ERASMO C, RAFAEL C-A. Multi-criteria optimal design of small wind turbine blades based on deep learning methods[J]. Energy, 2024, 293: 130625.
[16] LIU H, ZHANG Z, JIA H, et al. A novel method to predict the stiffness evolution of in-service wind turbine blades based on deep learning models[J]. Composite Structures, 2020, 252: 112702.
[17] TIAN S, WANG H, SHANG L, et al. Structural optimization and influence factors on reliability for composite wind turbine blade[J]. Journal of Failure Analysis and Prevention, 2021, 21(6): 2305-2319.
[18] 陈耀然. 基于人工智能方法的垂直轴风机气动性能预测与优化的若干问题研究[D]. 上海: 上海交通大学, 2022.
[19] OGAILI A A F, HAMZAH M N, ABDULHADYJABER A. Integration of machine learning (ML) and finite element analysis (FEA) for predicting the failure modes of a small horizontal composite blade[J]. International Journal of Renewable Energy Research, 2022, 12(4): 2168-2179.
[20] GULLI A, PAL S. Deep learning with Keras[M]. Birmingham: Packt Publishing Ltd., 2017.
[21] RUMELHART D E, HINTON G E, WILLIAMS R J. Learning representations by back-propagating errors[J]. Nature, 1986, 323(6088): 533-536.
[22] RASMUSSEN C E. Gaussian processes in machine learning[M]//Summer School on Machine Learning. Berlin, Heidelberg: Springer Berlin Heidelberg, 2003: 63-71.
[23] HOERL A E, KENNARD R W. Ridge regression: biased estimation for nonorthogonal problems[J]. Technometrics, 1970, 12(1): 55-67.
[24] DEMIR-KAVUK O, KAMADA M, AKUTSU T, et al. Prediction using step-wise L1, L2 regularization and feature selection for small data sets with large number of features[J]. BMC Bioinformatics, 2011, 12: 412.
[25] PAN Y, WANG Y, ZHOU P, et al. Activation functions selection for BP neural network model of ground surface roughness[J]. Journal of Intelligent Manufacturing, 2020, 31: 1825-1836.
[26] GU G X, CHEN C-T, RICHMOND D J, et al. Bioinspired hierarchical composite design using machine learning: simulation, additive manufacturing, and experiment[J]. Materials Horizons, 2018, 5(5): 939-945.