Abstract: Composite materials now make up over 80% of the materials used in wind-turbine blades because they are lightweight, strong, and corrosion resistant. However, their fatigue behavior is more complex than that of metals, soconventional empirical S-N curves often cannot predict fatigue life accurately, leading to errors in full-scale tests and serious safety risks. To improve prediction accuracy, this work presents an enhanced neural-network method. This method uses a Back-Propagation Neural Network (BPNN) owing to its strong nonlinear fitting ability and overcomes its overfitting and poor generalization by optimizing network weights with an Improved Northern Goshawk Optimization (INGO) algorithm. Based on the original Northern Goshawk Optimization (NGO), INGO adds three enhancements:(i)A global search guided by the best solution. (ii)A subtractive-mean position update. (iii)An adaptive Cauchy mutation. Tests conducted on multiple fatigue datasets of different composite laminates demonstrate the following:(i) INGO-BPNN achieves a 2.226 9% average error on the validation set—4.341 6% lower than standard BPNN. (ii) It offers much better prediction stability, producing similar S-N curves across runs. (iii) It cuts validation-set errors by 49.43% and 31.22% in two case studies compared to traditional S-N formulas. The INGO-optimized BPNN method thus greatly improves both the accuracy and stability of fatigue-life predictions for composite materials, providing a reliable new tool for assessing wind-turbine blade fatigue.
董俊涛, 安宗文, 马强, 白学宗. 改进的北方苍鹰算法优化BP神经网络在风电叶片复合材料疲劳预测中的应用[J]. 材料导报, 2026, 40(7): 25020176-7.
DONG Juntao, AN Zongwen, MA Qiang, BAI Xuezong. Application of BP Neural Network Optimized by Improved Northern Goshawk Algorithm in Fatigue Prediction of Composite Materials for Wind Turbine Blades. Materials Reports, 2026, 40(7): 25020176-7.
1 Cheng X Q, Du X Y. Journal of Beijing University of Aeronautics and Astronautics, 2021, 47(7), 1311 (in Chinese). 程小全, 杜晓渊. 北京航空航天大学学报, 2021, 47(7), 1311. 2 Wu Y, Xu F. Engineering & Test, 2024, 64(3), 41 (in Chinese). 吴阳, 许飞. 工程与试验, 2024, 64(3), 41. 3 Feng W S, Yang C P, Jia F. Materials Reports, 2024, 38(2), 23050190 (in Chinese). 冯炜森, 杨成鹏, 贾斐. 材料导报, 2024, 38(2), 23050190. 4 Zhang J B, Hu Z, Zhang J L, et al. Science Technology and Engineering, 2023, 23(13), 5390 (in Chinese). 张金豹, 胡铮, 张金乐, 等. 科学技术与工程, 2023, 23(13), 5390. 5 Tuo H L, Wu T, Lu Z X, et al. Journal of Northwestern Polytechnical University, 2022, 40(3), 651 (in Chinese). 拓宏亮, 吴涛, 卢智先, 等. 西北工业大学学报, 2022, 40(3), 651. 6 Yu F A. Study on fatigue damage mechanism and life prediction of wind turbine blade under random loading. Ph. D. Thesis, Shenyang University of Technology, China, 2022 (in Chinese). 俞方艾. 随机载荷条件下风力机叶片疲劳损伤机理及寿命预测研究. 博士学位论文, 沈阳工业大学, 2022. 7 Chen J, Liu Y. Fatigue & Fracture of Engineering Materials & Structures, 2022, 45(4), 945. 8 Liu X S, Liang Z H, Fu J Y, et al. Acta Materiae Compositae Sinica, 2021, 38(10), 3323 (in Chinese). 廖兴升, 梁智洪, 傅继阳, 等. 复合材料学报, 2021, 38(10), 3323. 9 Xu R X, Gao J X, Zhu P N, et al. Journal of Shanghai Jiao Tong University, 2024, 58(3), 400 (in Chinese). 徐蓉霞, 高建雄, 朱鹏年, 等. 上海交通大学学报, 2024, 58(3), 400. 10 Zhang Z, Yang Q M. Computer Applications and Software, 2018, 35(3), 151 (in Chinese). 张志, 杨清海. 计算机应用与软件, 2018, 35(3), 151. 11 Liu X, Zhao X, Shangguan W B. Fatigue & Fracture of Engineering Materials & Structures, 2022, 45(6), 1678. 12 Dehghani M, Hubálovský Š, Trojovský P. Ieee Access, 2021, 9, 162059. 13 Han S, Khatir S, Wang C, et al. Tribology International, 2023, 183, 108411. 14 Wu C Y, Zhang J. Intelligent Computer and Applications, DOI:10.20169/j.issn.2095-2163.24101005. 吴昌寅, 张军. 智能计算机与应用, DOI:10.20169/j.issn.2095-2163.24101005. 15 Trojovský P, Dehghani M. Biomimetics, 2023, 8(2), 149.
渝公网安备50019002502923号 版权所有：《材料导报》编辑部
2026, Vol. 40 
