基于深度学习的复合材料铺层优化方法

复合材料科学与工程 ›› 2020, Vol. 0 ›› Issue (7): 68-73.

基于深度学习的复合材料铺层优化方法

白国栋¹, 童小燕², 姚磊江^2*

1.西北工业大学航空学院,西安710072;
2.西北工业大学无人机特种技术重点实验室,西安710072

收稿日期:2019-10-18 出版日期:2020-07-28 发布日期:2020-07-28
通讯作者: 姚磊江(1973-),男,博士,教授,主要从事陶瓷基复合材料方面的研究,yaolj@nwpu.edu.cn。
作者简介:白国栋(1995-),男,硕士研究生,主要从事复合材料损伤方面的研究。
基金资助:
国家重点研发计划(2016YFB0700503);国家自然科学基金项目(51772244,11072195)

LAYUP OPTIMIZATION METHOD OF COMPOSITE WING BASED ON DEEP LEARNING

BAI Guo-dong¹, TONG Xiao-yan², YAO Lei-jiang^2*

1. School of Aeronautics, Northwestern Polytechnical University, Xi′an 710072, China;
2. Science and Technology on UVA Laboratory, Northwestern Polytechnical University, Xi′an 710072, Chin

Received:2019-10-18 Online:2020-07-28 Published:2020-07-28

摘要/Abstract

摘要： 复合材料在航空工程中有着广泛的应用,其中铺层优化对提高结构承载能力至关重要。本文中首先运用基于神经网络的深度学习方法快速预测设计结果,然后依据预测结果,通过遗传算法搜索复合材料层合板设计问题的最优解,提高铺层设计效率,最后对比应用深度学习技术与传统优化算法得到的优化结果,证明了深度学习技术在层合板铺层优化中的便捷性和有效性。

关键词: 深度学习, 复合材料, 遗传算法, 多任务学习, 铺层优化

Abstract: Composite wings have been widely used in engineering, the optimization of the ply is essential to improve the carrying capacity of the structure. In this paper, a neural network is trained by samples, and then the neural network mode is used to quickly predict the design result. A genetic algorithm evolution operation is used according to the prediction result to search for the optimal solution of the composite laminate design problem and accelerate the design efficiency of the laminate. Applying deep learning technology to quickly solve the optimal layup order of laminates, and comparing with the optimization results obtained by traditional optimization algorithms, it proves a convenience and effectiveness of deep learning technology in the optimization of laminates. The trained neural network can quickly predict the layering sequence optimization index and provide guidance for the optimization process, which has certain guiding significance for further development of the potential and engineering design of the laminate.

Key words: deep learning, genetic algorithm, multi-task learning, layup optimization, skin design

中图分类号:

TB332

白国栋, 童小燕, 姚磊江. 基于深度学习的复合材料铺层优化方法[J]. 复合材料科学与工程, 2020, 0(7): 68-73.

BAI Guo-dong, TONG Xiao-yan, YAO Lei-jiang. LAYUP OPTIMIZATION METHOD OF COMPOSITE WING BASED ON DEEP LEARNING[J]. Composites Science and Engineering, 2020, 0(7): 68-73.

参考文献

[1]杜善义. 先进复合材料与航空航天[J]. 复合材料学报, 2007, 24(1): 1-12.
[2]薛忠民. 中国玻璃钢/复合材料发展回顾与展望[J]. 玻璃钢/复合材料, 2015(1): 5-12.
[3]邢丽英, 包建文, 礼嵩明, 等. 先进树脂基复合材料发展现状和面临的挑战[J]. 复合材料学报, 2016, 33(7): 1327-1338.
[4]钟文亮. 民机复合材料铺层优化方法[J]. 中国科技信息, 2017(12): 24-26.
[5]Ghiasi H, Fayazbakhsh K, Pasini D, et al. Optimum stacking sequence design of composite materials Part Ⅱ: Variable stiffness design[J]. Composite Structures, 2010, 93(1): 1-13.
[6]郭琰. 基于遗传算法的开孔复合材料层合板铺层优化[J]. 玻璃钢/复合材料, 2018(12): 5-10.
[7]郑国文, 谢习华. 基于改进双种群遗传算法的复合材料层合板铺层优化设计[J]. 玻璃钢/复合材料, 2017(6): 28-32.
[8]Maheri M R. The effect of layup and boundary conditions on the modal damping of FRP composite panels[J]. Journal of Composite Materials, 2011, 45(13): 1411-1422.
[9]Guo S, Cheng W, Cui D. Aeroelastic tailoring of composite wing structures by laminate layup optimization[J]. AIAA Journal, 2006, 44(12): 3146-3150.
[10]Yokozeki T, Aoki T, Ogasawara T, et al. Effects of layup angle and ply thickness on matrix crack interaction in contiguous plies of composite laminates[J]. Composites Part A: Applied Science and Manufacturing, 2005, 36(9): 1229-1235.
[11]冯消冰, 黄海, 王伟, 等. 基于遗传算法的层压板强度优化设计[J]. 玻璃钢/复合材料, 2012(3): 7-13.
[12]王刚, 贾普荣, 黄涛, 等. 含切口复合材料层合板拉伸应变集中与失效分析[J]. 材料工程, 2018, 46(2): 134-141.
[13]李国正, 李丹. 集成学习中特征选择技术[J]. 上海大学学报(自然科学版), 2007, 13(5): 598-604.
[14]谭娟, 王胜春. 基于深度学习的交通拥堵预测模型研究[J]. 计算机应用研究, 2015, 32(10): 2951-2954.
[15]曹俊. 遗传算法及其在复合材料层合板设计中应用的研究[D]. 南京: 南京航空航天大学, 2003.