基于数据驱动的纤维增强复合材料高效多尺度损伤分析方法

马鹏辉; 胡殿印; 刘茜; 刘昱; 王荣桥

doi:10.13224/j.cnki.jasp.20230051

基于数据驱动的纤维增强复合材料高效多尺度损伤分析方法

doi: 10.13224/j.cnki.jasp.20230051

马鹏辉^1,,
胡殿印^{2, 3, 4},
刘茜^{1, 3, 4,*, ,},
刘昱¹,
王荣桥^{1, 3, 4}

1.
北京航空航天大学能源与动力工程学院，北京 100191
2.
北京航空航天大学航空发动机研究院，北京 100191
3.
中小型航空发动机联合研究中心，北京 100191
4.
北京航空航天大学航空发动机结构强度北京市重点实验室，北京 100191

基金项目: 国家自然科学基金（52022007，52275142）

详细信息

作者简介:
马鹏辉（1998-），男，博士生，主要从事纤维增强复合材料损伤分析方面的研究

通讯作者:
刘茜（1994-），女，副教授，博士，研究领域为航空发动机结构强度及可靠性。E-mail：liuxi@buaa.edu.cn

中图分类号: V231.91
计量
- 文章访问数: 695
- HTML浏览量: 354
- PDF量: 75
- 被引次数: 0
出版历程
- 收稿日期: 2023-02-05
- 网络出版日期: 2025-04-11

Data-driven approach for efficient multiscale damage analysis of fiber reinforced composites

MA Penghui^1
,,
HU Dianyin^{2, 3, 4},
LIU Xi^{1, 3, 4,*
, ,},
LIU Yu¹,
WANG Rongqiao^{1, 3, 4}

1.
School of Energy and Power Engineering，Beihang University，Beijing 100191，China
2.
Research Institute of Aero-Engine，Beihang University，Beijing 100191，China
3.
United Research Center of Mid-Small Aero-Engine，Beijing 100191，China
4.
Beijing Key Laboratory of Aero-Engine Structure and Strength，Beihang University，Beijing 100191，China

摘要

摘要:
为实现纤维增强复合材料的损伤分析，建立了一种高效多尺度损伤分析方法。首先，基于通用单胞理论，分别针对层合板和平纹编织复合材料构建了多尺度损伤分析框架，研究了这类复合材料在单轴拉伸载荷下细、微观尺度的损伤过程。结果表明：平纹编织复合材料复杂的编织结构导致其细、微观损伤演化过程较为复杂，同层合板的损伤过程具有显著差异。在此基础上，引入神经网络，提出了一种基于数据驱动的多尺度损伤分析策略，进一步实现了平纹编织复合材料的高效损伤模拟。模拟与试验结果相比，建立的高效多尺度损伤分析方法预测拉伸强度误差小于7%；且与传统多尺度损伤分析方法相比，宏观尺度计算效率提升12.47倍。
- 复合材料 /
- 通用单胞模型 /
- 多尺度仿真 /
- 数据驱动 /
- 神经网络
Abstract:
In order to realize damage analysis of the fiber reinforced composites, an efficient multiscale damage analysis method was developed. Firstly, based on the generalized method of cells, a multiscale damage analysis framework was constructed for laminate and plain weave composites, and the damage processes at the mesoscale and microscale under uniaxial tension were analyzed. The results showed that the complex weave structure of plain weave composites led to a more complex mesoscale and microscale damage evolution process, which was significantly different from the damage process of laminates. Based on this, neural networks were introduced, and a data-driven multiscale damage analysis strategy was proposed to realize the efficient damage simulation of plain weave composites. Compared with the experimental and simulant results, the error of predicting tensile strength by the efficient multiscale damage analysis method was less than 7%; and compared with traditional multiscale damage analysis method, the efficiency of macroscale calculations can be improved by about 12.47 times.
- composites /
- general method of cells model /
- multiscale simulation /
- data-driven /
- neural networks

HTML

图 1 二维通用单胞模型（GMC）

Figure 1. 2D model of generalized method of cells （GMC）

下载: 全尺寸图片幻灯片

图 2 平纹编织复合材料细观通用单胞模型

Figure 2. GMC for plain weave composites at mesoscale

下载: 全尺寸图片幻灯片

图 3 界面应力场分析

Figure 3. Stress fields of interface

下载: 全尺寸图片幻灯片

图 4 层合板拉伸试样

Figure 4. Laminate tension experimental sample

下载: 全尺寸图片幻灯片

图 5 试验与模拟结果对比

Figure 5. Comparison of experimental and calculated results

下载: 全尺寸图片幻灯片

图 6 层合板宏观损伤扩展

Figure 6. Macroscale damage expansion of laminate

下载: 全尺寸图片幻灯片

图 7 层合板细观损伤扩展

Figure 7. Mesoscale damage expansion of laminate

下载: 全尺寸图片幻灯片

图 8 开孔平纹编织复合材料有限元模型^[40]

Figure 8. Finite element model of open hole plain weave composites^[40]

下载: 全尺寸图片幻灯片

图 9 试验与模拟结果对比

Figure 9. Comparison of experimental and calculated results

下载: 全尺寸图片幻灯片

图 10 开孔平纹编织复合材料宏观损伤扩展

Figure 10. Macroscale damage expansion of open hole plain weave composites

下载: 全尺寸图片幻灯片

图 11 开孔平纹编织复合材料细、微观损伤扩展

Figure 11. Mesoscale and microscale damage expansion of open hole plain weave composites

下载: 全尺寸图片幻灯片

图 12 平纹编织复合材料高效多尺度损伤分析策略

Figure 12. Efficient multiscale analysis strategy for plain weave composites

下载: 全尺寸图片幻灯片

图 13 高效多尺度损伤分析流程

Figure 13. Efficient multiscale analysis process

下载: 全尺寸图片幻灯片

图 14 神经网络结构

Figure 14. Neural network architecture

下载: 全尺寸图片幻灯片

图 15 神经元个数及隐藏层层数对神经网络性能的影响

Figure 15. Effect of number of neurons and hidden layers on the neural network training performance

下载: 全尺寸图片幻灯片

图 16 训练过程中的均方误差演化

Figure 16. MSE evolution during the training

下载: 全尺寸图片幻灯片

图 17 训练过程中的决定系数演化

Figure 17. R² evolution during the training

下载: 全尺寸图片幻灯片

图 18 神经网络模型训练过程

Figure 18. Neural network model training process

下载: 全尺寸图片幻灯片

图 19 应力-应变曲线随神经网络迭代次数的变化

Figure 19. Evolution of the stress-strain curve with the number of iterations of neural network model

下载: 全尺寸图片幻灯片

图 20 开孔平纹编织复合材料两种多尺度分析方法模拟结果对比

Figure 20. Comparison of simulation results of two multiscale methods for open hole plain weave composites

下载: 全尺寸图片幻灯片

表 1 基于失效准则各组分材料刚度退化方案

Table 1. Material stiffness degradation scheme based on failure criteria

失效模式	D₁	D₂	D₃	D₄	D₅	D₆
纤维失效	0.99	0.99	0.99	0.99	0.99	0.99
1方向基体失效	0.9			0.9	0.9
2方向基体失效		0.9		0.9		0.9
3方向基体失效			0.9		0.9	0.9
界面失效	0.9	0.9	0.9	0.9	0.9	0.9

下载: 导出CSV

表 2 层合板组分材料参数

Table 2. Mechanical properties of component materials in laminate

组分	E/GPa	v	G/GPa	X_t/MPa	X_c/MPa	S/MPa
纤维	E₁₁=295 E₂₂=17.03	v₁₂=0.25 v₂₃=0.45	G₁₂=54.2 G₂₃=58.7	5100	2700
基体	4.01	0.29	1.55	58.6	170	73
界面	4.01	0.29	1.55	45*	60*	60*
注：*表示界面法向、切向剪切和轴向剪切强度。

下载: 导出CSV

表 3 平纹编织组分材料参数

Table 3. Mechanical properties of component materials in plain weave composites

组分	E/GPa	v	G/GPa	X_t/MPa	X_c/MPa	S/MPa
纤维	E₁₁=210 E₂₂=15	v₁₂=0.2 v₂₃=0.07	G₁₂=27 G₂₃=7	3530	3530
基体	3	0.35	1.11	25	100	100
界面	3	0.35	1.11	25*	40*	40*
注：*表示界面法向、切向剪切和轴向剪切强度。

下载: 导出CSV

表 4 两种多尺度分析方法计算精度及效率对比

Table 4. Comparison of calculation arruracy and efficiency of two multiscale methods

计算方法	误差/ %	调用单胞模型次数	求解时间/h	计算效率提高
传统多尺度	4.76	516100	212
高效多尺度	6.94	45000	17	12.47

下载: 导出CSV

参考文献(44)

[1]	BILISIK K. Three-dimensional braiding for composites: a review[J]. Textile Research Journal,2013,83(13): 1414-1436. doi: 10.1177/0040517512450766
[2]	吴邵庆,范刚,李彦斌,等. 复合材料梁弹性参数不确定性量化及试验验证[J]. 东南大学学报(自然科学版),2018,48(6): 1004-1012. WU Shaoqing,FAN Gang,LI Yanbin,et al. Uncertainty quantification on elastic parameters of composite beams and its experimental verification[J]. Journal of Southeast University (Natural Science Edition),2018,48(6): 1004-1012. (in Chinese WU Shaoqing, FAN Gang, LI Yanbin, et al. Uncertainty quantification on elastic parameters of composite beams and its experimental verification[J]. Journal of Southeast University (Natural Science Edition), 2018, 48(6): 1004-1012. (in Chinese)
[3]	秦福溶,姜东,曹芝腑,等. 基于灵敏度分析的复合材料组分参数识别方法[J]. 复合材料学报,2018,35(12): 3350-3359. QIN Furong,JIANG Dong,CAO Zhifu,et al. Parameter identification for components of composites based on sensitivity analysis[J]. Acta Materiae Compositae Sinica,2018,35(12): 3350-3359. (in Chinese QIN Furong, JIANG Dong, CAO Zhifu, et al. Parameter identification for components of composites based on sensitivity analysis[J]. Acta Materiae Compositae Sinica, 2018, 35(12): 3350-3359. (in Chinese)
[4]	王猛. 碳纤维增强复合材料宏-细-微观损伤失效研究[D]. 南京: 东南大学,2020. WANG Meng. Study on macro-micro-damage failure of carbon fiber reinforced composites[D]. Nanjing: Southeast University,2020. (in Chinese WANG Meng. Study on macro-micro-damage failure of carbon fiber reinforced composites[D]. Nanjing: Southeast University, 2020. (in Chinese)
[5]	JENKIN C F. Report on materials of construction used in aircraft and aircraft engines[M]. London: Legare Street Press,2022.
[6]	HILL R. A theory of the yielding and plastic flow of anisotropic metals[J]. Proceedings of the Royal Society of London Series A Mathematical and Physical Sciences,1948,193(1033): 281-297.
[7]	TOYODA M. Strength characteristics of composite materials[J]. Welding international,1991,5(5): 341-345. doi: 10.1080/09507119109446748
[8]	HOFFMAN O. The brittle strength of orthotropic materials[J]. Journal of Composite Materials,1967,1(2): 200-206. doi: 10.1177/002199836700100210
[9]	HASHIN Z,ROTEM A. A fatigue failure criterion for fiber reinforced materials[J]. Journal of Composite Materials,1973,7(4): 448-464. doi: 10.1177/002199837300700404
[10]	HASHIN Z. Failure criteria for unidirectional fiber composites[J]. Journal of Applied Mechanics,1980,47(2): 329-334. doi: 10.1115/1.3153664
[11]	PUCK A,SCHÜRMANN H. Failure analysis of FRP laminates by means of physically based phenomenological models[J]. Composites Science and Technology,2002,62(12/13): 1633-1662.
[12]	边天涯,关志东,刘发齐. 三维碳/碳复合材料板件拉伸强度预报[J]. 复合材料学报,2017,34(2): 430-437. BIAN Tianya,GUAN Zhidong,LIU Faqi. Prediction on the tensile strength of 3D carbon/carbon composite plate[J]. Acta Materiae Compositae Sinica,2017,34(2): 430-437. (in Chinese BIAN Tianya, GUAN Zhidong, LIU Faqi. Prediction on the tensile strength of 3D carbon/carbon composite plate[J]. Acta Materiae Compositae Sinica, 2017, 34(2): 430-437. (in Chinese)
[13]	LI Xing,GUAN Zhidong,LI Zengshan,et al. A new stress-based multi-scale failure criterion of composites and its validation in open hole tension tests[J]. Chinese Journal of Aeronautics,2014,27(6): 1430-1441. doi: 10.1016/j.cja.2014.10.009
[14]	张力,王猛,陈强,等. 考虑微观界面的2D编织SiC/SiC复合材料宏-细-微多尺度渐进损伤失效分析[J]. 工程力学,2022,39(3): 233-248. ZHANG Li,WANG Meng,CHEN Qiang,et al. Macro-meso-micro multiscale analysis for progressive damage failure of 2D braided SiC/SiC composites considering microscale interfaces[J]. Engineering Mechanics,2022,39(3): 233-248. (in Chinese doi: 10.6052/j.issn.1000-4750.2021.01.0050 ZHANG Li, WANG Meng, CHEN Qiang, et al. Macro-meso-micro multiscale analysis for progressive damage failure of 2D braided SiC/SiC composites considering microscale interfaces[J]. Engineering Mechanics, 2022, 39(3): 233-248. (in Chinese) doi: 10.6052/j.issn.1000-4750.2021.01.0050
[15]	LIU Zeliang,BESSA M A,LIU W K. Self-consistent clustering analysis: an efficient multi-scale scheme for inelastic heterogeneous materials[J]. Computer Methods in Applied Mechanics and Engineering,2016,306: 319-341. doi: 10.1016/j.cma.2016.04.004
[16]	王涛,侯玉亮,铁瑛,等. 基于ECPL模型的平纹机织复合材料低速冲击多尺度模拟[J]. 振动与冲击,2020,39(20): 295-304. WANG Tao,HOU Yuliang,TIE Ying,et al. Multi-scale simulation of low-velocity impact on plain woven composites based on an ECPL model[J]. Journal of Vibration and Shock,2020,39(20): 295-304. (in Chinese WANG Tao, HOU Yuliang, TIE Ying, et al. Multi-scale simulation of low-velocity impact on plain woven composites based on an ECPL model[J]. Journal of Vibration and Shock, 2020, 39(20): 295-304. (in Chinese)
[17]	皮晓璠,铁瑛,胡明浩. 基于多尺度分析的平纹机织复合材料贴补结构抗冲击性能研究[J]. 振动与冲击,2022,41(19): 188-197. PI Xiaofan,TIE Ying,HU Minghao. Anti-impact performance of plain woven composite patching structure based on multi-scale analysis[J]. Journal of Vibration and Shock,2022,41(19): 188-197. (in Chinese PI Xiaofan, TIE Ying, HU Minghao. Anti-impact performance of plain woven composite patching structure based on multi-scale analysis[J]. Journal of Vibration and Shock, 2022, 41(19): 188-197. (in Chinese)
[18]	ZHAO Qiaoli,WANG Weihan,LIU Yutong,et al. Multiscale modeling framework to predict the low-velocity impact and compression after impact behaviors of plain woven CFRP composites[J]. Composite Structures,2022,299: 116090. doi: 10.1016/j.compstruct.2022.116090
[19]	UDHAYARAMAN R,MULAY S S. Multi-scale approach based constitutive modelling of plain woven textile composites[J]. Mechanics of Materials,2017,112: 172-192. doi: 10.1016/j.mechmat.2017.06.007
[20]	DANG Haoyuan,LIU Peng,ZHANG Yinxiao,et al. Theoretical prediction for effective properties and progressive failure of textile composites: a generalized multi-scale approach[J]. Acta Mechanica Sinica,2021,37(8): 1222-1244. doi: 10.1007/s10409-021-01098-8
[21]	PINEDA E J,BEDNARCYK B A,RICKS T M,et al. Multiscale failure analysis of a 3D woven composite containing manufacturing induced voids and disbonds[J]. Composites Part A: Applied Science and Manufacturing,2022,156: 106844. doi: 10.1016/j.compositesa.2022.106844
[22]	PINEDA E J,BEDNARCYK B A,RICKS T M,et al. Efficient multiscale recursive micromechanics of composites for engineering applications[J]. International Journal for Multiscale Computational Engineering,2021,19(4): 77-105. doi: 10.1615/IntJMultCompEng.2021039732
[23]	BALOKAS G,CZICHON S,ROLFES R. Neural network assisted multiscale analysis for the elastic properties prediction of 3D braided composites under uncertainty[J]. Composite Structures,2018,183: 550-562. doi: 10.1016/j.compstruct.2017.06.037
[24]	YAN Shibo,ZOU Xi,ILKHANI M,et al. An efficient multiscale surrogate modelling framework for composite materials considering progressive damage based on artificial neural networks[J]. Composites Part B: Engineering,2020,194: 108014. doi: 10.1016/j.compositesb.2020.108014
[25]	LIU Xin,TIAN Su,TAO Fei,et al. A review of artificial neural networks in the constitutive modeling of composite materials[J]. Composites Part B: Engineering,2021,224: 109152. doi: 10.1016/j.compositesb.2021.109152
[26]	ABOUDI J. Micromechanical analysis of composites by the method of cells[J]. Applied Mechanics Reviews,1989,42(7): 193-221. doi: 10.1115/1.3152428
[27]	ABOUDI J,ARNOLD S M,BEDNARCYK B A. Micromechanics of composite materials: a generalized multiscale analysis approach[M]. Boston: Butterworth-Heinemann,2012.
[28]	孙杰,孙志刚,宋迎东,等. 基于高精度通用单胞模型的材料细观结构拓扑优化设计[J]. 航空学报,2009,30(11): 2106-2112. SUN Jie,SUN Zhigang,SONG Yingdong,et al. Microstructure topological optimization based on high-fidelity generalized method of cell[J]. Acta Aeronautica et Astronautica Sinica,2009,30(11): 2106-2112. (in Chinese doi: 10.3321/j.issn:1000-6893.2009.11.015 SUN Jie, SUN Zhigang, SONG Yingdong, et al. Microstructure topological optimization based on high-fidelity generalized method of cell[J]. Acta Aeronautica et Astronautica Sinica, 2009, 30(11): 2106-2112. (in Chinese) doi: 10.3321/j.issn:1000-6893.2009.11.015
[29]	胡殿印,杨尧,郭小军,等. 一种平纹编织复合材料的三维通用单胞模型[J]. 航空动力学报,2019,34(3): 608-615. HU Dianyin,YANG Yao,GUO Xiaojun,et al. A 3D general method of cells model for plain weave composites[J]. Journal of Aerospace Power,2019,34(3): 608-615. (in Chinese HU Dianyin, YANG Yao, GUO Xiaojun, et al. A 3D general method of cells model for plain weave composites[J]. Journal of Aerospace Power, 2019, 34(3): 608-615. (in Chinese)
[30]	TANG Zhanwen,ZHANG Boming. Prediction of biaxial failure envelopes for composite laminates based on Generalized Method of Cells[J]. Composites Part B: Engineering,2012,43(3): 914-925. doi: 10.1016/j.compositesb.2012.01.003
[31]	田志强,李彦斌,张培伟,等. 复合材料层合板多尺度交互渐进损伤分析[J]. 工程力学,2019,36(12): 247-256. TIAN Zhiqiang,LI Yanbin,ZHANG Peiwei,et al. Multi-scale interactive progressive damage analysis of composite laminates[J]. Engineering Mechanics,2019,36(12): 247-256. (in Chinese TIAN Zhiqiang, LI Yanbin, ZHANG Peiwei, et al. Multi-scale interactive progressive damage analysis of composite laminates[J]. Engineering Mechanics, 2019, 36(12): 247-256. (in Chinese)
[32]	HA S K,JIN K K,HUANG Yuanchen. Micro-mechanics of failure (MMF) for continuous fiber reinforced composites[J]. Journal of Composite Materials,2008,42(18): 1873-1895. doi: 10.1177/0021998308093911
[33]	LIU Zhun,GUAN Zhidong,TAN Riming,et al. Analysis of open-hole compressive CFRP laminates at various temperatures based on a multiscale strategy[J]. Applied Composite Materials,2019,26(3): 923-944. doi: 10.1007/s10443-019-9759-8
[34]	WANG Liang,ZHENG Chuanxiang,WEI Shuang,et al. Micromechanics-based progressive failure analysis of carbon fiber/epoxy composite vessel under combined internal pressure and thermomechanical loading[J]. Composites Part B: Engineering,2016,89: 77-84. doi: 10.1016/j.compositesb.2015.11.018
[35]	WANG Liang,WANG Baicun,WEI Shuang,et al. Prediction of long-term fatigue life of CFRP composite hydrogen storage vessel based on micromechanics of failure[J]. Composites Part B: Engineering,2016,97: 274-281. doi: 10.1016/j.compositesb.2016.05.012
[36]	LIAO B B,TAN H C,ZHOU J W,et al. Multi-scale modelling of dynamic progressive failure in composite laminates subjected to low velocity impact[J]. Thin-Walled Structures,2018,131: 695-707. doi: 10.1016/j.tws.2018.07.047
[37]	CHOWDHURY N T,BALASUBRAMANI N K,PEARCE G M,et al. A multiscale modelling procedure for predicting failure in composite textiles using an enhancement approach[J]. Engineering Failure Analysis,2019,102: 148-159. doi: 10.1016/j.engfailanal.2019.04.013
[38]	XU Lei,HUANG Yuanchen,ZHAO Chao,et al. Progressive failure prediction of woven fabric composites using a multi-scale approach[J]. International Journal of Damage Mechanics,2018,27(1): 97-119. doi: 10.1177/1056789516663613
[39]	WANG Meng,ZHANG Peiwei,FEI Qingguo,et al. Modified micro-mechanics based multiscale model for progressive failure prediction of 2D twill woven composites[J]. Chinese Journal of Aeronautics,2020,33(7): 2070-2087. doi: 10.1016/j.cja.2019.10.009
[40]	MAO J Z,SUN X S,RIDHA M,et al. A modeling approach across length scales for progressive failure analysis of woven composites[J]. Applied Composite Materials,2013,20(3): 213-231. doi: 10.1007/s10443-012-9266-7
[41]	陈滨琦. 基于多尺度方法的复合材料层合板结构失效机理研究[D]. 南京: 南京航空航天大学,2016. CHEN Binqi. Study on structural failure mechanism of composite laminates based on multi-scale method[D]. Nanjing: Nanjing University of Aeronautics and Astronautics,2016. (in Chinese CHEN Binqi. Study on structural failure mechanism of composite laminates based on multi-scale method[D]. Nanjing: Nanjing University of Aeronautics and Astronautics, 2016. (in Chinese)
[42]	GOODFELLOW I,BENGIO Y,COURVILLE A,et al. Deep Learning: Adaptive Computation and Machine Learning series[M]. Cambridge: The MIT Press,2016.
[43]	JAWAD J,HAWARI A H,ZAIDI S. Modeling of forward osmosis process using artificial neural networks (ANN) to predict the permeate flux[J]. Desalination,2020,484: 114427. doi: 10.1016/j.desal.2020.114427
[44]	GULIKERS T. An integrated machine learning and finite element analysis framework,applied to composite substructures including damage[D]. Dutch: Delft University of Technology,2018.