2021
DOI: 10.1016/j.compstruct.2021.113871
|View full text |Cite
|
Sign up to set email alerts
|

Deep-learning based damage sensing of carbon fiber/polypropylene composite via addressable conducting network

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1

Citation Types

0
12
0
1

Year Published

2022
2022
2024
2024

Publication Types

Select...
8

Relationship

1
7

Authors

Journals

citations
Cited by 20 publications
(13 citation statements)
references
References 22 publications
0
12
0
1
Order By: Relevance
“…In recent years, there has been increased use of CFRP using thermoplastic polymers, such as poly(phenylene sulfide) (PPS), 3840 poly(ether-ether-ketone) (PEEK), 4143 polypropylene (PP), 4446 polyethylene (PE), 47,48 polyamide (PA), 49,50 polyethylene terephthalate (PET), 51,52 and thermosets, such as phenolic formaldehyde 5355 and epoxy, 5658 because the CFRP presents numerous advantage, due to its high specific tensile strength, high modulus and outstanding wear resistance. 8,5961…”
Section: Carbon Fiber-reinforced Thermoset and Thermoplastic Compositesmentioning
confidence: 99%
“…In recent years, there has been increased use of CFRP using thermoplastic polymers, such as poly(phenylene sulfide) (PPS), 3840 poly(ether-ether-ketone) (PEEK), 4143 polypropylene (PP), 4446 polyethylene (PE), 47,48 polyamide (PA), 49,50 polyethylene terephthalate (PET), 51,52 and thermosets, such as phenolic formaldehyde 5355 and epoxy, 5658 because the CFRP presents numerous advantage, due to its high specific tensile strength, high modulus and outstanding wear resistance. 8,5961…”
Section: Carbon Fiber-reinforced Thermoset and Thermoplastic Compositesmentioning
confidence: 99%
“…Localized damage sensing can be evaluated by simply showing resistance change between top ( tx )- and bottom ( bx )-electrodes one by one (i.e., t1–b1 , t2–b2 …, tx–bx ) (resistance change in thickness direction). 12 For the self-healing, the use of a thermoplastic matrix has an advantage as the damage in the thermoplastic matrix can be easily melted and solidified for healing through joule heating. In this study, the damage sensing and healing behavior using carbon fiber/nylon composites via ACN was studied.…”
Section: Introductionmentioning
confidence: 99%
“…Figure 2(a) shows the damage sensing process of CFRP following our previous work. 12 The resistance change data with respect to the damage of the composite laminate were gathered and trained by machine learning to identify the location and size of the damage. In order to increase the accuracy of damage sensing through machine-learning, the resistance network model with Kirchhoff’s law was devised and the resistance change data with respect to the damage for training, were generated.…”
Section: Introductionmentioning
confidence: 99%
See 1 more Smart Citation
“…technology of composite components for equipment engineering applications, performance design technology of composite components under dynamic load, data-driven design and simulation technology of composite components, and strength and lifespan simulation software platform of composite components, thereby comprehensively improving the design and engineering application level of composite components in China. Keywords: composite material components; multi-field and multi-scale design; dynamic design; data-driven simulation; evaluation of strength and lifespan 力耦合等方面获得初步应用 [2] 。多场耦合环境下复 合材料构件的变形分析方法、损伤失效机理、服役 安全及耐久性多尺度设计,是重大装备发展过程中 亟待解决的关键科学问题 [3] 。因此,发展多物理场 耦合条件下复合材料 -结构一体化多尺度设计理论 体系,是重要且具有挑战性的科学任务。 (二)复合材料构件动力学设计理论及仿真方法 爆炸、冲击、振动等动态服役环境,对复合材料 构件动力学设计理论与仿真方法提出了严峻挑战 [4] 表面、孔洞缺陷对氧化过程中力学行为的影响 [8] ; 基于热力学框架建立了一种考虑质量扩散、氧化反 应、黏弹性有限变形耦合连续介质模型,研究了 SiC 纤维在超高温氧化过程中表面裂纹的形成机理; 建立了化 -力耦合的本构理论,据此预测热氧化复 合材料的本构响应 [9] 。在此基础上,提出了具有热 力学一致性的化 -力耦合大变形理论,考虑了氧扩 散、化学反应与力学的耦合作用,完成聚合物热氧 化与力学耦合变形行为数值模拟 [10] ;进一步地,将 这种化 -力本构模型推广到纤维增强聚合物复合材 料,对热氧化、力学行为进行了数值模拟 [11] ;提出 了热 -化 -力耦合的黏塑性理论模型,应用于金属的 力 -氧化耦合问题研究 [12] 。针对航空发动机用钛 / 环 氧树脂 / 钛胶合接头,采用热 -化 -力模型研究了热 氧化引起的应力状态与梯度,发现热氧化梯度对接头 内部应力具有重要影响 [13] 。 冲击下的力学和破坏行为 [18] 。在复合材料构件动力 学设计中引入结构层次、长度尺度,可丰富传统构 件的抗冲击、减隔振功能;利用该策略,总结了分 级纤维增强复合材料、分级多孔材料、分级功能表 面和高韧性陶瓷复合材料构件的设计方法,研判了 分级复合材料设计和优化方向的研究挑战 [19] 。将多 级复合材料构件作为变形结构动响应设计平台,设 计各层的约束、自适应、预应力以实现构件形状与 刚度的大范围动态控制,为开发形状和功能可动态 调控的复合材料构件提供了新思路 [20] 。 据拟合了非均质材料的显式等效应变能密度函数, 进而获得材料的等效本构关系 [21] ;使用神经网络来 训练材料的隐式等效本构关系(见图 1) ,在降低训 练所需样本的同时,显著提高了仿真效率 [22] 。基于 距离泛函的数据驱动算法,对本构数据点与满足守 恒定量点的距离求极值,可在无需数据拟合的情况 下求得响应 [23] 。 二是基于聚类分析的数据驱动仿真方法。对复 合材料构件的细观代表性胞元(RUC)进行分块或 粗粒化,降低仿真的求解规模。基于数据驱动的自 洽聚类分析方法(SCA) [24] 是此方向上的开拓性工 作,先在离线阶段计算高保真 RUC 模型中各个单 元的弹性应变响应,后据此进行聚类分析,将高保 真 RUC 模型压缩为基于聚类的缩减 RUC 模型。进 一步,考虑材料弹塑性损伤的数据驱动自洽聚类分 析 SCA 方法 [25,26] [31] 。影响复合材料层合板疲劳性能的因素 包括:纤维和基体类型、纤维方向、纤维体积分 数、平均应力与应力比、加载频率、环境因素(湿 热)等。近年来,结合复合材料结构的数值分析结 果来评估疲劳寿命和强度行为的方法主要有:疲劳 寿命模型,基于剩余刚度、剩余强度的唯象模型, 渐进损伤模型(机理模型) [32] 。例如,基于连续损 伤力学理论,发展了复合材料内部损伤的萌生、扩 展模型以及疲劳损伤模型,适用于多级载荷工况下 的复合材料疲劳失效数值模拟分析 [33] ;将剩余应变 作为损伤参量并表征疲劳寿命,建立基于应变的疲 劳损伤判据并以剩余强度为失效判据,实现了复合 材料疲劳寿命预测 [34] ;通过数值模拟方式建立纤维 复合材料界面开裂、纤维开裂模型,评估了复合材 料疲劳寿命与性能 [35] 。 随着 AI 技术的进步,基于 AI 预测复合材料与 结构损伤、疲劳性能等效率很高。借鉴统计理论、 基于机器学习的分析方法 [36] ,在复合材料与结构的 剩余寿命预测方向获得了成功探索。采用深度神经 网络(如卷积神经网络、自动编码器、循环神经网 络) ,针对原始数据进行深度学习,进而表征复合 材料的强度特性 …”
unclassified