COMPRESSIVE FAILURE BEHAVIOR OF THREE-DIMENSIONAL WOVEN COMPOSITES

​ABSTRACT

Three-dimensional woven composite materials have a wide range of applications in aerospace fields and become the focus of emerging materials development due to their outstanding mechanical properties in the thickness direction and excellent damage resistance. In terms of further expanding the application prospects of the materials, it is necessary to accurately predict the mechanical properties and failure modes of 3D woven composites, which requires the establishment of high-fidelity solid models. In this paper, a semi-automatic modeling method is proposed by synthesizing the advantages of currently different modeling methods, and the feasibility of the modeling method is verified by the comparative analysis of experimental and simulation results. Since now this model is applied to predict the strength and damage modes of the material under compression load, and explain the relevant mechanism of size effect. The model will be further applied to other aspects of material property prediction in the future.

Keywords: 3D woven textile composite; digital element method; failure mode; Compressive strength.

1 Introduction

Three dimensional woven composites are produced by 3D woven technology combined with resin transfer molding (RTM) technology. Since the 1960s, a variety of fiber-reinforced composites with three dimensional structures have been developed. The common feature of these structures is the addition of reinforced fiber in the direction of thickness.

As a high-performance material by combining different materials, 3D woven composites have advantages such as effective delamination inhibition, strong damage tolerance and excellent impact resistance compared with two-dimensional laminates, which further over- comes the problems in manufacturing and assembly and becomes a novel material with more research value and application prospect. As a reliable load-bearing structural material, it overcomes many design problems that are difficult to be solved by other materials, and is widely used in the aerospace field, for example, H-shaped connectors used to connect the wing panels of aircraft, stiffeners for the aircraft inlet panels, rocket nose cones and ramjet combustion chambers, etc. [4, 5].

In recent years, the mechanical properties of 3D woven composites under different loads have become the focus of research. Great progress has been made in the study of mechanical properties of 3D woven composites [6, 7]. At the same time, due to the uneven distribution inside the material, the overall anisotropy, coupled with the possible cracks and voids caused by the woven forming process, the damage evolution law of this material is very complicated, and the research on the progressive damage mechanism has become a great challenge. With the help of X-ray, microscope and acoustic emission, failure process of three-dimensional woven composites under various loads are studied [8, 9].

The meso-structure of 3D woven composites is quite complex, and the precast is easy to deform in the process of weaving, which makes it difficult to obtain the final meso-geometric shape of the composites. In order to improve the prediction accuracy of the mechanical properties of 3D woven composites, the meso-geometric model is the key point. The traditional methods for modeling meso-structure of three-dimensional woven composites are usually based on certain assumptions that the path and section shape of yarn are ideal, for example, assuming that the path is sinusoidal curve or other periodic function, and the section of yarn is oval or rectangle, etc. [23, 24, 25]. At present, there are many commercial software on the market that can generate ideal geometry, such as WiseTex and TexGen [26, 27]. However, these assumptions to some extent ignore the deformation of yarn and torsion of section, resulting in a decline in the accuracy of the model and a low flexibility of model adjustment, which can- not effectively and accurately set the internal fiber volume content. In recent years, with the development of digital image processing technology, some researchers use μ-CT technology to obtain the internal structure of materials and generate real meso-geometric models through 3D geometric reconstruction technology. This method can effectively improve the calculation accuracy of material properties, but it still has some shortcomings. The accurate segmentation of image depends on the scanning resolution, which will limit the dimensional accuracy of generated geometry and introduce unpredictable and uncontrollable errors [30, 31, 32].

The use of numerical methods to establish meso-geometric models is also worth considering. Wang and Sun et al. proposed a digital element method to realize meso-geometric modeling of fiber composite materials on the virtual fiber scale by simulating the process. The whole weaving process is modeled as a nonlinear solid mechanics problem with boundary displacement conditions. A large number of studies show that this method reduces the assumptions of the model, and the geometric shape of the yarn is obtained from the deformation simulation calculation, so the geometric model is more rea- sonable and the prediction accuracy is greatly improved [33, 34].

In this paper, the aim of the present work is to study the basic mechanical properties and failure modes of 3D woven composites under compression load based on the meso-structural model generated by the digital element method. The layout of the paper is as follow. The basic information and application of 3D woven composites are introduced firstly. Then Section 2 introduces the progress and results of experiments. Section 3 introduces the concrete modeling method and meso-structural model. Section 4 presents the results of simulation, including the performance parameters and failure modes of the structure. Finally, we analyze the results of experiment and simulation in detail.

2 Method

2.1 Test

In this experiment, a set of experimental devices for the study of compression failure behavior of three-dimensional woven composites were designed and built, with the purpose of obtaining the compression modulus and strength in different directions under different thick- nesses, and observing and recording the initiation and development of failure.

There are a total of 30 test pieces, all of which are three-dimensional woven carbon fiber reinforced resin matrix composites. The designed length of test pieces L=150mm (including the clamping section of about 100mm), width b=25mm, and thickness of 2.4mm, 3.4mm and 5.3mm, respectively. In addition, according to the main yarns along the loading direction, the test pieces can be divided into warp yarns and weft yarns. The dimensions of the in-plane compression test piece are shown in Table 1.

Тable 1.

Size and quantity of compression test pieces

The measurement part of compression test is mainly divided into two parts: compression parameter measurement part and shooting recording part. The former is used to record the compression load and deformation degree of the test part in real time during the test process, and the modulus and strength of the test part can be further obtained. The latter is used to record the formation and development of the failure of the test piece. 

Picture 1. Compression test clamping part

According to the position of the strain gauge in the figure, mark the corresponding position on the surface of the test piece. The direction of strain gauge should be parallel to the loading direction of testing machine. Before the test, the experimental procedure was prepared according to the task requirements. The test was carried out according to the displacement control method, and the displacement rate was 2mm/min. Clamp the test part, as shown in Picture 1, the formal test was carried out. The termination condition of the test was that the load instantly dropped to 40% of the maximum tensile load. After loading, the displacement, load and strain data recorded during the test were stored, and the fracture of the test piece was photographed to record the failure mode.

According to the relationship between strain and load in the test process and the initial size of the test part, the compression modulus and compression strength of the test part can be calculated by the following formula:

                                                                             (1)

                                                                                   (2)

F₁ — Recorded force value at position 1;

F₂ — Recorded force value at position 2;

ε₁ — Recorded strain value at position 1;

ε₂ — Recorded strain value at position 2;

A — Section-area of the test pieces;

t — Thickness of the test pieces;

F_max — Maximum force value during the entire experiment.

2.2 Numerical simulation

The finite element analysis is to simulate the real physical system by mathematical ap- proximation, and transform the infinite unknowns that are difficult to calculate into the finite unknowns that are easy to solve. In this study, the finite element software ABAQUS will be used to simulate the machining process and conduct failure analysis.

Three-dimensional woven composite material is mainly composed of two parts, which are the yarn of T800 carbon fiber and the filler matrix of PR520 epoxy resin material. SEM material measurement method is a commonly used method to measure the fiber volume fraction in yarns. Picture 2 shows the distribution of fibers inside the yarn obtained by SEM measurement. The fiber volume fraction represented in the figure is about 78.6%.

Picture 2. Yarn fiber volume measured by SEM

Yarns can be divided into warp and weft yarns according to their orientation. The materials of the two yarns are the same, and the cross-section shape is ellipsoid. The section parameters of warp and weft yarns can be obtained by measuring the test pieces, as shown in Picture 3.

Picture 3. Architecture and cross-section images of a 3D woven composite. (a) warp cross-section; (b) weft cross-section; (c) side view.

After defining the geometric parameters of the yarn, it is necessary to sum up the weaving interspersed ways of the two yarns. Depending on the three kinds of warp interspersed with each other, all the weft yarn can be bound, to achieve the effect of machine weaving.

The way to simulate the forming process of weaving is to apply periodic boundary conditions for yarn insertion and tension forming, and finally the yarn model can be obtained conforming to the spatial distribution law. This method does not need to consider the com- plex structure of yarn fibers, and can be calculated only after the line segment representing the yarn is moved to the appropriate position, greatly simplifying the initial modeling process. In each analysis step, the yarn is moved by setting periodic boundary conditions，as shown in the Picture 4.

Picture 4. Boundary conditions for yarn movement

The process of generating solid yarn is a work with a high repetition rate. Its basic idea is to generate solid yarn in space with the help of node coordinate parameters after woven forming. Considering the large number of yarns and certain regularity, the reverse engineering technology of CATIA is used to conduct parametric modeling of solid yarn.

3 Result

According to the test requirements, the compression test of three-dimensional woven composite materials was completed. The deformation and failure mode of 3D woven composites were captured and recorded by a high-speed camera. Causes of compression failure mainly include layering of test parts and bending and fracture of fiber tubes. Typical failure modes are mainly layering and shear failure, as shown in Fig 3-2. After processing the test data, the stress-strain curve of three-dimensional woven composites under compression load can be obtained. The obtained modulus, strength, and ultimate load of each test component are shown in Table 2.

Тable 2.

Summary of compression test results of woven composites

After comparing the results of several groups of test pieces with different thicknesses, it can be seen that the compression modulus of test pieces with the same thickness is less discrete and the compression strength is relatively stable. When the thickness is within a certain range, its modulus and strength will not be greatly affected by the change of thickness. When the thickness reaches a certain value, its compressive strength will increase with the increase of the thickness of the test part. During the failure process of the test piece, two obvious failure modes can be seen. One is shear failure, which indicates that the failure port is flat and the fracture is along the 45° direction; the other is layered failure, in which obvious compression and bending phenomenon of yarn can be seen at the failure fracture, and obvious separation between yarns can be seen. The test pieces of both directions have similar failure phenomena, such as fracture of warp line, bending failure of weft line, resin delamination and resin cracking.

(a) Layering failure                         (b) Shear failure

Picture 5. Compression failure recorded by a high-speed camera

Picture 6. Compression failure mode of Test piece

The final woven yarn model is shown in Picture 7, which better realizes the effect of woven molding. Relevant parameters of yarn section are similar to actual yarn parameters, achieving the effect of simulated woven molding. The compared results are shown in Table 3. It can be seen that the yarn formed by weaving method can fit the actual yarn size well.

Picture 7. Yarn weaving model

Тable 3.

Verification of yarn weaving method

After correcting the embedding problem between yarns, we generate a 3D solid yarn model by CATIA reverse engineering technology. As shown in the Picture 8.

Picture 8. Solid yarn model generated by CATIA

After obtaining the solid yarn model, we need to define the interaction properties between yarn and resin for simulation calculation. Here, the cohesion model is adopted. The default cohesion model in ABAQUS software is based on the theory of tractor-separation Law. The separation displacement-force relationship can be expressed as a linear triangle, which is a bilinear constitutive model. Picture 9 shows the results of the numerical simulation.

Picture 9. Compression deformation results of 3D woven composites

In addition to the above problems in calculating the mechanical properties of materials, it is also necessary to pay attention to the possible damage modes of yarn and matrix. The first step is to determine whether the yarn and matrix have been damaged, which can be determined by setting the output variables in ABAQUS. In this case, through CSDMG (scalar stiffness degradation for cohesive surface) and CSQUADSCRT (quadratic traction damage initiation criterion for cohesive surface). Where, the total value of damage variables of the CSDMG scale model, and whether the CSQUADSCRT scalar contact points meet the secondary contact stress damage initiation criterion. For the above two variables, when the value is less than 1.0, the criterion of damage occurrence has not been met, and when the value reaches 1.0, the damage has occurred. It can be seen from the output results that local damage has occurred in the matrix and yarn. In contrast, the damage of yarn is more obvious, and the main damage occurs in the contact between yarn and matrix.

4 Discussion

3D woven composites are produced by 3D woven technology combined with resin transfer molding process. Their excellent mechanical properties in thickness direction and damage impedance make them widely used in aerospace and other fields. At present, the research on mechanical properties and failure behavior of 3D woven composites have been relatively mature. On the basis of the experimental study, the finite element simulation method is used to replace the test method to measure the mechanical properties of three-dimensional woven composite materials, which can greatly reduce the cost of the test process, and is morely close to the needs of engineering design, which has more important research significance.

The accuracy of simulation results is closely related to the accuracy of finite element model. At present, the common modeling software still has some shortcomings, which can only consider yarn from the geometric point of view, but cannot pay attention to the actual yarn path and section of non-idealized defects. Based on the DEM method to simulate the yarn units generated in the real weaving process, this research attempts to further transform the virtual fiber units into a solid model that fits the real yarn shape, and then carries out a simple failure analysis on the solid model and compares it with the test results. The specific research content of this paper is as follows:

(1) Based on ASTM standard test, a set of test device for testing the compression properties of three-dimensional woven composites is built. The device can record the load and displacement data during compression test in real time, and capture the initiation and development of the damage of the test parts through high-speed camera.

(2) Solid yarn modeling was realized through ABAQUS software. The yarn model with real spatial distribution was obtained by simulating the actual yarn weaving process with fiber element. Finally, the solid yarn model with real spatial distribution relationship was generated by using the reverse design function of CATIA.

(3) The simulated solid yarn model was applied to failure analysis, and the cohesive contact model was used to define the damage behavior between yarn and matrix, which verified the reliability of the model in predicting the mechanical properties and failure behavior of 3D woven composites.

Reference:

1. Anqi Li, Karim EhabMoustafa Kamel, Badadjida Wintiba. A level set-based procedure for the cohesive modeling of yarn-yarn contacts in woven composite RVEs[J]. Compos- ite Structures 304 (2023) 116356, compstruct, 2022, 116356.
2. Zengfei Liu, Jingran Ge, Kai Liu, Mengran Li. High-fidelity modeling of 3D woven composites considering inhomogeneous intra-yarn fiber volume fractions[J]. Composite Structures, Volume 290, 2022, 115505, ISSN 0263-8223.
3. Gerrit Pierreux, Danny Van Hemelrijck. Automated generation of 3D orthogonal wo- ven composites RVEs including yarn cross-section variations, Composites Science and Technology, Volume 176, 2019, Pages 90-102, ISSN 0266-3538.
4. Wong R. Sandwich construction in the Starship[C]. International SAMPE Symposium and Exhibition, 37 th , 1992: 186-197.
5. Kandebo S W. France, Russia to join in scramjet flight tests[J]. Aviation Week & Space Technology, 2001(March 26): 60-62.
6. Krenkel, W. C/C–SiC composites for space applications and advanced friction systems Mater. Sci. Eng. A 2005, 412, 177–181.
7. Gras R , Leclerc H , Roux S , et al. Identification of the Out-of-Plane Shear Modulus of a 3D Woven Composite[J]. Experimental Mechanics, 2013, 53(5):719-730.
8. S Buchanan, et al. Determination of in-plane shear modulus of 3D woven composites with large repeat unit cells[J]. Plastics Rubber & Composites, 2013:194-198.
9. Bogdanovich A E, Karahan M, Lomov S V, et al. Quasi-static tensile behavior and damage of carbon/epoxy composite reinforced with 3D non-crimp orthogonal woven fabric[J]. Mechanics Of Materials, 2013(62):14-31.
10. Coppola A M, Thakre P R, Sottos N R, et al. Tensile properties and damage evolution in vascular 3D woven glass/epoxy composites[J]. Composites Part A: Applied Science and Manufacturing, 2014(59):9-17.
11. Ploeckl M, Kuhn P, Grosser J, Wolfahrt M. A dynamic test methodology for analyz- ing the strain-rate effect on the longitudinal compressive behavior of fiber-reinforced composites[J]. Composite Structures 2017;180:429–38.
12. Tornabene F, Fantuzzi N, Bacciocchi M, Reddy JN. A posteriori stress and strain re- covery procedure for the static analysis of laminated shells resting on nonlinear elastic foundation[J]. Compos Part B-Eng 2017;126:162–91.
13. LopatinAV, Morozov EV. Buckling of compressed rectangular orthotropic plate resting on elastic foundation with nonlinear change of transverse displacement over the thick- ness[J]. Compos Struct 2021;261:113535.
14. He S, Huang X, Yu P. The Investigation on Static Stability Analysis for Reticulated Shellwith Initial Defect Value Using Stochastic Defect Mode Method[J]. Buildings 2022;12:615.
15. Zhu SP, Liu Q, Lei Q, Wang Q. Probabilistic fatigue life prediction and reliability as- sessment of a high pressure turbine disc considering load variations[J]. Int J Damage Mech 2018;27:156-88.
16. He JC, Zhu SP, Luo C. Size effect in fatigue modelling of defective materials: Applica- tion of the calibrated weakest-link theory[J]. Int J Fatigue 2022;165:107213.
17. M. Hughes, C. Hill, J. Hague, The fracture toughness of bast fibre reinforced polyester composites Part 1 Evaluation and analysis, [J]. Mater. Sci. 37 (21) (2002) 4669–4676.
18. B. Green, M. Wisnom. An experimental investigation into the tensile strength scaling of notched composites[J]. Compos. Part A: Appl. Sci. Manuf. 38 (3) (2007)867–878.
19. Z.P. Bažant, I.M. Daniel, Z. Li, Size effect and fracture characteristics of composite laminates[J]. Eng. Mater. Technol. 118 (3) (1996) 317–324.
20. Tian Qiu, LiHong Liang, YiYi Hao. Thickness effect on the mechanical properties of nacre in Hyriopsis cumingii under three-point bending. Engineering Fracture Mechan- ics[J]. Volume 276, Part A, 2022, 108869, ISSN 0013-7944.
21. Weixin Li, Yao Qiao. Elastic and fracture behavior of three-dimensional ply-to-ply angle interlock woven composites: Through-thickness, size effect, and multiaxial tests,Composites Part C: Open Access,Volume 4, 2021, 100098, ISSN 2666-6820.
22. Cory Hage Mefford, Yao Qiao, Marco Salviato. Failure behavior and scaling of graphene nanocomposites[J]. Composite Structures, Volume 176, 2017, Pages 961-972, ISSN 0263-8223.
23. Lomov S V, Gusakov A V, Huysmans G, Prodromou A, Verpoest I. Textile geometry preprocessor for meso-mechanical models of woven composites [J]. Composites Science and Technology, 2000, 60(11): 2083-2095.
24. Dong W, Xiao J, Li Y. Finite element analysis of the tensile properties of 2.5D braided composites [J]. Materials Science & Engineering A, 2007, 457(1/2):199-204.
25. Isart N, Mayugo J A, Blanco N, Ripoll L. Geometric model for 3D through-thickness orthogonal interlock composites [J]. Composite Structures, 2015, 119: 787-798.
26. Brown L P. Modeling the geometry of textile reinforcements for composites: TexGen [J]. Composite Reinforcements for Optimum Performance (Second Edition), 2021: 237- 265.
27. Stepan V. Lomov. Modeling the geometry of textile composite reinforcements: Wise- Tex[J]. Composite Reinforcements for Optimum Performance (Second Edition), 2021: 199-236.
28. Emerson MJ, Jespersen KM, Dahl AB, Conradsen K. Individual fibre segmentation from 3DX-ray computed tomography for characterising the fibre orientation in unidirectional composite materials[J]. Compos A Appl Sci Manuf 2017;97:83–92.
29. Schell JSU, RenggliM, van Lenthe GH, Müller R, Ermanni P. Micro-computed tomog- raphydetermination of glass fibre reinforced polymer meso-structure[J]. Compos Sci Technol 2006;66(13):2016–22.
30. Huang W, Causse P, Brailovski V, Hu H, Trochu F. Reconstruction of mesostructural material twin models of engineering textiles based on micro-CT aided geometric mod- eling [J]. Composite Part A: Applied Science and Manufacturing, 2019, 124: 105481.
31. Wintiba B, Vasiukov D, Panier S, Lomov S V, Kamel K E M, Massart T J. Automated reconstruction and conformal discretization of 3D woven composite CT scans with local fiber volume fraction control [J]. Composite structures, 2020, 248: 112438.
32. Ge L, Li H, Zhong J. Micro-CT based trans-scale damage analysis of 3D braided com- posites with pore defects [J]. Composites Science and Technology, 2021, 211(28): 108830.
33. Döbrich O, Gereke T, Cherif C. Modeling of textile composite reinforcements on the micro-scale [J] . Autex Research Journal 2014;14:28-33.
34. Isart N, Said B, Ivanov b DS, Hallett SR. Internal geometric modelling of 3D woven composites: A comparison between different approaches[J]. Compos Struct 2015;132:1219-1230.
35. Doebrich O, Gereke T, Cherif C. Modeling the mechanical properties of textile- reinforced composites with a near micro-scale approach[J]. Compos Struct 2016;135(JAN.):1-7.
36. Pickett A, Sirtautas J, Erber A. Braiding Simulation and Prediction of Mechanical Prop- erties[J]. Appl Compos Mater 2009;16:345-364.
37. Wang Y, Sun X. Digital-element simulation of textile processes [J]. Composites Science and Technology, 2001, 61: 311-319.
38. Sun X, Sun C. Mechanical properties of three-dimensional braided composites [J], Com- posite structures, 2004, 65 (3/4):485-492.