Scholarly article on topic 'Static Pull and Push Bending Properties of RTM-made TWF Composite Tee-joints'

Static Pull and Push Bending Properties of RTM-made TWF Composite Tee-joints Academic research paper on "Materials engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Chinese Journal of Aeronautics
OECD Field of science
Keywords
{fabrics / "mechanical properties" / "finite element analysis" / "resin transfer moulding" / tee-joint}

Abstract of research paper on Materials engineering, author of scientific article — Chuyang LUO, Junjiang XIONG

Abstract This paper deals with static pull and push bending tests on two-dimensional (2D) orthogonal EW220/5284 twill weave fabric (TWF) composite tee-joints processed with the resin transfer moulding (RTM) technique. Static pull and push bending properties are determined and failure initiation mechanism is deduced from experimental observations. The experiments show that the failure initiation load, on average, is greater for push bending than for pull bending, whereas the scatter is smaller for push bending than for pull bending. The failure mode of RTM-made tee-joints in pull bending tests can be reckoned to be characteristic of debonding of resin matrix at the interface between the triangular resin-rich zone and the curved web of tee-joint until complete separation of the curved web from the bottom plate. In contrast, as distinct from the products subject to pull bending loading, the RTM tee-joints in push bending tests experience matrix cracking and fibre fracture from outer layers to inner layers of the bottom plate until catastrophic collapse resulting from the bending. Three-dimensional finite element (FE) models are presented to simulate the load transfer path and failure initiation mechanism of RTM-made TWF composite tee-joint based on the maximum stress criterion. Good correlation between experimental and numerical results is achieved.

Academic research paper on topic "Static Pull and Push Bending Properties of RTM-made TWF Composite Tee-joints"

Chinese Journal of Aeronautics 25 (2012) 198-207

ELSEVIER

Contents lists available at ScienceDirect

Chinese Journal of Aeronautics

journal homepage: www.elsevier.com/locate/cja

JOURNAL OF

AERONAUTICS

Static Pull and Push Bending Properties of RTM-made TWF

Composite Tee-joints

LUO Chuyanga'b, XIONG Junjianga'*

aSchool of Transportation Science and Engineering, Beihang University, Beijing 100191, China bChina Airborne Missile Academy, Luoyang 471009, China Received: 29 March 2011; revised: 17 May 2011; accepted: 20 July 2011

Abstract

This paper deals with static pull and push bending tests on two-dimensional (2D) orthogonal EW220/5284 twill weave fabric (TWF) composite tee-joints processed with the resin transfer moulding (RTM) technique. Static pull and push bending properties are determined and failure initiation mechanism is deduced from experimental observations. The experiments show that the failure initiation load, on average, is greater for push bending than for pull bending, whereas the scatter is smaller for push bending than for pull bending. The failure mode of RTM-made tee-joints in pull bending tests can be reckoned to be characteristic of debonding of resin matrix at the interface between the triangular resin-rich zone and the curved web of tee-joint until complete separation of the curved web from the bottom plate. In contrast, as distinct from the products subject to pull bending loading, the RTM tee-joints in push bending tests experience matrix cracking and fibre fracture from outer layers to inner layers of the bottom plate until catastrophic collapse resulting from the bending. Three-dimensional finite element (FE) models are presented to simulate the load transfer path and failure initiation mechanism of RTM-made TWF composite tee-joint based on the maximum stress criterion. Good correlation between experimental and numerical results is achieved.

Keywords: fabrics; mechanical properties; finite element analysis; resin transfer moulding; tee-joint

1. Introduction

Tee-joints are commonly used as stiffening structure in aircraft engineering with typical examples being illustrated in Figs. 1-2. The T-sections are needed to provide the rigidity to the otherwise thin and relatively flexible composite plates. The stiffening can be done in longitudinal or/and transverse directions with typical basic elements being I-section (see Fig. 1), T-section (see Fig. 2), L-section (see Fig. 3) and tc-section (see Fig. 4). It is well known that mechanical connections, widely used in metallic structures, generally reduce the

^Corresponding author. Tel.: +86-10-82316203. E-mail address: jjxiong@buaa.edu.cn

Foundation items: National Natural Science Foundation of China (E050603); Aeronautical Science Foundation of China (20095251024)

1000-9361/$ - see front matter © 2012 Elsevier Ltd. All rights reserved. doi: 10.1016/S1000-9361(11)60379-8

strength of composite structures. Thus integral moulding technology is more and more widely applied to

Fig. 1 T-section (or Y-section, I-section) stiffening skin element in aerofoil structure.

Fig. 2 T-section (or Y-section) stiffening skin element in RTM-made tail.

Fig. 3 L-section element in T-section (or Y-section) stiffening skin structure.

Fig. 4 rc-section stiffening skin element in multi-wall box of aircraft.

curing the skin and stiffeners in modern aeronautical material and structure manufacture for improving connection strength and decreasing structural weight and manufacturing cost.

A significant body of research is available dealing with the mechanical behavior and failure mechanism of the tee-joints under static loadings using experimental, numerical simulation and analytical methods. Falzon, et al. [1] investigated experimentally the post-buckling mechanical behavior of T-section stiffened skin structure and found that the structure with buckling did not fail immediately; rather mid-plane delamination occurred initially at free edge of the stiffener web and then propagated along the interface until inter-laminar shear stress failure. Orifici, et al. [2] investigated the post-buckling failure of composite tee-joint by using a combination of experiment and finite element (FE) simulation. It was shown that flange failure was the dominant mode. The FE model with cohesive element was used to simulate flange failure and good agreement between experiments and simulations was achieved. As mentioned above, final failure of stiffening skin structure usually results from local damage between the skin and stiffener. Therefore, failure strength and mechanism of the T-joint under static loading received a wide attention from a number of researches. Shenoi, et al. [3] investigated failure behavior of tee-joints using experiment and FE method and found that gap size and backfill angle had limited effect on the performance of the joint while fillet radius and web thickness had a significant influence on mechanical behavior of the joint. Stickler, et al. [4] investigated the failure mechanism of stitched resin transfer moulding-made (RTM-made) tee-joints under bending and found that the failure initially occurred at resin-rich fillet and the final failure was caused by fibre pullout and breakage at the web-to-flange interface. Vijayaraju [5] and Greenhalgh [6], et al. carried out the experiments on tensile properties of RTM-made and bonded tee-joints, where it was clear that the initial damage usually occurred at the resin-rich zone to result in the decreasing of rigidity and increasing non-linear-

ity of the stress-strain curve. After this, the stress-strain curve continued to rise in a smooth and approximately linear manner up to the peak-ultimate strength at a lower rate than in the elastic regime. From previous literature, it is clear that the some tee-joint configurations with initial damage could continue to withstand increased loading in an approximately linear manner up to the peak-ultimate strength with the presence of cracks not apparently causing a difference in the numerical simulations.

Dharmawan, et al. [7] predicted failure strength of tee-joints subject to tensile loading based on FE method by introducing crack tip element method. Blake, et al. [8] conducted experimental investigations on static tensile properties of tee-joints with visco-elastic constituents and proposed a progressive damage model by means of FE method to simulate crack formation and propagation of joint subjected to static loading. Falzon [9] and Li [10], et al. determined fracture properties of tee-joint subjected to tensile and compressive loading through experiment and simulated crack propagation in multi-damage modes based on virtual crack closure technique (VCCT) and strain energy release rate (SERR). Li, et al. [11] predicted damage in the triangular resin zone of the tee-joint subjected to tensile loading by means of FE analysis based on the Hart-Smith matrix failure criterion and by considering residual stress effect.

It is clear from the literature review that the mechanical behavior of composite tee-joint has been studied comprehensively. It goes without saying that different production methods would result in distinct mechanical properties in products [12-13]. By using the RTM technique, numerous fasteners are eliminated and bearing races in the fittings are integrally moulded in place. These result in the reduction in the fabrication cost over the previous prepreg panel with the hand lay-up and autoclave curing. In addition, the interface adhesion properties between the laminate and resin-rich zone of RTM-made tee-joint have a significant improvement over those of a tee-joint with other techniques [14-15]. Therefore, in order to provide important information as a basis for technologists to decide what method is the best choice, a study on mechanical behavior of RTM-made tee-joint in stiffened skin wing (see Fig. 1) is urgently needed in aircraft industry [12-13]. However, there seems to be precious few works done on this subject, from the above review. The paper, therefore, aims to investigate the mechanical properties and failure initiation mechanism of 2D orthogonal EW220/5284 twill weave fabric (TWF) tee-joints processed using the RTM technique through pull and push bending tests and numerical simulation, and finally to make a quantitative prediction of failure initiation load on the basis of FE analysis.

2. Static Mechanical Tests of RTM-made Tee-joint

2.1. Material and specimens

The tee-joint specimens for pull and push bending

tests were constructed out of 2D orthogonal EW220/ 5284 TWF composites, prepared with the same geometry and dimensions as shown in Fig. 5(a) and manufactured using the RTM technique. Figure 5(b) shows a detailed geometric configuration in the resin-rich zone and the fabric stacking sequence of the tee-joint. From Fig. 5(b), it can be observed that the tee-joint was made up of both left and right curved webs and one bottom plate, 1.5 mm thick, which were stacked together with nine 0° fabric lay-ups. Three bundles of EW220 glass thread were twisted into a single strand, with the warps and wefts of the fabric amounting to 14 strands/cm and 18 strands/cm (i.e., 42 bundles/cm and 54 bundles/cm) respectively. 5284 epoxy resin was selected as the matrix of the composite tee-joints for long-term application in 150 °C hy-grothermal environment.

Lower angular points (b) Detailed triangular zone

(c) Specimen

Fig. 5 RTM-made EW220/5284 tee-joint.

The permeability of resin through fibres has a significant influence on mechanical properties of the RTM products, therefore it is essential for the permeability simulation of resin through fibre architecture to opti-

mize the processing parameters. From the dimension and detailed geometric configuration in the resin-rich zone of tee-joint (see Figs. 5(a)-5(b)), the FE model for permeability simulation was generated to determine optimized injection mode and parameters of RTM-made tee-joint by using the PAM-RTM code [16]. From the simulated results of resin flowing [17], an optimized injection groove was determined at the bottom and four vents were at the top of mould for discharging air. With the initial preliminary processing parameters of injection pressure p=0.3 MPa, resin viscosity rj= 0.1 Pa-s, fibre volume fraction Vf=55%, fabric permeability ^ = 1.0*10"" m2 and permeability of resin-rich zone k2=1.0*10~10 m2, the optimized processing parameters were obtained as follows: resin viscosity 7=0.1 Pa-s, injection pressure p=0.3 MPa and injection temperature T=60-80 °C.

The 2D orthogonal EW220 glass fibre TWF layers were first stacked up one over the other in a steel mould following the sequence as shown in Fig. 5(b). The 5284 epoxy resin was then infused in the closed steel mould as a matrix following the above optimized processing parameters. After this, the consolidation was conducted. Finally, all RTM-made tee-joints (see Fig. 5(c)) were produced. During consolidation, the composite pre-form was first heated up to 160 °C at a rate of 5 °C/min and held for 1 h at this temperature, and then heated up again at the same rate to the desired temperature of 180 °C, at which the specimen was held for 2 h. Next, the mould was cooled to room temperature and the consolidation was completed.

2.2. Experimental results and discussion

All pull and push bending tests were carried out on MTS880-100kN servo-hydraulic machine in a dry state and at room temperature and a continuous displacement rate of 3 mm/min. Figure 6 shows the boundary conditions and load directions in the tests. From Fig. 6, it is clear that the clamped and pin supports were applied to pull and push bending tests respectively. During the tests, no observable amount of slipping on the boundary conditions was detected. The load-displacement (F-S) curves of specimens in the tests were recorded (see Fig. 7). From Fig. 7, it is apparent that the load-displacement curves of the tested RTM-made tee-joints are almost identical in the linear elastic region. An existence of more than one peak on the F-S curve of all specimens marks the debonding initiation and propagation of resin debonding along the interface between the resin-rich zone and the curved web, which causes load drop on the F-S curve. In order to discuss the failure initiation mechanism, the failure initiation load is defined as the initial load drop, which can be obtained from the F-S curves determined by tests. Table 1 and Fig. 7 illustrate failure initiation loads for pull and push bending. From Table 1, it is obvious that the failure initiation load, on average, is greater for push bending than for pull bending, whereas the scatter is smaller for push bending than for pull bending.

Table 1 Experimental data of failure initiation load

Specimen

Load/N

(b) Push bending test Fig. 6 Boundary condition and load direction

(b) Push bending

Fig. 7 Load-displacement curves

From experimental observation, in pull bending tests, at mean value of pull bending load of about 1 916 N, the resin debonding of all specimens firstly appeared near the upper angular point of the resin-rich zone at the interface between the resin-rich zone and the curved web (see Fig. 8(a)), and there was a small load

Pull bending

Push bending

1 2 429 3 007

2 1 934 2 786

3 1 845 2 717

4 1 567 2 925

5 1 803 2 848

Parameter Pull bending Push bending

Mean/N 1 916 2 857

Standard deviation/N 317 114

Coefficient of variation 0.17 0.04

drop on the load-displacement curves (see Fig. 7(a)). Then the resin debonding propagated along the interface between the triangular resin-rich zone and the adjacent curved web from the upper angular point to the bottom plate (see Fig. 8(b)). This process of debonding propagation ended in the complete separation of the curved web from the bottom plate (see Fig. 8(c)). During debonding propagation process after the first small load drop, the joint continued to carry load until complete separation of the curved web from the bottom plate. Hence, the failure mode of the tested RTM-made tee-joints in pull bending tests can be reckoned to be characteristic of debonding near the upper angular point and debonding propagation along the interface between the triangular resin-rich zone and

(c) Final failure

Fig. 8 Failure sequence under pull bending loading.

the curved web of tee-joint until complete separation of the curved web from the bottom plate.

From the experimental observation of push bending tests (see Fig. 9), it is clear that string after string of white spots (i.e., matrix cracking) first occurred in outer layers of the curved webs (see Fig. 9(a)) and more strings of white spots (or further matrix cracking) were found with increasing load in tests (see Fig. 9(b)). After this, matrix cracking occurred and the fibre fracture then appeared in outer layers of the bottom plate (see Fig. 9(c)) with very small but audible cracking noises and a small load drop on the F-S curves of Fig. 7(b). Furthermore, matrix cracking and fibre fracture continued to take place from outer layers to inner layers of the bottom plate with increasing load until catastrophic collapse of the bottom plate (see Fig. 9(d)). In fact, there is a multi-path of load transfer in the tri-

angular zone of the tee-joint, which is illustrated by a sketch (see Fig. 10). As shown in Fig. 10, the left, intermediate, right and bottom bars represent the simplified load transfer sections of the left curved web, intermediate resin, right curved web and bottom plate of the tee-joint respectively.

Bottom plate Initial crack (d) Final failure

Fig.9 Failure sequence under push bending loading.

Fig. 10 Multi-path sketch of load transfer.

From Fig. 10, it is obvious that 1) under push bending loading, the compression load is transferred to the bottom plate through the multi-path of load transfer (or the left, intermediate and right bars), thus the load-carrying capability of the tee-joint is governed by the failure initiation load of the bottom plate and the failure of the bottom plate is critical failure mode of the tee-joint; 2) the intermediate bar is the shortest and most direct path to transfer the compression load to the bottom plate. Therefore, the intermediate bar (i.e., the resin in the triangular zone) is the dominant load transfer section of the tee-joint, and the damage of the left or right bar cannot change the load transfer mode and transferred load amount significantly. In other words, matrix cracking in outer layers of the curved webs cannot significantly modify the transferred load amount of the resin in the triangular zone to the bottom plate as the dominate load transfer section, and cannot alter the load-displacement curve as well as load-carrying capability of the tee-joint, either. Subsequently, the dominate transferred load through the resin in the triangular zone to the bottom plate induces matrix cracking and fibre fracture from outer layers to inner layers of the bottom plate, until catastrophic collapse of the bottom plate (see Fig. 9(d)). The critical failure mode of the tested RTM tee-joints in push bending tests is matrix cracking and fibre fracture from outer layers to inner layers of the bottom plate until catastrophic collapse and the load-carrying capability is governed by the failure initiation load of the bottom plate of tee-joint. It is evident that in push bending tests, the failure process and mode for the RTM-made EW220/5284 specimens are distinctly different from those in pull bending tests.

Due to the discontinuities of material and owing also to the geometric configuration between the curved webs and the triangular resin-rich zone, there is a stress concentration to cause the debonding at the interface between the curved web and the triangular resin-rich zone. For a tee-joint subjected to a pull bending loading, the curved webs are subjected to through-thickness normal stress and interlaminar shear stress and thus show a tendency to open. When the resin debonding

approaches the bottom plate, there are peeling and opening on the interface between the curved webs and the bottom plate until complete separation of the curved web from the bottom plate. Under push bending loading, the bottom plate is subjected to a bending normal stress. With the increase of push bending loading, the maximum bending normal stress in outer layers of the bottom plate exceeded the longitudinal tensile strength of EW220/5284 to induce matrix cracking and fibre fracture from outer layers to inner layers until catastrophic collapse of the bottom plate.

From the above analysis, it is obvious that the failure initiation load under pull bending loading is dependent on the interfacial strength between plies, while the failure initiation under push bending loading depends on the longitudinal tensile strength of EW220/5284. It is well-known that the tensile strength of a lamina in the fibre direction is much greater than the interfacial strength between plies. Thus the load-carrying capability of the tested RTM-made EW220/5284 tee-joint is greater for push bending than for pull bending.

3. FE Analysis

3.1. Analysis methodology

In order to provide numerical analysis to validate the experimental results, the tee-joint shown in Figs. 5(a)-5(b) is chosen to be modeled and local coordinate systems are then set up to ensure the fibres with correct 3D orientation, i.e., to keep three axial directions 1-3 of the coordinate system consistent with the three normal stresses oi-o3 (see Fig. 11). Figure 11 illustrates the definitions of three axial directions 1-3 of the coordinate system and three normal stresses 01-03 both for the plane and curved overlaminated parts of the plain weave fabric, where the coordinate axes 1-3 denote the longitudinal, transverse and through-thickness directions of the overlaminate respectively. Based on the definitions of three normal stresses 01-03, one has the definitions of three shear stress components Z12, t13 and r23. From Fig. 11, it can be seen that there is a geometrically triangular resin-rich zone bordered by both curved webs and one bottom plate. In order to well suit to model triangular zone, the higher order 3D, 20-node SOLID95 element of ANSYS code [18] with quadratic displacement behavior is implemented to model the zone to attain a high accuracy of simulation. Moreover, 3D, 20-node layered solid element SOLID 191 with three degrees of freedom per node is employed to model the composite layup. The isotropic 5284 resin is used for the triangular zone, while the orthotropic EW220/5284 composite is used for the web, curved web and bottom plate. Thus, a 3D FE model (see Fig. 11) which includes 20 180 hexahedral elements and 94 328 nodes is generated to model stress or strain patterns of tee-joint in association with relevant material properties listed in Table 2. It is worth noting that the material data from the ply is used to define the failure initiation at the ply-resin interface due to the

absence of material data of the adhesion between the ply-resin materials. That is, the resin debonding between the laminate and the resin-rich zone has been assumed to have the same properties as the resin debonding of the ply [19]. The loading and boundary conditions in pull and push bending tests are defined as the clamped and pin supports (see Fig. 6) respectively.

Fig. 11 Tee-joint 3D FE model.

Table 2 Mechanical properties of EW220/5284 TWF composites for FE analysis [19]

Property Value

Longitudinal elastic modulus E^GPa 14.2

Transverse elastic modulus E2/GPa 19.3

Through-thickness elastic modulus E3/GPa 5.0

Poisson's ratio V12 0.15

Poisson's ratio V13 0.01

Poisson's ratio V23 0.01

In-plane shear modulus G^/GPa 4.3

Inter-laminar shear modulus G^/GPa 3.0

Inter-laminar shear modulus G23/GPa 3.0

Ply thickness/mm 0.17

Longitudinal tensile strength X1t/MPa 380

Transverse tensile strength X2t/MPa 493

Through-thickness tensile strength X3t/MPa 50

Longitudinal compressive strength X1c/MPa 312

Transverse compressive strength X2c/MPa 417

Through-thickness compressive strength X3c/MPa 199

In-plane shear strength X^/MPa 111

Inter-laminar shear strength X13/MPa 25

Inter-laminar shear strength X23/MPa 30

Fibre volume fraction/% 57.4

In order to verify the numerical results of tee-joint with above-described FE model, a comparison is carried out between the experimental load-displacement curves and the numerical results (see Fig. 7). Figure 7 demonstrates that, the predicted values to construct the F-S curve from FE results are in good agreement with those from the test in the linear elastic region. The experimental results appear relatively less stiff as against the numerical results. This is probably due to the significant changes in experimental structural stiffness. In fact, even non-observable amounts of slipping and lack of true clamp conditions can change experimental structural stiffness significantly, however, the numerical model applies perfect boundary conditions. This results in the curves of calculated results being stiffer than those shown in the experimental ones.

3.2. Stress pattern and failure prediction

As it considers the interaction between longitudinal, transverse and through-thickness strengths of material, the Tsai-Wu criterion [20] seems more appropriate and effective for predicting the failure of E-glass/epoxy composites, etc. as compared with, for example, the maximum stress or strain rule. However, Tsai-Wu is not suitable for isolating individual damage modes and has fallen out of favor over the years for this reason and for its history in metal failure theories. In fact, the maximum stress approach, which although simplistic, is effective in identifying damage modes, or at least dominant stress components. Thus the maximum stress rule is used for identifying failure modes associated with the individual stress components in this work.

The stress patterns and failure index in the curved webs and the resin-rich zone of tee-joints obtained from the FE model at the mean value of failure initiation load of 2 160 N under pull bending loading are shown in Fig. 12. From Fig. 12, it can be seen that stress concentration occurs along the interfaces between the triangular resin-rich zone and curved web of the joint. This implies that the failure initiation likely appears first along the interface. Six stress components at likely failure initiation location (see Fig. 8(a)) are listed in Table 3. From Tables 2-3, it can be shown that the through-thickness normal stress 03 and interlaminar shear stress r13 are closer to the corresponding strength values than other components. So it can be concluded that the stress components 03 and r13 are the primary reasons to cause the debonding along the interface between the triangular resin-rich zone and curved web. The damage pattern (see Fig. 12(i)) in the curved webs is then determined from applying the maximum stress criteria to the stress patterns for indicating the failure initiation location near the upper angular point at the interface between the triangular resin-rich zone and the curved web (see Fig. 8(a)) and the maximum failure ratio of R for maximum stress criterion at relevant failure initiation location is 1.017 (see Table 3). This is exactly consistent with the findings about the failure initiation process and mode in tests introduced in the above section.

Table 3 Stress components used in failure criterion

around 1.57. The predictions correlate well with the findings about failure initiation process and mode in tests (see Figs. 9(a)-9(b)). From Fig. 13(g), it is also

Parameter Pull bend- Push bend- Parameter Pull bend- Push bend-

ing mg ing ing

Load/N 2 160 -2 820 ü2/MPa 0 -4.7

Location Curved web Bottom plate ^3/MPa 0 - 0.5

CTi/MPa 18.5 418.7 ri3/MPa 2i.9 - 0.7

a2/MPa -3.1 3.0 R i.0i7 i.ii

CT3/MPa 8.5 1.7

Fig. 13 shows stress patterns and failure index in the bottom plate at a mean value of failure initiation load of 2 820 N under push bending loading. Fig. 13(g) shows that the maximum stress of the joint is a compressive stress of around 491 MPa, which occurs in the curved webs and exceeds the longitudinal compression strength listed in Table 2, and the failure index is

Fig. 12 Stress and failure ratio in the curved webs and the resin-rich zone at the mean value of failure initiation load of 2 160 N under pull bending loading.

apparent that stress concentration exists in the bottom plate near the corner of the angular zone, where the failure initiation likely appears as shown in Fig. 9(c). Six stress components at the likely failure initiation location (see Fig. 9(c)) are listed in Table 3. From Tables 2-3, it is clear that the maximum normal stress o\ in the bottom plate near the corner of the angular zone equals 418.7 MPa, which exceeds the longitudinal tensile strength (or 380 MPa) of EW220/5284 TWF composites. Consequently, the failure ratio R for the maximum stress criterion in the bottom plate is greater than 1.0 (see Fig. 13(h) and Table 3) and this corresponds to matrix cracking and fibre fracture resulting from the push bending that appears in outer layers of the bottom plate (see Fig. 9(c)). Thus the predictions

correlate well with the findings about failure initiation process and mode in experiments.

From the experimental and calculated results of failure initiation loads in Table 4, the predicted values of failure initiation loads of the tested RTM-made EW220/ 5284 tee-joint under pull and push bending loading amount to about 2 160 N and 2 820 N respectively. These agree fairly well with the experimental mean values of failure initiation load, 1 916 N and 2 857 N, implying that the FE models can give an insight to understand the load transfer path and failure initiation mechanism in the tested RTM-made EW220/5284 TWF tee-joint under pull and push bending loading. However, in Table 4, it can be seen that the difference for the failure initiation load under pull bending loading is 12.7% and there is much higher difference for the

Fig. 13 Stress and failure ratio in the bottom plate at the mean value of failure initiation load of 2 820 N under push bending loading.

pull bending loading case compared with the push bending loading case. This is because failure initiation load for pull bending is dependent on the interfacial strength between plies, while failure initiation for push bending depends on longitudinal tensile strength of

Table 4 Failure initiation loads of tee-joint

Method Pull bending/N Relative deviation/%

Prediction 2 160 12.7

Experiment 1 916

Method Push bending/N Relative deviation/%

Prediction 2 820 1.3

Experiment 2 857

EW220/5284; actually, the interfacial strength between plies is more sensitive to the process accuracy and has greater scatter than the longitudinal tensile strength of EW220/5284.

4. Conclusions

This paper focuses on experimental and numerical studies of the behavior of 2D orthogonal EW220/5284 TWF composite tee-joints manufactured by the RTM technique under pull and push bending loading. It is shown that the proposed FE model could be used in practice and possesses acceptable accuracy. The significant results emerging from the studies are as follows.

1) It is clear that failure process and mode for 2D orthogonal EW220/5284 TWF composite tee-joints in push bending tests are distinct different from those in pull bending tests. The failure mode of the tested RTM-made tee-joints in pull bending tests can be reckoned to be characteristic of resin debonding at the interface between the triangular resin-rich zone and the curved web of tee-joint until complete separation of the curved web from the bottom plate, which is in agreement with references. In contrast, the critical failure initiation in push bending tests can be concluded to be characteristic of bending induced matrix cracking and fibre fracture from outer layers to inner layers until catastrophic collapse of the bottom plate.

2) The test show that the failure initiation load of 2D orthogonal EW220/5284 TWF composite tee-joints manufactured by the RTM technique, on average, is greater for push bending than for pull bending, whereas the scatter is smaller for push bending than for pull bending. Failure initiation load for pull bending is dependent on the interfacial strength between plies, consistent with the lessons of previous works, while the failure initiation for push bending depends on longitudinal tensile strength of EW220/284 TWF laminate. Actually, the longitudinal tensile strength of EW220/5284 TWF lamina that governs the failure initiation of the tested RTM-made tee-joints in push bending, is much greater than the interfacial strength between plies, which dominates in pull bending. Thus the load-carrying capability of 2D orthogonal EW220/5284 TWF composite tee-joints manufactured by the RTM technique is greater for push bending than for pull bending.

3) The FE models can give an insight to understand the load transfer path and failure initiation mechanism in 2D orthogonal EW220/5284 TWF composite tee-joints manufactured by the RTM technique under pull and push bending loading. The failure initiation load and pattern predicted by FE analysis are well consistent with that of experimental, demonstrating that the through-thickness normal stress 03 and interlaminar shear stress r13 result in the failure initiation of tee-joint under pull bending loading while the bending normal stress causes the critical failure initiation under push bending loading.

References

[1] Falzon B G, Stevens K A, Davies G O. Postbuckling behaviour of a blade-stiffened composite panel loaded in uniaxial compression. Composites: Part A 2000; 31(5): 459-468.

[2] Orifici A C, Shah S A, Herszberg I, et al. Failure analysis in postbuckled composite T-sections. Composite Structures 2008; 86(1-3): 146-153.

[3] Shenoi R A, Hawkins G L. Influence of material and geometry variations on the behaviour of bonded tee connections in FRP ships. Composites 1992; 23(5): 335-345.

[4] Stickler P B, Ramulu M, Johnson P S. Experimental and numerical analysis of transverse stitched T-joints in bending. Composite Structures 2000; 50(1): 17-27.

[5] Vijayaraju K, Mangalgiri P D, Dattaguru B. Experimental study of failure and failure progression in T-stiffened skins. Composite Structures 2004; 64(2): 227-234.

[6] Greenhalgh E, Lewis A, Bowen R, et al. Evaluation of toughening concepts at structural features in CFRP—Part I: stiffener pull-off. Composites: Part A 2006; 37(10): 1521-1535.

[7] Dharmawan F, Li H C H, Herszberg I, et al. Applicability of the crack tip element analysis for damage prediction of composite T-joints. Composite Structures 2008; 86(1-3): 61-68.

[8] Blake J I R, Shenoi R A, House J, et al. Progressive damage analysis of tee joints with viscoelastic inserts. Composites: Part A 2001; 32(5): 641-653.

[9] Falzon B G, Davies G A O, Greenhalgh E. Failure of thick-skinned stiffener run out sections loaded in uniaxial compression. Composite Structures 2001; 53(2): 223-233.

[10] Li H C H, Dharmawan F, Herszberg I, et al. Fracture behaviour of composite maritime T-joints. Composite Structures 2006; 75(1-4): 339-350.

[11] Li R, Kelly D, Crosky A. An evaluation of failure criteria for matrix induced failure in composite materials. Composite Structure 2002; 57(1-4): 385-391.

[12] Ford T. Aerospace composites. Aircraft Engineering and Aerospace Technology 1997; 69(4): 334-342.

[13] Yi X S, An X F. Developments of high-performance composites by innovative ex situ concept for aerospace application. Journal of Thermoplastic Composites Materials 2009; 22(1): 29-49. [in Chinese]

[14] Pandiyan K R R, Kundu G, Neogi S, et al. Development of manufacturing technology for cab front using

resin transfer molding process. Journal of Composite Materials 2010; 44(18): 2217-2231.

[15] Gong Y, Yang G S. All-polyamide composites prepared by resin transfer molding. Journal of Materials Science 2010; 45(19): 5237-5243.

[16] PAM-RTM. User's guide & tutorials: Version 2004.1. ESI Group, 2004.

[17] Bai J B, Xiong J J, Cheng X, et al. Processing parameter optimization and experimental investigation on mechanical properties for RTM composite T-joint. Acta Materiae Compositae Sinica 2009; 26(3): 13-17. [in Chinese]

[18] Release 11.0. Documentation for ANSYS. SAS IP, Inc., 2007.

[19] Peng B. Mechanical properties of tee-joints processed by co-RTM. Beijing: School of Aeronautic Science and Engineering, Beihang University, 2007. [in Chinese]

[20] Daniel I M, Ishai O. Engineering Mechanics of Composite materials. 2nd ed. Oxford: Oxford University Press, 2006.

Biographies:

LUO Chuyang graduated with a B.S. degree in China Three Gorges University in 2006. He received his Ph.D. degree at Beihang University in 2011. Since then he has worked for the China Airborne Missile Academy on Composite Structure Design.

E-mail: cyluo@ase.buaa.edu.cn

XIONG Junjiang graduated in 1986 with a B.S. degree from Northwest Polytechnic University, then worked for the Institute of Helicopter Design and Research from 1986 to 1989. From 1989 to 1995, he worked for academic degrees at Beihang University where he earned master's and doctor's degrees. Since then he has been working at the same University as a post-doctoral researcher, an associate professor and a professor. He is also a doctoral supervisor in Aircraft Design. As a chief investigator, he has engaged in a number of research projects sponsored by National Natural Science Foundation, National Defense Science Foundation, Aeronautics Science Foundation of China, etc. He also received funding from the industry. He published 7 books and over 70 papers on archival journals concerning fatigue and fracture reliability engineering. As a chief contributor, he has won 6 awards from the competent ministry of China. E-mail: jjxiong@buaa.edu.cn