Effect of out-of-plane wrinkles in curved multi-directional carbon/epoxy laminates

https://doi.org/10.1016/j.compscitech.2020.108282Get rights and content

Abstract

Defects such as out-of-plane wrinkles are known to strongly affect in-plane strength but there has been very little research on their effect on out-of-plane properties. Experimental and numerical studies of multi-directional curved-beam laminates were thus carried out to understand the effects of out-of-plane wrinkles on through-thickness tensile strength. The initially selected layup saw free-edge delamination interacting with transverse cracking, which is undesirable. After suppressing the free-edge delamination by dispersing the plies near the specimen surfaces, through-thickness tensile failure was observed near the mid-plane. The effects of out-of-plane wrinkles could be studied with this appropriate layup, showing a 16% reduction in strength. A High-fidelity Finite Element Method (Hi-FEM) has been used to distinguish between the different failure modes and to understand the effects of wrinkles. Good agreement was achieved between the numerical and experimental results in terms of through-thickness tensile strengths and delamination locations.

Introduction

Laminated composite materials have outstanding in-plane properties such as high stiffness and strength-to-weight ratios [1], but relatively low interlaminar properties. This is because of a lack of through-thickness reinforcement normal to the fibre direction, so the laminate relies on the matrix to withstand loading. Delamination is a commonly observed type of failure in laminated composites, which can occur due to high interlaminar stresses developed at free edges and discontinuities [2], through-thickness tensile loading and geometry curvature [3]. Through-thickness tensile strength (σImax) is a key characteristic of a composite laminate's resistance to delamination initiation and so its accurate prediction is important for composite structures.

An ASTM D6415/D6415M – 06a standard [4] which utilizes a curved-beam configuration under pure bending introduced by a 4-point bending jig is available to characterize σImax for fibre-reinforced polymer-matrix composites. The curved-beam configuration can generate high through-thickness tensile stresses near the mid-plane the specimen, and σImax can be determined from the critical bending moment at the first load drop based on Lekhnitskii [5] or Kedward et al. [6]. It is still quite challenging to know when the measured σImax can be used as a material property. For example, Jackson and Martin [7] tested Unidirectional (UD) AS4/3501-6 carbon/epoxy curved bean specimens of different thicknesses (ply thickness is 0.123 mm in the thin specimens). They found that the thinner specimens showed higher mean σImax than the thicker ones, owing to the observed large macroscopic flaws such as voids or resin pockets, which could be explained by volume scaling laws. This is consistent with [8] in which significant size effects on interlaminar tensile strength were demonstrated in volume scaled UD glass/epoxy curved-beam laminates.

The ASTM D6415/D6415M – 06a standard [4] also highlights the difficulty in using non-unidirectional specimens in which matrix cracks and free-edge stresses may cause errors in the σImax calculations. Lekhnitskii's [5] equations assume homogeneity, so σImax is expected to be independent of specimen stacking sequences. This was found to be true for specimens of UD and different Quasi-isotropic (QI) stacking sequences with a laminate thickness between 4.2 and 8.4 mm by Charrier et al. [9], who reported a σImax value of 47 MPa for T700GC/M21 carbon/epoxy composite laminates of different stacking sequences (ply thickness is 0.262 mm). In contrast, González-Cantero et al. [10] found that it is not the case for the specimens of different QI stacking sequences (ply thickness is 0.184 mm) with a laminate thickness between 1.7 and 8.8 mm, in which the transverse cracking in the 90° plies under high in-plane tensile stresses plays a key role in delamination onset. Their stress analyses were backed by up by the observation of different delamination locations with different laminate thicknesses and stacking sequences. Tasdemir and Coker [11] highlighted a similar effect of transverse cracks in 90° plies in their AS4/8552 carbon/epoxy cross-ply laminates (ply thickness is 0.184 mm). They reported the σImax values of 55 MPa and 62 MPa for the two stacking sequences used. Nguyen et al. [12] tested and modelled L-shape carbon/epoxy (USN–150B, SK Chemicals, Korea) specimens with a QI stacking sequence. They compared different input parameters such as the interlaminar tensile and shear strengths in their FE analyses and correlated satisfactorily against the experimental results. However, they [12] did not consider the effect of transverse cracks.

Free edges give rise to stress concentrations due to high interlaminar shear and tensile stresses in some composite laminates [13]. Free-edge effects have been studied extensively [2,[14], [15], [16], [17]]. In some laminates, these effects can lead to premature failure initiating at the free edges. Free-edge effects in curved-beam tests have been studied by Fletcher et al. [18], and were significantly reduced by using an edge layer of resin, achieving 16% higher Curved Beam Strength (CBS) which is equivalent to approximately 49 MPa for IMA/M21 carbon/epoxy multidirectional laminates with a ply thickness of 0.25 mm. In contrast, Charrier et al. [9] claimed that the free-edge effects are negligible in their T700GC/M21 carbon/epoxy QI laminates. The change in strength caused by changing ply block thickness is usually more significant than by just changing the laminate volume like in Ref. [7,8]. Laminates with thicker ply blocks have more energy available to drive failure due to matrix cracking and delamination [19,20]. Therefore, dispersing the plies may suppress free-edge delamination, potentially resulting in a different failure mode.

Defects and features, including out-of-plane wrinkles, can occur during composites manufacture. Previously, the effects of defects and features in flat coupons have been studied experimentally [[21], [22], [23], [24], [25]] and numerically [26]. It was concluded that the effects on in-plane tensile and compressive strengths are most significant when gaps and overlaps are combined in a complex network [23,26]. Among these defects and features, the commonly observed out-of-plane wrinkles were found to influence the tensile and compressive strengths [27,28]. However, it has not been studied yet how such defects influence the through-thickness tensile strength.

In the current work, a method adapted from the ASTM D6415/D6415M – 06a standard has been used to determine σImax by using multidirectional curved-beam specimens. After observing undesirable delamination near the inner radius with an initial stacking sequence, an improved stacking sequence was used to characterize σImax. This was done by dispersing the plies near both surfaces, and the delamination location was changed to be nearer the mid-plane. Based on the second layup, a further study of the effects of out-of-plane wrinkles on through-thickness tensile strength was carried out. A High-fidelity Finite Element Method (Hi-FEM) which can predict damage evolution was used to help understand the competing failure modes. Hi-FEM can distinguish between complex failure mechanisms such as free-edge effects, transverse cracking and through-thickness tensile failure. Good agreement has been achieved between the Hi-FEM predictions and the experimental results in terms of through-thickness tensile strength, the failure mode and delamination location.

Section snippets

Experimental study

The test method follows the ASTM D6415/D6415M – 06a standard [4], but with increased radii to accommodate out-of-plane wrinkles and increased width to minimize the interaction between the two free edges. The test set-up comprises a 4-point bending test jig and a curved-beam specimen as shown in Fig. 1. In the test jig, four rollers are in contact with the specimen, which have a diameter of 20 mm. The diameter is larger than that suggested by the ASTM standard, and the purpose is to reduce the

Numerical study

A High-fidelity Finite Element Method (Hi-FEM) has been adopted using cohesive interface elements between each adjacent ply and within each ply along pre-defined potential split paths to help understand the competing failure modes in the curved-beam tests and predict overall failure in detail. This is difficult to examine experimentally due to the complex nature of the different failure mechanisms e.g. matrix cracking, free-edge delaminations and through-thickness tensile failure. Only after

Discussion

González-Cantero et al. [10] pointed out that while through-thickness tension is a main reason for the unfolding failure in curved laminates, transverse cracks in 90° plies is another secondary failure mechanism. The experimental and numerical results presented here are consistent with the previous findings. For example, in the Pristine Layup 1 case, delamination was near the inner radius where the transverse tensile stress is high. The Hi-FEM model predicts the cohesive elements representing

Summary and conclusions

This work demonstrated the effect of out-of-plane wrinkles on through-thickness tensile strength in curved multi-directional laminates. This was done by carefully choosing the appropriate stacking sequence to highlight and understand the two competing failure modes.

An adaptation of the ASTM D6415/D6415M – 06a standard was successfully applied to multidirectional curved-beam specimens. Initially, an undesirable delamination located near the inner radius was observed in the curved-beam specimens

CRediT authorship contribution statement

Xiaodong Xu: Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing - original draft, Writing - review & editing. Mike I. Jones: Methodology, Validation, Investigation, Resources. Hafiz Ali: Investigation, Resources. Michael R. Wisnom: Conceptualization, Writing - review & editing, Supervision, Funding acquisition. Stephen R. Hallett: Conceptualization, Methodology, Writing - review & editing, Supervision, Project administration, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

The authors would like to acknowledge the support of this work by Rolls-Royce, through the Composites University Technology Centre (UTC) at the University of Bristol, UK.

References (33)

Cited by (27)

  • Consolidation mechanism of composite corners cured on convex and concave tools

    2023, Composites Part A: Applied Science and Manufacturing
View all citing articles on Scopus
View full text