betaFeedback

WorksComputer Science

Structural cohesive element for the modelling of delamination in composite laminates without the cohesive zone limit

Abstract

Delamination is a critical mode of failure that occurs between plies in a composite laminate. The cohesive element, developed based on the cohesive zone model, is widely used for modeling delamination. However, standard cohesive elements suffer from a well-known limit on the mesh density-the element size must be much smaller than the cohesive zone size. This work develops a new set of elements for modelling composite plies and their interfaces in 3D. A triangular Kirchhoff-Love shell element is developed for orthotropic materials to model the plies. A structural cohesive element, conforming to the shell elements of the plies, is developed to model the interface delamination. The proposed method is verified and validated on the classical benchmark problems of Mode I, Mode II, and mixed-mode delamination of unidirectional laminates, as well as on the single-leg bending problem of a multi-directional laminate. All the results show that the element size in the proposed models can be ten times larger than that in the standard cohesive element models, with more than 90% reduction in CPU time, while retaining prediction accuracy. This would then allow more effective and efficient modeling of delamination in composites without worrying about the cohesive zone limit on the mesh density.

Introduction

The accurate prediction of delamination is of critical importance for the reliable design of fiber-reinforced composite structures. The cohesive element (CE) is a widely used finite element technology to model delamination. CE is developed based on the Cohesive Zone Model proposed by Dugdale and Barenblatt [1, 2]. A fracture process zone, generally called the cohesive zone, exists along the interface, ahead of the stress-free crack tip. A traction-separation relationship, namely the cohesive law, describes how the interfacial stresses and damage evolve with respect to the interfacial openings. Standard CEs are usually developed for use between two solid elements to model their debonding [3, 4, 5, 6, 7, 8, 9, 10].1

While being a popular and versatile tool to model delamination, standard CEs suffer from a well-known limit on the mesh density – the element size must be much smaller than the Cohesive Zone Length (CZL) to accurately predict delamination. According to the previous analysis [3, 4], high-stress gradients could be produced within the cohesive zone during delamination in composites. A very fine CE mesh must be used there to sufficiently capture the stress gradients such that the internal virtual work of the CEs can be accurately integrated. So far, there is no fixed rule on how fine the mesh should be in the literature. In some research [5, 6, 7], the authors have demonstrated that at least two or three CEs should be used inside the cohesive zone. With coarser meshes, simulations would significantly over-predict the peak load [8]. In a typical mode I delamination test of a unidirectional composites coupon, i.e., the Double Cantilever Beam (DCB) test, the CZL is less than 1 mm.

The above-mentioned problem of cohesive zone limit on the mesh density of CEs has drawn the attention of many researchers in the past. Turon et al. [8] adopted an engineering method to solve the problem by reducing the material strength to numerically extend the CZL. While shown to be promising in the Mode I DCB case, in cases such as the pure mode II End-Notched-Flexure (ENF) test, decreasing the cohesive strength can cause excessive under-prediction of the overall strength [11, 12]. Yang [13] and Do [14]’s research work demonstrated that larger CE sizes can be achieved by placing enough integration points in the cohesive zone. Their method could predict the peak load correctly for CE size up to 1.43 times the estimated CZL. However, its first-order shape function limits the effectiveness of this method in further expanding CE beyond the scale of the cohesive zone [15]. Guiamatsia et al. [16, 17, 18] used the beam on elastic foundation solution as an enrichment function and tested it in the mixed mode delamination. However, the enrichment method could lead to inaccuracies in interpolation for elements larger than 3 mm [17]. Another enrichment approach with the piecewise linear shape functions was proposed by Samimi [19, 20, 21]. However, the over-prediction problem of peak load under large CE has not been solved. Van der Meer et al. [22] used the level set method with a energy-based criterion to propagate delamination without the cohesive zone limit. However, this method is limited to the case of a single delamination. Lu et al. [23] proposed a adaptive version of the Floating Node Method (FNM) [24] to adaptively refine the CEs in the cohesive zone. However, this method is currently implemented for 2D problems. Alvarez et al. [25] used quadratic CEs between quadratic solid elements and used higher number of integration points, which increased the grid size to be comparable to CZL but not beyond [15]. Mukhopadhyay and Bhatia [26] developed a hp refinement strategy to simulate the delamination between two solid elements. However, it is only implemented in 2D in their work. Daniel [27] developed an ERR-Cohesive method to simulate the delamination with large elements by estimating the energy release rate by means of the virtual crack closure technique (VCCT). However, this method is only presented in 2D, and the VCCT approach relies on the existence of an initial crack and the assumption of self-similar crack propagation.

Inspired by the earlier works in the literature, Russo and Chen [15] developed a so-called structural CE, which conforms to Kirchhoff-Love structural elements for the neighbouring plies. Their work was done in 2D. The Euler-Bernoulli beam elements were used to model the plies. The structural CE, sharing its nodes with the beam elements, was developed to model the delamination. An adaptive integration scheme was used to place more integration points in CEs containing the cohesive zone. Their results showed that the structural CE could overcome the cohesive zone limit on mesh density, allowing the element size to be ten times larger than that of the standard linear CEs. Motivated by Russo and Chen’s work, Tosti Balducci and Chen [28] extended the structural CE to 3D DCB problem by developing a structural CE compatible with the TUBA3 plate elements [29]. Their results showed that the TUBA3-based structural formulation of CE could overcome the cohesive zone limit. However, the curvature degrees of freedom (DoFs) make it more complicated to set boundary conditions on TUBA3 elements than on more common elements such as solid, beam or Reissner-Mindlin shell elements, thereby impeding the adoption of such elements by engineers in practice.

From the above reviews, we can see that a composites delamination model which:

  1. does not suffer from the cohesive zone limit on mesh density;

  2. works in 3D space;

  3. does not require artificial reduction of strengths;

  4. does not require a pre-crack and the self-similar propagation assumption;

  5. can model delamination along multiple interfaces; and

  6. does not require working with curvature DoFs,

is still missing. Based on the earlier works [15, 28], this article aims to develop a new set of elements for the plies and interfaces of composites. The idea is to develop a triangular Kirchhoff-Love shell element without curvature DoFs to model the composite plies, then formulate the conforming structural CE to be used between the plies to model their delamination. Allman [30] proposed a simple triangle Kirchhoff-Love plate element named cubic displacement element. There are only three DoFs at each node, one for out-of-plane displacement and two for rotations. These DoFs are commonly used by engineers and are much easier to handle than the curvature ones when setting boundary conditions. Therefore, this article will extend this triangle plate element for the modeling of composite plies and develop the corresponding structural CE for the interfaces. If such a structural CE could overcome the cohesive zone limit on its mesh density, then it would achieve all the requirements listed above.

The rest of the paper will be structured in the following way. § 2 presents the proposed element formulations in detail. § 3 demonstrates the performance of these elements on a series of benchmarks on delamination in both unidirectional and multi-directional laminates. In the end, § 4 draws the conclusions of this work and discusses some potential future work.

Method

Overall illustration

The overall idea of the proposed modelling approach is to represent the composite plies by Kirchhoff-Love shell elements and the interfaces by structural CEs. Figure 1 shows the geometrical comparison between the conventional modelling approach with solid elements and the proposed structural approach with shells. Looking at the structural CE, the difference from the conventional CE is that the nodes of the CE are placed at the mid-plane of the two shell elements, which are not the actual surfaces of the cohesive interface. Hence, the opening of the actual interface does not equate to the distance between the upper nodes and lower ones of the structural CE. This opening shall be calculated using the shell kinematics to be detailed in later subsections. The rest of this section will firstly present the cubic plate element formulation by Allman [30] and its adaptation for flat composites shell, then move on to derive the structural CE formulation.

Geometrical comparison between conventional CE and structural CE

Figure 1. Geometrical comparison between conventional CE and structural CE

Cubic plate element

Geometric definitions of the plate element

The plate part of the shell element in this work is based on the triangular cubic plate element developed by Allman [30], as shown in Figure 2. A local coordinate system is used in this element, with origin at the centroid of the triangle element. The local axis are represented by lowercase letters, xx and yy, to distinguish them from the global ones, XX and YY. The area of this triangular element is AA. ss is the anti-clockwise coordinate along the element boundary. nn is the exterior normal. The angle between the normal nn and the local axis xx is γ\gamma. The out-of-plane displacement defined over the domain is w(x,y)w(x,y). An independent out-of-plane boundary displacement, wˉ(s)\bar{w}(s), and its compatible normal derivative, ∂wˉ∂n(s)\frac{\partial\bar{w}}{\partial n}(s), are assumed along the boundary. The DoFs at each node include the displacement wˉ\bar{w} and two rotations ∂wˉ∂x\frac{\partial\bar{w}}{\partial x}, ∂wˉ∂y\frac{\partial\bar{w}}{\partial y}.

Image not extracted

Figure 2. Coordinate system and DoFs for the triangular cubic plate element [30]

The cubic plate element in Allman’s work only considered the case of isotropic material [30]. This work extends the cubic plate element using the classical laminate theory, such that symmetric composite laminates can also be modelled by this plate element. The original cubic element formulation will be presented in detail, with adaptations for composites specified along the way.

Modified potential energy for the plate element

The minimum potential energy principle is used to derive the finite element formulation of the triangular plate element. The potential energy used is referred to as the modified potential energy in Allman’s work [30]:

π=∬AU0 dxdy+∑N=13RN(wˉN−wN)+∫∂AVn(wˉ−w) ds−∫∂AMn(∂wˉ∂n−∂w∂n) ds−∬Ap∗wˉ dxdy−∑N=13RN∗wˉN−∫∂AVn∗wˉ ds+∫∂AMn∗∂wˉ∂n ds(1)\begin{aligned} \pi={}& \iint_A U_0\,\mathrm{d}x\mathrm{d}y + \sum_{N=1}^{3} R_N(\bar{w}_N-w_N) + \int_{\partial A} V_n(\bar{w}-w)\,\mathrm{d}s - \int_{\partial A} M_n\left(\frac{\partial\bar{w}}{\partial n}-\frac{\partial w}{\partial n}\right)\,\mathrm{d}s \\ &- \iint_A p^*\bar{w}\,\mathrm{d}x\mathrm{d}y - \sum_{N=1}^{3} R_N^*\bar{w}_N - \int_{\partial A} V_n^*\bar{w}\,\mathrm{d}s + \int_{\partial A} M_n^*\frac{\partial\bar{w}}{\partial n}\,\mathrm{d}s \tag*{(1)} \end{aligned}

where U0U_0 is the strain energy density. MnM_n is the normal bending moment resultant and VnV_n is the Kirchhoff shear force distribution on the element boundary, respectively. RNR_N (N=1,2,3N=1,2,3) are the concentrated forces at the element vertices. In addition, RN∗R_N^*, Vn∗V_n^*, and Mn∗M_n^* are the values of prescribed concentrated force, Kirchhoff shear force, and normal bending moment resultant, respectively. p∗(x,y)p^*(x,y) denotes the prescribed distributed pressure load on the element. This potential energy essentially uses RNR_N, VnV_n, and MnM_n as Lagrange multipliers to enforce the compatibility between the two fields ww and wˉ\bar{w} and their normal derivatives along the boundary. The variational principle based on this modified potential energy satisfies the equilibrium equations, the boundary conditions, and the compatibility requirement between ww and wˉ\bar{w} [30].

The expression of U0U_0 for composites will be different from that for isotropic materials in Allman’s work [30]. The classical laminate theory can be used to describe the constitutive relationship of a composite laminate under Kirchhoff-Love kinematic assumptions.

For the case of a symmetric laminate (the smallest of which would be a single composite ply) considered in this work, there would be no membrane-bending coupling. Therefore, the expressions of the moments are:

Mx=−D11∂2w∂x2−D12∂2w∂y2−2D16∂2w∂x∂yMy=−D12∂2w∂x2−D22∂2w∂y2−2D26∂2w∂x∂yMxy=−D16∂2w∂x2−D26∂2w∂y2−2D66∂2w∂x∂y(2)\begin{aligned} M_x &= -D_{11}\frac{\partial^2 w}{\partial x^2} - D_{12}\frac{\partial^2 w}{\partial y^2} - 2D_{16}\frac{\partial^2 w}{\partial x\partial y} \\ M_y &= -D_{12}\frac{\partial^2 w}{\partial x^2} - D_{22}\frac{\partial^2 w}{\partial y^2} - 2D_{26}\frac{\partial^2 w}{\partial x\partial y} \\ M_{xy} &= -D_{16}\frac{\partial^2 w}{\partial x^2} - D_{26}\frac{\partial^2 w}{\partial y^2} - 2D_{66}\frac{\partial^2 w}{\partial x\partial y} \tag*{(2)} \end{aligned}

Using the above moment-curvature relations (equation (2)), the strain energy density U0U_0 can be written as:

U0=12[D11(∂2w∂x2)2+D22(∂2w∂y2)2+2D12∂2w∂x2∂2w∂y2+4D16∂2w∂x2∂2w∂x∂y+4D26∂2w∂y2∂2w∂x∂y+4D66(∂2w∂x∂y)2](3)\begin{aligned} U_0 = \frac{1}{2}\biggl[ &D_{11}\left(\frac{\partial^2 w}{\partial x^2}\right)^2 + D_{22}\left(\frac{\partial^2 w}{\partial y^2}\right)^2 + 2D_{12}\frac{\partial^2 w}{\partial x^2}\frac{\partial^2 w}{\partial y^2} \\ &+ 4D_{16}\frac{\partial^2 w}{\partial x^2}\frac{\partial^2 w}{\partial x\partial y} + 4D_{26}\frac{\partial^2 w}{\partial y^2}\frac{\partial^2 w}{\partial x\partial y} + 4D_{66}\left(\frac{\partial^2 w}{\partial x\partial y}\right)^2\biggr] \tag*{(3)} \end{aligned}

The sign conventions for the shear force and bending moment resultants are shown in Figure 3. For the triangular plate element, the normal bending moment resultant MnM_n, Kirchhoff shear force distribution VnV_n and concentrated forces RNR_N (N=1,2,3N=1,2,3) can be expressed as [30]:

Image not extracted

Figure 3. Sign conventions for shear force and bending moment resultants

Mn=Mxcos⁡2γ+Mysin⁡2γ+Mxysin⁡2γ(4)M_n = M_x\cos^2\gamma+ M_y\sin^2\gamma+ M_{xy}\sin2\gamma \tag*{(4)}
Vn=Qn+∂Mns∂s(5)V_n = Q_n + \frac{\partial M_{ns}}{\partial s} \tag*{(5)}
RN=[Mns]s−s+(6)R_N = [M_{ns}]_{s^-}^{s^+} \tag*{(6)}

where

Mns=12(My−Mx)sin⁡2γ+Mxycos⁡2γ(7)M_{ns} = \frac{1}{2}(M_y - M_x)\sin2\gamma+ M_{xy}\cos2\gamma \tag*{(7)}
Qn=∂Mn∂n+∂Mns∂s(8)Q_n = \frac{\partial M_n}{\partial n} + \frac{\partial M_{ns}}{\partial s} \tag*{(8)}

In this element, linear variation of the normal bending moment resultant and constant Kirchhoff shear force distribution are assumed along the edges:

References

  1. [1]G. I. Barenblatt, The mathematical theory of equilibrium cracks in brittle fracture, Advances in applied mechanics 7 (1962) 55–129.DOI
  2. [2]D. S. Dugdale, Yielding of steel sheets containing slits, Journal of the Mechanics and Physics of Solids 8 (1960) 100–104.DOI
  3. [3]Y. Qiu, M. A. Crisfield, G. Alfano, An interface element formulation for the simulation of delamination with buckling, Engineering Fracture Mechanics 68 (2001) 1755–1776.
  4. [4]Q. Yang, B. Cox, Cohesive models for damage evolution in laminated composites, International Journal of Fracture 133 (2005) 107–137.DOI
  5. [5]P. P. Camanho, C. G. Dávila, Mixed-mode decohesion finite elements for the simulation of delamination in composite materials, NASA Technical Reports 211737 (2002).
  6. [6]N. Moës, T. Belytschko, Extended finite element method for cohesive crack growth, Engineering fracture mechanics 69 (2002) 813–833.DOI
  7. [7]X.-P. Xu, A. Needleman, Void nucleation by inclusion debonding in a crystal matrix, Modelling and Simulation in Materials Science and engineering 1 (1993) 111.DOI
  8. [8]A. Turon, C. G. Davila, P. P. Camanho, J. Costa, An engineering solution for mesh size effects in the simulation of delamination using cohesive zone models, Engineering fracture mechanics 74 (2007) 1665–1682.DOI
  9. [9]A. Pirondi, G. Giuliese, F. Moroni, A. Bernasconi, A. Jamil, Comparative study of cohesive zone and virtual crack closure techniques for three-dimensional fatigue debonding, The Journal of Adhesion 90 (2014) 457–481.DOI
  10. [10]J. Chen, M. Crisfield, A. Kinloch, E. P. Busso, F. L. Matthews, Y. Qiu, Predicting progressive delamination of composite material specimens via interface elements, Mechanics of composite materials and structures 6 (1999) 301–317.DOI
  11. [11]P. W. Harper, S. R. Hallett, Cohesive zone length in numerical simulations of composite delamination, Engineering Fracture Mechanics 75 (2008) 4774–4792.DOI
  12. [12]X. Lu, M. Ridha, B. Chen, V. Tan, T. Tay, On cohesive element parameters and delamination modelling, Engineering Fracture Mechanics 206 (2019) 278–296.DOI
  13. [13]Q. Yang, X. Fang, J. Shi, J. Lua, An improved cohesive element for shell delamination analyses, International journal for numerical methods in engineering 83 (2010) 611–641.DOI
  14. [14]B. Do, W. Liu, Q. Yang, X. Su, Improved cohesive stress integration schemes for cohesive zone elements, Engineering Fracture Mechanics 107 (2013) 14–28.DOI
  15. [15]R. Russo, B. Chen, Overcoming the cohesive zone limit in composites delamination: modeling with slender structural elements and higher-order adaptive integration, International Journal for Numerical Methods in Engineering 121 (2020) 5511–5545.DOI
  16. [16]I. Guiamatsia, J. Ankersen, G. Davies, L. Iannucci, Decohesion finite element with enriched basis functions for delamination, Composites Science and Technology 69 (2009) 2616–2624.DOI
  17. [17]I. Guiamatsia, G. Davies, J. Ankersen, L. Iannucci, A framework for cohesive element enrichment, Composite structures 92 (2010) 454–459.DOI
  18. [18]I. Guiamatsia, J. Ankersen, L. Iannucci, A study of mixed-mode composite delamination using enriched interface elements, The Aeronautical Journal 117 (2013) 959–967.DOI
  19. [19]M. Samimi, J. Van Dommelen, M. Geers, An enriched cohesive zone model for delamination in brittle interfaces, International journal for numerical methods in engineering 80 (2009) 609–630.DOI
  20. [20]M. Samimi, J. Van Dommelen, M. Geers, A self-adaptive finite element approach for simulation of mixed-mode delamination using cohesive zone models, Engineering Fracture Mechanics 78 (2011) 2202–2219.DOI
  21. [21]M. Samimi, J. Van Dommelen, M. Geers, A three-dimensional self-adaptive cohesive zone model for interfacial delamination, Computer Methods in Applied Mechanics and Engineering 200 (2011) 3540–3553.DOI
  22. [22]F. Van der Meer, N. Moës, L. J. Sluys, A level set model for delamination–modeling crack growth without cohesive zone or stress singularity, Engineering Fracture Mechanics 79 (2012) 191–212.DOI
  23. [23]X. Lu, B.-Y. Chen, V. B. Tan, T.-E. Tay, Adaptive floating node method for modelling cohesive fracture of composite materials, Engineering Fracture Mechanics 194 (2018) 240–261.DOI
  24. [24]B. Chen, S. Pinho, N. De Carvalho, P. Baiz, T. Tay, A floating node method for the modelling of discontinuities in composites, Engineering Fracture Mechanics 127 (2014) 104–134.DOI
  25. [25]D. Álvarez, B. Blackman, F. Guild, A. Kinloch, Mode I fracture in adhesively-bonded joints: A mesh-size independent modelling approach using cohesive elements, Engineering Fracture Mechanics 115 (2014) 73–95.DOI
  26. [26]S. Mukhopadhyay, S. Bhatia, Accurate coarse mesh simulation of delamination in composites using a novel hp-adaptive cohesive element, Journal of Composite Materials 57 (2023) 2201–2218.DOI
  27. [27]P. M. Daniel, J. Främbÿ, M. Fagerström, P. Maimí, An efficient err-cohesive method for the modelling of delamination propagation with large elements, Composites Part A: Applied Science and Manufacturing 167 (2023) 107423.DOI
  28. [28]G. Tosti Balducci, B. Chen, Overcoming the cohesive zone limit in the modelling of composites delamination with tuba cohesive elements, arXiv preprint arXiv:2403.09895 (2024).DOI
  29. [29]K. Bell, A refined triangular plate bending finite element, International journal for numerical methods in engineering 1 (1969) 101–122.DOI
  30. [30]D. Allman, A simple cubic displacement element for plate bending, International Journal for Numerical Methods in Engineering 10 (1976) 263–281.DOI
  31. [31]J. S. Przemieniecki, Theory of matrix structural analysis, Courier Corporation, 1985.
  32. [32]A. Turon, P. P. Camanho, J. Costa, C. Dávila, A damage model for the simulation of delamination in advanced composites under variable-mode loading, Mechanics of materials 38 (2006) 1072–1089.DOI
  33. [33]A. Turon, P. Camanho, J. Costa, J. Renart, Accurate simulation of delamination growth under mixed-mode loading using cohesive elements: Definition of interlaminar strengths and elastic stiffness, Composite structures 92 (2010) 1857–1864.DOI
  34. [34]G. Cowper, Gaussian quadrature formulas for triangles, International Journal for Numerical Methods in Engineering 7 (1973) 405–408.DOI
  35. [35]B. Davidson, R. Krüger, M. König, Three-dimensional analysis of center-delaminated unidirectional and multidirectional single-leg bending specimens, Composites Science and Technology 54 (1995) 385–394.DOI
  36. [36]R. Krueger, A summary of benchmark examples to assess the performance of quasi-static delamination propagation prediction capabilities in finite element codes, Journal of Composite Materials 49 (2015) 3297–3316.DOI
  37. [37]J. Williams, The fracture mechanics of delamination tests, The Journal of strain analysis for engineering design 24 (1989) 207–214.DOI
  38. [38]ASTM D5528 /D5528M-21, Standard test method for mode I interlaminar fracture toughness of unidirectional fiber-reinforced polymer matrix composites, ASTM-International, West Conshohocken, PA, 2007.
  39. [39]ASTM D6671/D6671M-22, Standard test method for mixed mode I -mode II interlaminar fracture toughness of unidirectional fiber reinforced polymer matrix composites, ASTM-International, West Conshohocken, PA, 2022.
  40. [40]M. Bruyneel, J.-P. Delsemme, P. Jetteur, F. Germain, Modeling inter-laminar failure in composite structures: illustration on an industrial case study, Applied Composite Materials 16 (2009) 149–162.DOI
  41. [41]R. Krueger, L. Deobald, H. Gu, A benchmark example for delamination growth predictions based on the single leg bending specimen under fatigue loading, Advanced Modeling and Simulation in Engineering Sciences 7 (2020) 1–25.

Paper details

vlnc:W_qbwxyuxf
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 8

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

Not recorded for this paper

View statement
Looking for related papers…
Contents