6th European Solid Mechanics Conference

ESMC 2006
28 August – 1 September, 2006
Budapest, Hungary

A COMPUTATIONAL MULTISCALE MICROMODELLING FOR LAMINATED

COMPOSITES: BASIC ASPECTS AND VALIDATION.

G. Lubineau1 , P. Ladevèze1,2 and D. Violeau1

1
Laboratory of Mechanics and Technology (LMT-Cachan)
ENS Cachan / Paris 6 Univ. / C.N.R.S
61 Avenue du Président Wilson 94235 Cachan Cedex France
E-mail: ladeveze@lmt.ens-cachan.fr
2
EADS Foundation chair Advanced Computational Structural Mechanics

Keywords: damage, fracture, micromechanics, multiscale

Summary. A computational damage micromodel for laminates has been introduced recently at the LMT-Cachan.
That hybrid modelling, based on the one hand on the discrete fracture machanics, and on the other hand on contin-
uous damage mechanics is here detailled. Several examples are exposed to illustrate how such a micromechanical
based modelling can improve predictions of strong state of degradation.

1 INTRODUCTION

The last quarter century has witnessed considerable research efforts in the mechanics of composites in order to
understand their behavior and to model or calculate them - the ultimate goal being the design of the materi-
als/structures/manufacturing processes. Even in the case of stratified composites (which are the most studied and,
therefore, the best understood), the prediction of the evolution of damage up to and including final fracture remains
a major challenge in the modern mechanics of composite materials and structures.
Today, the use of stratified composites in the aerospace industry always involves characterization procedures
consisting of huge numbers of tests, which shows the low level of confidence in models. A significant improvement
in this situation, i.e. a drastic reduction in the number of industrial tests, could be achieved if one could create a real
synergy among the approaches on different scales which, today in the case of stratified composites, are followed
quite independently of one another. One could jokingly say that there is, on the one hand, the micromechanics
of laminates in which one counts cracks ([1], [2], [3], [4]) and, on the other, the meso- or macromechanics of
laminates in which one measures stiffnesses ([5], [6]) - with only few links between the two. How to bridge the
micro- and mesomechanics aspects was a first and important issue discussed in a previous work ([7]).
Establishing a synergy between these two approches is so proposed, through a computational micromodelling,
which is in fact an hybrid approach. That modelling allows the computation of the degradation evolution under
general loading. For a part (transverse microcracking and local delamination), it is discrete and based on the main
results of the numerous works in micromechanics. For another part, it is based on continuous damage mechanics,
to predict the evolution of diffuse mechanims which are usually forgotten in micromechanical analysis.

2 THE COMPUTATIONAL MICROMODELLING

The proposed computational damage micromodel is relatively simple, yet semi-discrete and probabilistic ([9]).
It depends on only few material constants. To a certain degree, it is discrete and introduces minimum cracking
surfaces. The failure criterion for each elementary cracking area is defined based on well-known micromechanics
considerations ([1]). For the remainder, diffuse damage and (visco)plasticity are described through continuum
mechanics models. One can say that the proposed micromodel belongs, at least partially, in the “finite fracture
mechanics” domain.
Unfortunately, when applied to engineering structures using current industrial codes, this micromodel leads
to prohibitive calculation costs. Therefore, the use of a multiscale computational strategy is absolutely essential.
Here, we expand the method with homogenization in space and in time introduced at Cachan, which does not use
any periodic property ([8]) and is suitable for parallel computing ([10]).
 P. Ladevèze, G. Lubineau, and D. Violeau
3 ILLUSTRATIONS AND EXAMPLES
After presenting the bases of the proposed computational micromodel, we show that this micromodel is capable of
replicating classical experimental micro- and macrophenomena observed for different stacking sequences, such as
the thickness effects or the difference between the initiation and propagation of intralaminar cracks (figure 1).
loading direction
transverse cracks

A B

Figure 1: Comparison of the cracking patterns (a) in thin laminates [06 /90] s and (b) in thick laminates [03 /903 ] s .

Several examples are also discussed concerning the prediction of the degradation of holed plates. Despite that
it is a crucial industrial topic, few models are able to predict correctly the strong level of degradation for such a
case. Indeed, the coupling between intra and interlaminar degradations is in that case the major source for damage,
and its modelling is still a challenge. Ii is however shown that such an approach allows local analysis of holed
plates and the prediction of such complex situations.

Figure 2: Degradation pattern around a hole.

REFERENCES
[1] G.J. Dvorak and N. Laws, Analysis of progressive matrix cracking in composite laminates. Part 2: first ply
failure, Journal of Composite Materials, 21, (1987), 309–329.
[2] Z. Hashin, Analysis of cracked laminates: a variational approach, Mech. Mat., 4, (1985), 121–126.

[3] J. Nairn and S. H. Hu, Matrix microcracking, Damage Mechanics of Composite Materials, Taljera R. (Ed),
(1994), 187–243.

[4] J.M. Berthelot, Transverse cracking and delamination in cross-ply glass-fibers and carbon-fibers reinforced
plastic laminates: static and fatigue loading, Appl. Mech. Rev., 56(1), (2003), 1–37.

[5] P. Ladevèze, On damage mechanics of composites, Comptes-rendus des JNC5, C. Bathias and D. Menks
(Eds), (1986), 667–683 (in French).
[6] P. Ladevèze, O. Allix, J.F. Deü and D. Lévèque, A mesomodel for localization and damage computation in
laminates, Comput. Methods Appl. Mech. Engrg., 183, (2000), 105–122.
[7] P. Ladevèze and G. Lubineau, On a damage mesomodel for laminates: micro-meso relations, possibilities
and limits, Composite Science and Technology, 61, (2000), 2149–2158.
[8] P. Ladevèze and A. Nouy, On a computational strategy with time and space homogenization for structural
mechanics, Comp. Meth. Appl. Mech. Eng., 192, (2003), 3061–3087.
[9] P. Ladevèze, A bridge between the micro- and mesomechanics of laminates: fantasy or reality?, Sectional
lecture ICTAM2004, (2004), to appear.
 6th European Solid Mechanics Conference
ESMC 2006
28 August – 1 September, 2006
[10] P. Ladevèze, O. Loiseau and D. Dureisseix A micro-macro and parallel computational strategy for Hungary
Budapest, highly
heterogeneous structures, International Journal for Numerical Methods in Engineering, 52 (1/2), 121–138