An efficient approach to the modeling of compressive transverse cracking in composite laminates

Abstract

PREFACEAbout seven months ago, I was introduced to computational modeling of composite laminates. At that time, a poster that presented the work of Bert Sluys and Frans van der Meer caught my eye. I did not realize that this was going to be the basis of my master thesis! I'm very thankful that I've had to chance to work with Bert and Frans and that they have guided me through my final project. After a course in C++ and Jem Jive and a lot of hard work, the composite laminate failure model resulted that I describe in this thesis. To accompany it, I wrote a paper summarizing my findings, something I'm very excited about. As challenging as the computational modeling of composite laminates was, being able to explain what I spent my days doing to my friends and family was equally as difficult. I hope with this paper and my presentation I can make it clear to them.

Simulating the Clamped Tapered Beam Specimen under Quasi-Static and Fatigue Loading using Floating Node Method

“…In order to mitigate the effect of this approximation, the present study used two corrections. The first consists in determining the tractions in a rotated coordinate system, following [9,10]. In addition, since the matrix cracks were free to follow a mode I path as they propagate through-thickness, it was also considered that the shear component determined along direction 1 in Figure 4, if not zero, does not contribute to the mode-mixity.…”

“…However, as a consequence, Equation 4 may limit the ability of the approach to accurately capture damage initiation and small crack growth, since the actual material properties may not relate as given in Equation 4. In DCZ elements, the direction of the crack propagation, , is estimated and used to affect the elements constitutive response [5] or enhance an onset criterion [10]. In the present work, and following [10], the crack direction is assumed to be opposite to the gradient of the displacement jump determined within the element: (5) Contrarily to delamination modeling, where interface elements were pre-assigned along the interface between subelement 1 and 2, matrix cracks were not pre-defined.…”

“…In DCZ elements, the direction of the crack propagation, , is estimated and used to affect the elements constitutive response [5] or enhance an onset criterion [10]. In the present work, and following [10], the crack direction is assumed to be opposite to the gradient of the displacement jump determined within the element: (5) Contrarily to delamination modeling, where interface elements were pre-assigned along the interface between subelement 1 and 2, matrix cracks were not pre-defined. Under quasi-static loading conditions, matrix crack onset and formation were modeled by splitting the sub-elements along the fiber direction and inserting DCZ elements that open according to their traction-separation formulation.…”

“…In Fatigue, similarly to quasi-static, both delaminations and matrix cracks onset were determined via a Benzeggagh-Kenane (BK) criterion [11] written in stress space. Fatigue crack onset (delamination and matrix crack) is assumed to be approximated by an onset curve with the form [10]:…”

Simulating Matrix Crack and Delamination Interaction in a Clamped Tapered Beam

“…In order to attenuate the effect of this approximation, the present study uses two corrections. The first consists in determining the cohesive tractions in a rotated coordinate system, following [6,7]. In addition, since the matrix-cracks are free to seek a mode I path as they propagate through-thickness, the shear component determined along direction 1 in Figure 3 is ignored, if not zero, and not considered to contribute to the mode-mixity.…”