Three-dimensional BEM analysis to assess delamination cracks between two transversely isotropic materials

N.o. Larrosa, Jhonny E. Ortiz, Adrián P. CisilinoJournal of Mechanics of Materials and Structures 2011 · DOI

Abstract

Beyond the inherent attribute of reducing the dimensionality of the problem, the attraction of the boundary element method (BEM) for dealing with fracture mechanic problems is its accuracy in solving strong geometrical discontinuities. Within this context, a three-dimensional implementation of the energy domain integral (EDI) for the analysis of interface cracks in transversely isotropic bimaterials is presented in this paper. The EDI allows extending the two-dimensional J-integral to three dimensions by means of a domain representation naturally compatible with the BEM, in which the required stresses, strains, and derivatives of displacements are evaluated using their appropriate boundary integral equations. To this end, the BEM implementation uses a set of recently introduced fundamental solutions for transversely isotropic materials. Several examples are solved in order to demonstrate the efficiency and accuracy of the implementation for solving straight and curved crack-front problems.

Introduction: The Challenge of Delamination

Composite laminates are essential in modern aerospace and automotive engineering due to their exceptional strength-to-weight ratios. However, they are uniquely vulnerable to delamination—the formation and growth of cracks between layers (plies). These materials are "transversely isotropic," meaning their mechanical properties differ significantly along the fiber axis compared to the plane perpendicular to it. Modeling these cracks is computationally demanding because real-world delamination is a three-dimensional phenomenon influenced by complex boundary layer effects where the crack meets a free surface.

Traditional Finite Element Method (FEM) approaches often struggle with these problems, requiring extremely dense 3D meshes around the crack tip to achieve acceptable accuracy. This paper presents a more efficient alternative: a 3D numerical framework combining the Boundary Element Method (BEM) with a 3D Energy Domain Integral (EDI).

Methodology

The Boundary Element Method (BEM) Advantage

The researchers utilized BEM because it only requires the discretization (meshing) of surfaces rather than the entire volume. This dimensionality reduction is particularly advantageous for fracture mechanics, where high-precision data is needed at the crack interface.

Energy Domain Integral (EDI)

The core of the analysis is the EDI, a 3D extension of the classic $J$-integral. It calculates the energy release rate at any point along a crack front. Unlike surface-based calculations, the EDI is "path-independent," allowing researchers to calculate energy over a volume surrounding the crack tip. This volume-based approach smooths out local numerical errors and provides a more stable result.

Schematic of the 3D Energy Domain Integral (EDI) formulation showing the integration volume around a crack front.

Multidomain Formulation and Fundamental Solutions

To model the interface between two different materials, the BEM model is split into subdomains that share a common interface. This allows for different fiber orientations on either side of the crack. The implementation utilizes specialized fundamental solutions for transversely isotropic materials that rely on real functions, avoiding the computational complexity of complex-number formulations.

Multidomain BEM model representing an interface crack between two distinct material regions.

Validation and Key Results

The framework was validated using three rigorous benchmark problems: a thick bimaterial plate under tension, an edge interface crack in a laminate specimen, and a cylindrical bimaterial bar with a circumferential crack.

Key findings include:

  • High Accuracy: BEM results matched high-resolution FEM (ABAQUS) models within a 2% margin for homogeneous test cases.
  • Numerical Stability: The EDI demonstrated excellent path independence, with a standard deviation of only ~5% for integration domains with a normalized radius $r/a \ge 0.20$.
  • Orientation Sensitivity: The study confirmed that the fracture energy distribution becomes highly asymmetric if the principal axes of the materials are not aligned or perpendicular to the crack.
  • 3D Resolution: The model successfully captured periodic variations of fracture energy along curved crack fronts, a detail that 2D simplifications cannot replicate.

Limitations and Future Outlook

While robust, the method faces challenges at the exact point where a crack front intersects a free surface (the "boundary layer"), where the EDI loses strict path independence. Additionally, very small integration domains ($r/a = 0.1$) can suffer from discretization errors if not sufficiently refined.

Future work will focus on incorporating "contact mechanics" to account for crack faces touching at the tips and exploring improved auxiliary functions to stabilize numerical results at free surfaces. This research provides a critical tool for engineers to predict the structural integrity of advanced composite materials more accurately and efficiently.

Summary updated

Abstract

Beyond the inherent attribute of reducing the dimensionality of the problem, the attraction of the boundary element method (BEM) for dealing with fracture mechanic problems is its accuracy in solving strong geometrical discontinuities. Within this context, a three-dimensional implementation of the energy domain integral (EDI) for the analysis of interface cracks in transversely isotropic bimaterials is presented in this paper. The EDI allows extending the two-dimensional J -integral to three dimensions by means of a domain representation naturally compatible with the BEM, in which the required stresses, strains, and derivatives of displacements are evaluated using their appropriate boundary integral equations. To this end, the BEM implementation uses a set of recently introduced fundamental solutions for transversely isotropic materials. Several examples are solved in order to demonstrate the efficiency and accuracy of the implementation for solving straight and curved crack-front problems.
BibTeX
@article{larrosa2011three,
  author = {N.O. Larrosa and Jhonny E. Ortiz and Adrián P. Cisilino},
  title = {{Three-dimensional BEM analysis to assess delamination cracks between two transversely isotropic materials}},
  journal = {Journal of Mechanics of Materials and Structures},
  year = {2011},
  doi = {10.2140/jomms.2011.6.1103},
}
Figure