Fracture Behavior of Laminates in Mixed-Mode Delamination Resistance Testing

The purpose of this study is to investigate the factors contributing to the scatter of delamination resistance measurement results for composite laminates previously obtained at the Chair of Materials Science and Testing of Polymers at Montanuniversität Leoben focusing on the influence of data reduction methods and associated assumptions. Using established fracture mechanics methods, this research focuses on the analysis of fracture behavior under Mode I, Mode II, and mixed mode conditions. Five 3D finite element models have been developed to predict the crack behavior and to be validated against experimental results. These models include nonlinear geometry effects and use the Virtual Crack Closure Technique (VCCT) to calculate the critical energy release rate, which allows the study of mode transitions along the crack front with different mode mixities.
The delamination resistance of laminates with the same epoxy matrix but different fiber types - glass fiber (GF) and carbon fiber (CF) - is evaluated. Special attention is given to identifying the differences in fracture behavior resulting from the use of different fibers, aiming to understand how fiber type affects the fracture mechanics properties. In addition, discrepancies between experimental data and different data reduction methods are examined, which may stem from measurement inaccuracies or methodological variations.
This work provides a deeper understanding of the fracture behavior of composite laminates and improves the reliability of data reduction techniques, contributing to more accurate predictions of fracture mechanics properties in composite materials.
The combined experimental and VCCT-based simulations demonstrate that GFRP was more sensitive to anticlastic curvature than CFRP. The SBT reduction method best aligns with the FEM predictions because it does not account for additional parameters. The simulated force-displacement response of the CFRP was consistently stiffer than the experimental results across all fracture modes. This was likely due to the crack lengths being underestimated due to the precracks being visually indistinct. Despite accurate modelling, this led to lower calculated energy release rates.