Krisztián György Hertelendy, Computational Mechanics and Materials Engineering

Fibre-reinforced polymers (FRPs) are lightweight materials with high specific strength and stiffness. These materials are increasingly substituting metallic alloys in automotive, aerospace, and civil engineering structures. In applications such as aircraft fuselage panels, automotive monocoques, wind turbine blades, hydrogen pressure vessels, they are often exposed to significant compressive loads. However, their compressive strength is typically lower than their tensile strength in the primary load-bearing direction, making accurate predictions of compressive failure essential for safe and efficient structural design.

FRPs are inherently hierarchical materials, with reinforcing fibres 5-25 μm in diameter in most cases, orders of magnitude smaller than the structural scale. Unlike most metals, FRPs are highly anisotropic and feature many load-dependent complex failure modes. Recent advances in experimental and imaging techniques have made it possible to examine the failure processes post mortem and in real-time, contributing to understanding the material behaviour and giving rise to new, physically based material models.

This thesis presents a modelling framework that connects different compressive/shear failure modes based on the underlying micromechanics.
Using the physical connection between these separate failure modes simplifies material parameter calibration. Adoption of a simplified micromechanical representation makes it possible to efficiently account for spatial variation in morphological features, such as fibre misalignment and inhomogeneous fibre volume fraction.

The model is calibrated for a carbon-fibre reinforced polymer using simple load cases from the Third World-Wide Failure Exercise, followed by the generation of biaxial failure envelopes. The predictions are compared with phenomenological and physically based stress-based failure criteria. Additionally, a compression test campaign on a different carbon-fibre reinforced polymer material is also considered, featuring a number of unidirectional and multidirectional laminates. After parameter calibration on the unidirectional laminates, the model predicted the compressive strength of six multidirectional laminates with an average deviation of 6.14% from the experimental means.