A hybrid micro-macro mechanical damage model to consider the influence of resin-rich zones on the transverse tensile behaviour of unidirectional composites

This paper presents a hierarchical hybrid micro-macro mechanical damage model aimed to simulate progressive failure in fibre reinforced composite materials. The hybrid model works based on performing image processing on the SEM cross-sectional images of microstructure to generate a location base file with the information of scattering of local volume fraction of the microstructure. The information is moved to a macro mechanical model which is implemented in Abaqus/Explicit using a user-defined material model. Prior to the macro mechanical analysis, the user-defined material model estimates the effective mechanical properties of each material point by using the local volume fraction and the analytical micromechanical models such as the rule of mixture, Chamis and Bridging. The hybrid damage model was verified by comparison with results from macroscale simulations with different fibre volume fractions. The hybrid model was used to analyze the resin-rich uncertainty in the composite material. The effect of the discretization window size was mitigated by using microstructure images and point-to-point mapping for the estimation of the mechanical properties. The presence of the resin-rich zone led to a 25.2 % decrease in the transverse stiffness and a 27.0 % increase in the failure strain which was well predicted by the hybrid model.

Automated summary

This paper presents a hierarchical hybrid micro-macro mechanical damage model to simulate progressive failure in fibre reinforced composite materials. The model combines image processing on SEM cross-sectional images of microstructure with a macro mechanical model in Abaqus/Explicit using a user-defined material model. The hybrid model was verified by comparing results from macroscale simulations and was used to analyze the resin-rich uncertainty in the composite material. The presence of the resin-rich zone was accurately predicted by the hybrid model.