Purpose: This study aims to achieve a “pseudo-ductile” behaviour in the response of hybrid fibre reinforced composites under uniaxial traction by solving properly formulated optimization problems. Design/methodology/approach: The composite material model is based on the combination of different types of fibres (with different failure strains or strengths) embedded in a polymer matrix. The composite failure under tensile load is predicted by analytical models. An optimization problem formulation is proposed and a Genetic Algorithm is used. Multi-objective optimization problems balancing failure strength and ductility criteria are solved providing optimal mixtures of fibres whose properties may come either from a pre-defined list of materials, currently available in the market, or simply assuming their continuum variation within predefined bounds, in an attempt to attain unprecedented performance levels. Findings: Optimal solutions of hybrid fibre reinforced composites exhibiting pseudo-ductile behaviour are presented. It is found that a fibre made from a material exhibiting relatively low stiffness combined with high strength is preferred for hybridization. Furthermore, the ratio of the average failure/critical strains between the low and high elongation fibres to be hybridized must be equal or greater than two. Originality/value: Typically, a ductile failure is an inherent property of metals, that is, their typical response curve after the linear (elastic) region exhibits a yielding plateau still followed by an increase in stress till collapse. In stark contrast, composite materials exhibit (under some loading conditions) brittle failure that may limit their widespread usage. Therefore, a “pseudo-ductility” in composites is valued and targeted through optimization which is the main original contribution here.
|Number of pages||28|
|Journal||Engineering Computations (Swansea, Wales)|
|Publication status||Published - 1 Jan 2018|
|Event||ECCOMAS Thematic Conference on Computational Modelling of Multi-Uncertainty and Multi-Scale Problems - Porto, Portugal|
Duration: 12 Sep 2017 → 14 Sep 2017
- Genetic algorithm