Automatic versatile parametric procedure for a complete FEM structural analysis of composites having cylinder-shaped reinforcing fibres

Abstract

The aim of this work is the realisation of an automatic generalised procedure for creating a parametric easily manageable and fully automatic FEM model to predict the mechanical properties of composites having cylinder-shaped reinforcing fibres characterised by a complex morphology. A three-dimensional finite element RVE (Representative Volume Element) has been realised by means of an original modified RSA algorithm (Random Sequential Adsorption algorithm) which allows to reproduce, with high degree of approximation, the intrinsic geometrical and statistic characteristics of the examined composite, directly influenced by the particular real forming process and, therefore, the phenomena which characterise the mechanical response of the materials constituting the therein coexisting phases. The implemented procedure is completely automatic, because it requires an external operator barely for the insertion of a minimum number of geometrical information in Input. This insertion pertains to the geometrical information concerning the composite to simulate and to the information about the mechanical properties of the coexisting phases. In Output the procedure provides directly, through a single run, the whole stiffness matrix calculated with two different numerical homogenisation tools, using periodic boundary conditions. The procedure allows a minimisation of the RVE dimensions which give acceptable scatters, thanks to an original “isotropic criterion” based on the minimisation of a particular function. It was also introduced a method of calibration of those Input data which may be affected by geometrical constraints, introduced to face the basic assumptions of the algorithm. The first release of this algorithm has been already implemented, in a previous published work, see Cricrì et al. (2012). Such a procedure results potentially much more versatile than most of the existing models; it is facilely adaptable to investigations beyond the mechanical characterisation and covers other areas of interest related to composites.