Finite element creep prediction of polymeric voided composites with 3D statistical-based equivalent microstructure reconstruction

Homogenization is very efficient tool in modeling of complex phenomena in heterogeneous media due to different length scales between the characteristic length of the microstructural features and the actual macroscale dimensions. This is particularly important when dealing with voided composite structures due to the high degree of microstructural irregularity. One of the challenges in designing polymeric porous or voided structures is the absence of accurate and robust viscoelastic homogenization technique that takes into account the microstructural details. Another challenge underlies in constructing the microstructure for numerical homogenization. This work proposes reduced 3D reconstruction procedure of voided composite structures based on statistical considerations and granular mechanics. Voids' diameters and fractions of the actual microstructure were extracted with a deterministic locally adaptive thresholding technique. The diameters were categorized with Freedman-Diaconis method but preserving the overall voids' fractions. The simulation box was then created from the reduced voids' diameters and fractions with granular mechanics. Numerical experiments were conducted with periodic boundary conditions to the walls of the simulation box. Voided structures of Acrylonitrile Butadiene Styrene (ABS) were fabricated with physical foaming agent and tested for creep compliance to validate the proposed procedure. The agreement with experimental results for creep compliance is very good with maximum error of 8.62%. The contribution of the procedure is attributed to the simplicity and accuracy in developing representative voided structures from SEM images which otherwise need to be extracted from tedious processes like X-ray microtomography reconstruction.