Stochastic semi-analytical modeling of reinforced filaments for additive manufacturing

Abstract

This study presents a novel stochastic modeling approach that addresses two challenges encountered in micromechanical modeling of short-fiber composites. Firstly, the challenge of the time-consuming pre-processing required for extracting fibers from micro CT scans is tackled by introducing a new stochastic generation technique based on the kernel density estimation (KDE) method. This enables the generation of artificial fibers for micromechanical models, thus saving considerable time and effort. Secondly, the challenge of presenting a modeling approach that considers multiple fibers while reducing the computational effort associated with the simulation is addressed through a novel semi-analytical approach. To demonstrate the effectiveness of the stochastic modeling approach, it is applied to filaments of recycled poly(ethylene terephthalate) reinforced with recycled short carbon fibers that are used for additive manufacturing of composite parts. The results obtained from the stochastic modeling approach are compared with those from a direct modeling approach that considers 1050 fibers extracted from a micro CT scan. The novel approach is shown to provide similar predictions of elastic properties as the direct modeling approach while using only 40–50 fibers. Furthermore, the results are in close agreement with experimental data, highlighting the effectiveness of the proposed approach.