Perpustakaan Digital ITB

This research developed a multiscale numerical modeling of glass fiber reinforced thermoplastic composite material. The modeling process started from micro- to meso-scale level. The microscale level was done by modeling a hexagonal unit cell under 4 loading conditions, i.e. uniaxial tension, transverse tension, transversal shear, and longitudinal shear, to get a complete elastic and damage properties of the unit cell in all modes. The thermoplastic matrix was modeled using Drucker-Prager plasticity and Ductile Damage model while the fiber-matrix interface was modeled using cohesive element. The effect of fiber volume fraction variation to the damage modes and the mechanical properties was also studied. The mesoscale model was built of a n×n homogenized unit cells. The unit cell properties with different fiber volume fractions were distributed randomly across the mesoscale model with the expectation to represent the fiber nonuniformity in the actual composite material. It was proved that the proposed mesoscale model could predict the composite transverse tensile strength with error of 1% compared to the experiment result. The homogenized mesoscale model also had a lower computational cost compared to another reference model that model the heterogeneity of fibers and matrix. It was expected that this simplified method can be used to reduce the number of experiment to estimate the mechanical properties of heterogeneous material.