ASSESSMENT ON THE CAPABILITIES OF ABAQUS AND LS-DYNA TO PREDICT THE BEHAVIOR OF LAMINATED COMPOSITES SUBJECTED TO LOW-VELOCITY IMPACTS

Abstract

Composite materials have excellent in-plane mechanical properties. However, they have a low resistance to impact damage. Low-velocity impacts (LVI) produce barely visible impact damage (BVID) on the surface of the laminate but with the potential to produce significant internal damage, such as delamination and matrix cracking. Composite materials' impact damage evaluation still largely relies on experimental results rather than numerical simulations because of the material's multiple damage mechanisms. In an attempt to mitigate cost and time requirements for large-scale experimental studies, there have been significant strides toward enhancing computational simulations to incorporate and predict damage mechanisms observed in composite structures. However, current computational models have limitations in how they capture the impact response and damage spread in laminated composites. The present study compares the predictive capabilities of two commercially available software, ABAQUS and LS-Dyna, to assess their feasibility in capturing the impact response and damage spread on laminated composites. The computational responses will also be compared with the responses obtained experimentally. To the authors' best knowledge, there has not been any report in the literature that compares the predicting capabilities of these two softwares. IM7/977-3 graphite epoxy unidirectional laminates with a 32-ply layup [-45/0/45/90]4S were manufactured according to the ASTM D7136/D7136M-12. Drop weight impact tests were performed in an Instron CEAST 9350. The samples were subjected to LVI energies of 30 J. From each test, the contact force, displacement, velocity, energy and impact duration time were recorded to compare with the predicted responses from the computational models. To evaluate the internal damage area, nondestructive inspection (NDI) was performed on all the samples with X-ray. The objective is to compare the internal damage per layer of each experiment with the internal damage obtained from the ABAQUS and LS-Dyna computational models. For the ABAQUS model, the intralaminar damage (ply failure) model consisted of a continuum damage model, Hashin failure criterion, and a damage evolution model based on equivalent displacement. The interlaminar damage (delamination) was incorporated through a cohesive surface interaction with a bilinear traction-separation law. In the LS-Dyna model, MAT261 Laminated Fracture Daimler-Pinho material card was used as the intralaminar damage model. MAT261 is a continuum damage model with linear softening evolution based on fracture toughness. The interlaminar damage was incorporated through a Tiebreak contact algorithm with a bilinear traction-separation behavior. Preliminary studies have shown that the ABAQUS/Explicit model showed a good correlation with the experimental results in terms of contact force, impact duration time, and displacement. On the other hand, the LS-Dyna MAT261 material model underpredicted the contact force, impact duration time, and displacement. The extreme differences in the LS-Dyna simulation are attributed to MAT261’s algorithm requiring element deletion during the simulation to maintain stability. Therefore, the striker never rebounded and continued penetrating the laminate. Further impact energies will be explored with the ABAQUS computational model. For LS-Dyna, different venues need to be explored in calibration for MAT261’s material model to account for the element deletion’s energy loss or use a different material model.