Fiber Orientation Distribution Predictions for an Injection Molded Venturi-Shaped Part Validated Against Experimental Micro-Computed Tomography Characterization

Abstract

This work evaluates and compares the accuracy of different fiber orientation prediction models for a short fiber reinforced injection molded Venturi-shaped part which displays variable thickness. The experimental characterization of the specimen fiber orientation distribution (FOD) was carried out by the micro computed tomography technique (micro-CT). The computational study of fiber orientation predictions was performed using Moldex3D. All the possible combinations of the Folgar-Tucker (FT) and improved Anisotropic Rotary Diffusion (iARD) rotary diffusion models and the Hybrid (Hyb), Orthotropic (ORE), and Invariant Based Optimal Fitting (IBOF) closure approximations were considered. The relevance of the Retardant Principal Rate (RPR) model on predictions results was also evaluated. The values of the fiber-fiber (Ci), matrix-fiber (Cm) interaction coefficients and the alpha-RPR parameter were varied in a significant range in order to find the set of parameters that better fits the experimental fiber orientation data. The parameters' sensitivity effect over the second order orientation tensor components was quantified via the Analysis of Variance (ANOVA) statistical method. The experimental micro-CT results show an increase in the fiber orientation degree at the specimen constriction region due to the narrowed cavity and the Venturi effect. The comparison of the experimental and predicted orientation profiles demonstrates that the predictions of the iARD model, in combination with the IBOF closure approximation, are the most accurate for the case studied. However, simulations fail to estimate the change in orientation caused by variable thickness and section. ANOVA results prove that the orientation tensor component in the flow direction (a11) is more sensitive to changes in alpha-RPR and Ci coefficient, while the perpendicular components (a22, a33) are also significantly affected by Cm. From the predictive error analysis it is seen that the optimal parameters set to capture the orientation state of the specimen is: (i) for the FT model, Ci = 0.005, alpha-RPR = 0.7 and (ii) for the iARD model, Ci = 0.005, Cm = 0.2, and alpha-RPR = 0.7.