Fiber orientation during injection molding of glass-fiber-reinforced thermoplastics: An experimental and numerical investigation.

Analytical and numerical tools are developed for predicting fiber orientation and flow during injection molding of glass-filled engineering thermoplastics. The theoretical analysis is based upon a realistic constitutive equation that accurately describes the rheology of the material. The experimental investigation includes rheological measurements to determine the material parameters for the constitutive equation, and fiber orientation determination based upon microtomed sections of tensile-bar samples. The computational research is based upon the marker-and-cell (MAC) technique with two-dimensional, finite-difference solutions to the continuity, momentum, and energy equations. The numerical work modifies a commercial MAC program, FLOW-3D from Flow Science Inc., to include a viscosity that is a function of temperature, pressure, and fiber concentration. The research is focused on materials with fiber concentrations much higher than previously reported. Rheological measurements of engineering thermoplastics with glass fibers are compared to theoretical predictions based upon the Dinh-Armstrong constitutive equation. The theory is expanded to test non-Newtonian materials at semi-concentrated and concentrated suspensions of polycarbonate, polypropylene, and nylon 6/6 resins with short and long glass fibers. The transient viscosity and transient normal stress predictions demonstrate good agreement to the experimental values for the total strain of 0.1 to 100 and for materials within the semi-concentrated fiber limit. The viscosity model is modified to account for shear-thinning viscosities of the polymer, based upon combining the Dinh-Armstrong relationship with the Carreau viscosity model. The new model predicts reasonably well the steady-state viscosity. The numerical analysis fairly accurately predicts the orientation and viscous stresses caused by glass fibers. The pressure drop comparison is good for slow fill, but not as good for fast fill of a Himont polypropylene. The pressure drop comparison is good for fast fill, but not as good for slow fill of glass-filled resin, DSM polypropylene with 10% and 20% short fibers. The numerical calculations represent the typical orientation state at the flow-front (most random), gate, and midstream (most aligned) locations. The comparison of the predicted fiber orientation to the measured fiber-orientation in injection-molded parts was fair for the 10% and 20% fiber-filled polypropylene at all three locations.PhDChemical EngineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/103597/1/9332073.pdfDescription of 9332073.pdf : Restricted to UM users only