Masters Theses

Runjun Das

Date of Award

12-1991

Degree Type

Thesis

Degree Name

Master of Science

Major

Polymer Engineering

Major Professor

Donald C. Bogue

Committee Members

Edward S. Clark, John E. Fellers

Abstract

Several computer programs for dealing with the mold filling step in injection molding are now available commercially. In the present work two such programs (C-FLOW and Mold Filling) have been run to study their accuracy in various cases and to prepare case studies for classroom use. The programs were compared with each other and with experimental data (or experimental observations) from several sources. Three geometries were studied: (1) a flat rectangular mold with internal obstructions, a geometry used by the Cornell injection molding group; (2) a simulation of a 24" TV cabinet, a practical geometry of considerable complexity (2545 nodes, 4635 elements), provided to us by the Philips Consumer Electronics Company; and (3) a simplified version of the Philips geometry (1366 node 2399 elements), which maintains the overall shape of the complete simulation but omits many small details (grooves, slots, etc.). Both the C-FLOW and Mold Filling programs were run for the flat rectangular mold, giving results in reasonable agreement with each and with experimental data from the literature (pressure vs. time), excepting only the the pressure tap nearest the end of the mold gave predicted pressures quite a bit more than the measured ones. The full Philips simulation of the TV cabinet and a simplified version of it were run on C-FLOW, the polymer being a high impact polystyrene for which Philips had rheological data. Predictions for inlet pressure vs. time were in reasonable agreement between the two simulations. The final inlet pressures from the simulations were 39.MPa for the simplified model and 30 MPa for the Philips' model which are in reasonable agreement with in-plant (production) values of 30 - 40 MPa. The full simulation gives a spurious pressure "spike" at the end of the fill because of numerical problems associated with many small elements and some subjective judgement is necessary to determine the true value. The computer program predicted "short shots" at about 65% of full-fill pressure, compared with plant experience of about 80 - 85% of full-fill pressure, reasonable agreement in view of the un quantified nature of various kinds of short shots. The character of "weld lines" (junction lines between two flowing polymer fronts) was studied qualitatively. SEM photographs showed that bad (visible) weld lines showed fibrous structure when broken apart whereas near-invisible weld lines or samples free of weld lines broke with faceted, smoothly cleaved surfaces. No large structural differences across the thin dimension of the weld line were observed, which tends to discredit a criterion (for good or bad weld lines) based on temperature as a primary variable, temperature varying significantly as one moves from wall to centerline. In accord with suggestions from the program vendor the quality of the weld line did correlate reasonably well with the angle at which the two melt fronts meet. This cannot be the only variable in practical molding operations, however, and further studies of this kind are needed. The geometries produced and stored in computer memory cover a variety of practical shapes and will provide a useful "bank" to draw upon in future classroom projects involving C-FLOW.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS