IDENTIFICATION OF CRIMS MODEL PARAMETERS FOR WARPAGE PREDICTION IN INJECTION MOULDING SIMULATION

Polymer injection moulding is a process widely used to produce components in a lot of different applications. One of the most critical aspects related to this process is to control the warpage of the parts after the extraction from the mould. Numerical simulations can predict a part warpage by using specific warpage models. Among numerical codes, Autodesk Moldflow Insight (R) uses a Corrected In Mold Residual Stress (CRIMS) model, that calculate the residual stresses develop during the moulding process. Warpage is then predicted calculating the deformations of the component induced by the considered stresses. Using experimental and numerical techniques, a new identification procedure was introduced to evaluate the six parameters of the CRIMS model included in the Moldflow (R) material properties database. The study was conducted on a box for an automotive application made of polypropylene. On the base of a complete rheological, thermal and physical characterization of the employed material, a numerical simulation of the process was implemented, integrating it with an optimization procedure to identify the values of the CRIMS parameters that force numerical results to fit measured deformations. As this procedure was very time consuming, requiring to run several computationally intensive simulations, artificial neural networks were employed to approximate numerical results with lower computational time. Results were verified with independent samples, showing good correspondence between experimental results and numerical calculated deformations.