Analysis and design optimization of distributed parameter systems--An application to reaction injection molding.

摘要

Reaction Injection Molding (RIM) is a flexible and energy efficient polymer processing technique used in the manufacture of a variety of industrial and consumer products. The processing requirements, material characteristics, and performance criteria associated with each of these applications vary substantially. Analysis and control of processing and material properties in RIM are crucial design questions which to date have not been clearly addressed.;Reverse engineering applied to reaction injection molding provides a straightforward approach to process analysis and materials design. This approach involves accurate modeling of process behavior, analysis of the effects of design parameter variations on the system, and the characterization and optimization of RIM molded materials. Multidimensional, finite element models have been developed for the filling and curing stages of reaction injection molding. The accuracy of these formulations was verified through comparison of the predicted profiles to experimental results for a catalyzed polyurethane system. Parameter sensitivity analyses identified the principal design variables as well as the relative magnitude of their influence on processing. RIM material characterization required measuring the physical properties of a series of isothermally molded urethane films and developing expressions relating these properties to the processing temperature. The resulting empirical expressions were useful in property estimation based on the thermal behavior predicted by the models. Finally, a framework has been developed coupling the distributed parameter models to an optimization algorithm. Applied to representative moldability and materials design problems, this linked modeling-optimization procedure was found to be a flexible, practical approach to reaction injection molding process optimization.