This article provides the first multivariate analysis of reported thermal contact resistance (TCR) values from different characterization approaches (injection and compression molding) having missing data. The missing data are estimated by principal components analysis (PCA) with alternating least squares (ALS), and then the main effects are modeled using multiple regressions. Three models of increasing complexity are presented, each of which provides reasonable model fidelity relative to the TCR values reported in the literature. The primary cause of differences between reported TCR values as characterized by injection molding and compression molding can be explained by the melting of the polymer to make intimate contact with the metal mold during processing. Typical values of TCR are on the order of 0.0002 m2K/W, corresponding to heat transfer coefficients of 5000 W/(m2K). The analyses suggest that the TCR decreases significantly with increasing processing temperature and pressure but increases most significantly during cooling with the development of an air gap between the polymer and mold. Accordingly, it is strongly recommended that injection molding simulations incorporate modeling of the TCR as a function of the processing states for accurate prediction of the air gap, residual stresses, shrinkage, and warpage.Highlights
Multivariate analysis of thermal contact resistance data from the literature.
Sparse data handled by principal component analysis and alternating least square.
Comparison of injection molding and compression molding measurements.
Results suggest thermal contact resistance should be modeled as a function of the processing states.