Simulation models are used to design extruders in the polymer processing industry. This eliminates the need for prototypes and reduces development time for extruders and, in particular, extrusion screws. These programs simulate, among other process parameters, the temperature and pressure curves in the extruder. At present, it is not possible to predict the resulting melt quality from these results. This paper presents a simulation model for predicting the melt quality in the extrusion process. Previous work has shown correlations between material and thermal homogeneity and the screw performance index. As a result, the screw performance index can be used as a target value for the model to be developed. The results of the simulations were used as input variables, and with the help of artificial intelligence-more precisely, machine learning-a linear regression model was built. Finally, the correlation between the process parameters and the melt quality was determined, and the quality of the model was evaluated.
Keywords: machine learning; melt quality; polymer extrusion process; screw performance index; simulation.