Method Development for the Prediction of Melt Quality in the Extrusion Process

Area of Science:

• Polymer processing
• Materials science
• Computational modeling

Background:

• Simulation models are crucial for designing polymer extruders and extrusion screws, reducing the need for physical prototypes and shortening development cycles.
• Current simulation tools can predict temperature and pressure curves but cannot accurately determine the resulting melt quality.
• Melt quality is a critical factor in polymer product performance, and its prediction remains a challenge in the industry.

Purpose of the Study:

• To develop a novel simulation model capable of predicting melt quality in the polymer extrusion process.
• To establish a reliable method for assessing and improving melt quality based on simulation data.
• To leverage artificial intelligence, specifically machine learning, to bridge the gap between process simulation and melt quality prediction.

Main Methods:

• Utilized existing simulation data, including temperature and pressure curves, as input variables.
• Developed a machine learning model, specifically linear regression, to predict melt quality.
• Established the screw performance index as a key target value, correlating it with material and thermal homogeneity.

Main Results:

• A linear regression model was successfully built using simulation results as input.
• The model determined the correlation between extrusion process parameters and the final melt quality.
• The developed model's quality and predictive capabilities for melt quality were evaluated.

Conclusions:

• The study successfully demonstrates a simulation-based approach for predicting polymer melt quality.
• The integration of machine learning with extrusion process simulations offers a powerful tool for optimizing polymer processing.
• This model can significantly aid in the design of extrusion screws and the enhancement of overall melt quality in the polymer industry.