Dataset for Designing extrusion dies on the basis of eXplainable Artificial Intelligence

Published: 4 July 2022| Version 1 | DOI: 10.17632/dy97xr6t8h.1
Juan Marcos Llorca


This dataset contains the porthole die data used to develop a design support tool for aluminium extrusion porthole dies based on eXplainable IA. The tool is useful for dies with 4 cavities and 4 ports per cavity. Dataset includes the geometrical data related to 596 different ports from 88 first-trial 4 cavities and 4 ports per cavity dies. In terms of the R² metric and the results obtained with the application examples, the results obtained with this ML-based model are significantly better than those of a previous model based on linear regression. It also includes the results and geometries of the FEM simulations performed to validate the model.



Universitat d'Alacant


Experimental Design, Machine Learning, Die Design, Profiles Extrusion, Predictive Modeling