Enhancing extrusion die design efficiency through high-performance computing based optimization

被引：3

作者：

Vidal, J. P. O. ^{[1
,2
]}

Oliveira, M. ^{[2
]}

Srivastava, A. ^{[2
]}

Sacramento, A. ^{[1
]}

Aali, M. ^{[2
,3
]}

Guerrero, J. ^{[4
]}

Nobrega, J. M. ^{[2
]}

机构：

[1] Soprefa Componentes Industriais SA CJPlast, Zona Ind Mosteiro, P-4520909 Santa Maria Feira, Portugal

[2] Univ Minho, Inst Polymers & Composites, Campus Azurem, P-4804533 Guimaraes, Portugal

[3] Johannes Kepler Univ Linz, Inst Polymer Proc & Digital Transformat IPPD, Altenberger Str 69, A-4040 Linz, Austria

[4] Univ Genoa & Wolfdynam, I-16145 Genoa, Italy

来源：

MECCANICA | 2025年 / 60卷 / 06期

关键词：

Profile extrusion; Automatic optimization; Flow balance; HPC; OpenFOAM; Dakota; SHAPE OPTIMIZATION;

D O I：

10.1007/s11012-024-01919-7

中图分类号：

O3 [力学];

学科分类号：

08 ; 0801 ;

摘要：

This work presents a new computational framework for designing profile extrusion dies. The framework utilizes High-Performance Computing (HPC) resources to optimize parameterized die flow channels within a one-day time-frame, resulting in a significant reduction in the typical design time required for profile extrusion dies. By employing objective function-controlled convergence criteria, the framework achieved a 50% reduction in calculation time compared to runs where only the unknowns residuals were considered for the same purpose. Furthermore, it offers full optimization capability, requiring no user intervention once the CAD parameterization is complete. OpenFOAM and Dakota were employed for modeling and optimization, respectively. Fusion 360 and Onshape CAD software were used for drawing and parameterizing the flow channel. By leveraging HPC systems, the optimization framework can automatically test hundreds of alternative geometries within one day to find the optimal solution. This research demonstrates the feasibility and advantages of HPC-driven extrusion die optimization, which contributes to increased efficiency and competitiveness in the manufacturing industry.

引用

页码：1521 / 1532

页数：12