CFD-Based Optimization Methods Applied to Polymer Die Design

Laura Dietsche, Patrick Lee and Joe Dooley (The Dow Chemical Company)

CHALLENGE - Researchers explored the potential of CFD associated to RSM and optimization techniques in the design processes of the chemical industry. Two polymer die design examples were explored using different slot die geometries, different CFD codes, and different optimization packages and methods.

SOLUTION - In both cases, various geometric parameters were modified in order to meet the two competing objectives of uniform flow at the die exit while minimizing or limiting the pressure drop. In some trials, CFD was used to create RSMs and the optimization algorithms were directly coupled to the CFD.

Scatter plot of the Objective Parameters from  both the Single- and Multi-Objective Function Runs










BENEFITS - modeFRONTIER offered a Uniform Latin Hypercube space filling design as well as numerous algorithms for creating RSMs. Additionally, researchers utilized the NSGA-II (Non-dominated Sorting Genetic Algorithm) for both the single and multi-objective optimizations.