Numerical simulation of viscoelastic fluid flow in stirred vessels — Method and validation

Although mixing of viscoelastic fluids in stirred vessels is common in fields such as polymer processing and biotechnology, the applicability of viscoelastic constitutive models to reproduce these processes is still unclear. Therefore, flow simulation results for a viscoelastic xanthan gum solution agitated by a Rushton turbine in a centrally stirred vessel are presented, validated against particle image velocimetry (PIV) measurements. The simulations are conducted with a custom OpenFOAM finite-volume solver based on the PIMPLE algorithm, incorporating both-sides diffusion (BSD) and sliding-mesh techniques. Multi-mode Giesekus and exponential Phan-Thien-Tanner (using both upper-convected and Gordon-Schowalter derivatives) models are used as constitutive models, parametrized from rheological measurements for steady and oscillatory shear. Secondary flow reversal and elastic compartmentalization, which specifically result from the elasticity of the fluid, are reproduced by all models. Quantitative agreement is found in fully elastic flows, as evidenced by velocity fields and circulation flow numbers. Significant deviations occur in compartmentalized flows, in which compartment shapes are reproduced incorrectly. Based on the different model results and experimental opposed-nozzle rheometry, the elongational rheology is identified as an important characteristic for the flow behavior in stirred vessels. Overall, the results demonstrate that viscoelastic CFD is a useful tool to understand and characterize mixing processes with viscoelastic fluids. • OpenFOAM implementation of sliding-mesh method for viscoelastic CFD in stirred vessels. • Use of multi-mode Giesekus and exponential Phan-Thien-Tanner models. • Validation against PIV data for a xanthan gum solution agitated by a Rushton turbine. • Reproduction of secondary flow reversal and flow compartmentalization. • Qualitative difference between numerical and experimental compartment shapes.