Presentation pdf 6.3MB

Influence of Turbulence Model on CFD Simulations of Flow in the FDA Benchmark Blood Pump Model

Computational fluid dynamics (CFD) can be used to influence decisions early in the design process. In order to assess the state-of-the-art of CFD and its predictive capability for medical devices, the U.S. Food and Drug Administration (FDA) developed two benchmark models for validation. Previous work explored hybrid and multi-block structured meshing strategies and their impact on steady-state solution accuracy for one of the benchmark cases, a centrifugal blood pump. The focus of this work was to extend the previous study by examining the influence of the turbulence model. The k-ω SST, standard k-ε and realizable k-ε turbulence models were used. Results from all six pump test conditions were compared for each turbulence model.

The meshes, created with Pointwise, consisted of a hybrid viscous unstructured mesh containing 7.9 M cells and a multi-block structured mesh containing 8.9 M cells. Caelus v9.04, a derivative of OpenFOAM, was used to perform the simulations. A moving reference frame (MRF) source term was used to model the rotating impeller to enable a steady-state approach. The parametric run capability of the Python-based companion library to Caelus was used to automate running and post-processing the simulations.

When compared, simulation results trended well with experiment. Turbulence model influences were consistent with other published work where the flow through the diffuser was found to be most sensitive to the turbulence model. The peak velocity and location of the peak velocity was found to be turbulence model dependent. The computed pump head and velocity profile along a diagonal line in the first quadrant of the impeller appear to exhibit less sensitivity.

A transient simulation was also carried out to examine the modeling error of the MRF approach.