Validated Guidelines for Simulating Centrifugal Blood Pumps

Abstract

Rotary blood pumps (RBPs) employed as ventricular assist devices are developed to support the ventricles of patients suffering from heart failure. Computational Fluid Dynamics (CFD) is frequently used to predict the performance and haemocompatibility of these pumps during development, however different simulation techniques employed by various research groups result in inconsistent predictions. This inconsistency is further compounded by the lack of standardised model validation, thus it is difficult to determine which simulation techniques are accurate. To address these problems, the US Food and Drug Administration (FDA) proposed a simplified centrifugal RBP benchmark model. The aim of this paper was to determine simulation settings capable of producing accurate predictions using the published FDA results for validation. This paper considers several studies to investigate the impact of simulation options on the prediction of pressure and flow velocities. These included evaluation of the mesh density and interface position through steady simulations as well as time step size and turbulence models (k-ε realizable, k-ω SST, k-ω SST Intermittency, RSM ω-based, SAS and SBES) using a sliding mesh approach. The most accurate steady simulation using the k-ω turbulence model predicted the pressure to within 5% of experimental results, however experienced issues with unphysical velocity fields. A more computationally expensive transient simulation that used the Stress-Blended Eddy Simulation (SBES) turbulence model provided a more accurate prediction of the velocity field and pressure rise to within experimental variation. The findings of the study strongly suggest that SBES can be used to better predict RBP performance in the early development phase.