Impact of viscosity modeling on the simulation of aortic blood flow (Q6157921)

The authors study the modeling issues for the simulation of blood flow in an aortic coarctation and consider several viscosity models for non-Newtonian fluids as well as a Newtonian fluid model. They consider two different turbulence models in the simulations and compare the impact of both on clinically relevant biomarkers. The authors performed simulations with each of the viscosity models paired with both the \(\sigma\)-model and the RB-VMS model, and they used Taylor-Hood finite elements on the base mesh. The linear systems corresponding to each step of the Picard iteration were solved by FGMRES iteration, using a least-squares commutator preconditioner for the simulations using the \(\sigma\)-model and a hybrid FGMRES/BiCGSTAB approach for the RB-VMS computations. The authors conclude that, for time averaged quantities defined on various cross-sections of the aorta, the differences of the results between different viscosity models within the simulations with the same turbulence model were negligible. The instantaneous behavior, averaged over the periods, showed, in some cases, minor quantitative differences among the viscosity models, e.g., for the secondary flow degree, the normalized flow displacement, and the wall shear stress. However, the order of magnitude of these differences is not expected to affect diagnostic decisions.