The need for reliable approaches in numerical simulations stands out as a critical issue for the development and optimization of cardiovascular biomedical devices. This led the US Food and Drug Administration to undertake a programme of validation of computational fluid dynamics methods for transitional and turbulent flows. In the current investigation, large-eddy simulation is used to simulate the flow in the first benchmark medical device, and results are confronted to the existing laboratory experiments. This idealized medical device has the particularity to feature transition to turbulence after a sudden expansion. The effects of numerical parameters and low-level inlet perturbations are investigated. Results indicate a considerable impact of numerical aspects on the prediction of the location of the transition to turbulence. The study also demonstrates that injecting small perturbations at the inflow greatly improves the streamwise velocity estimation in the transition region and substantially contributes to the robustness of the flow statistical data. Copyright © 2016 John Wiley & Sons, Ltd.

International audienceTransverse secondary gas injection into the supersonic flow of an axisymmetric convergent-divergent nozzle is investigated to describe the effects of the fluidic thrust vectoring within the framework of a small satellite launcher. Cold-flow dry-air experiments are performed in a supersonic wind tunnel using two identical supersonic conical nozzles with the different transverse injection port positions. The complex three-dimensional flow field generated by the supersonic cross-flows in these test nozzles was examined. Valuable experimental data were confronted and compared with the results obtained from the numerical simulations. Different nozzle models are numerically simulated under experimental conditions and then further investigated to determine which parameters significantly affect thrust vectoring. Effects which characterize the nozzle and thrust vectoring performances are established. The results indicate that with moderate secondary to primary mass flow rate ratios, ranging around 5 %, it is possible to achieve pertinent vector side forces. It is also revealed that injector positioning and geometry have a strong effect on the shock vector control system and nozzle performances

A 4th-order accurate, low dissipative flow solver is used to perform Large-Eddy Simulations of three typical hemodynamic situations: the flow through the idealized medical device proposed by the American Food and Drug Administration; the intracardiac flow within an actual human left heart whose morphology and deformations are deduced from medical imaging; the flow downstream of an artificial aortic valve which arises from the blood-leaflets interaction problem. In all the cases, the subgrid scale model designed to handle wall-bounded transitional flows is successfully used and the numerical simulations compare favourably with the experimental data available. These results illustrate the potential of the Large-Eddy Simulation methodology to properly handle blood flows. They also support the idea that turbulence, even if not fully developed, may be present in cardiovascular flows, including under non pathological conditions.