Artificial boundaries and formulations for the incompressible Navier-Stokes equations: applications to air and blood flows (Q489965)

This article deals with numerical simulations of the air flow in the respiratory system. The physical model consists of the incompressible Navier-Stokes equations in truncated domains with suitable boundary conditions on the artificial boundaries. The author uses the standard formulation for the momentum equation of the Navier-Stokes system based on the basic convective form for the advection term. Mixed Dirichlet-Neumann boundary conditions on each corner of the computational domain lead to numerical difficulties which are investigated throughout this paper. The paper is organized in 4 sections. After the introduction with physical motivation, the author gives a theoretical overview where the basic notation is introduced and the tool model is described. Furthermore, the mathematical formulation of the tool problem is given with the corresponding weak formulation of the pressure drop problem. Moreover, the author refers to the existence and uniqueness results from the cited literature, which state that there exists a unique local-in-time solution for any data, and a unique global-in-time solution for small data. Using the basic form with natural boundary conditions implies the lack of energy conservation and, as a consequence, one deals with a restriction on the data. It is observed that if one uses the variational form which conserves the energy, existence theorems hold for less restrictive data. On the other side, changing the weak formulation also changes the associated boundary conditions and that can lead to the solutions which are not satisfactory from a physical point of view. Numerical tests are performed using the software package Felisce. When applying too high pressures, the author observes that the GMRES method treating nonlinearities does not converge and even leads to a blow up of the solution. Therefore, different stabilization methods, based on streamline diffusion or on direct handling of kinetic fluxes, are used and reviewed here in detail.