The article is devoted to the simulation of viscous incompressible turbulent fluid flow based on solving the Reynolds averaged Navier-Stokes (RANS) equations with different k − ω models. The isogeometrical approach is used for the discretization based on the Galerkin method. Primary goal of using isogeometric analysis is to be always geometrically exact, independent of the discretization, and to avoid a time-consuming generation of meshes of computational domains. For higher Reynolds numbers, we use stabilization SUPG technique in equations for k and ω. The solutions are compared with the standard benchmark example of turbulent flow over a backward facing step.
We deal with numerical solution of the incompressible Navier–Stokes equations discretized using the isogeometric analysis (IgA) approach. Similarly to finite elements, the discretization leads to sparse nonsymmetric saddle‐point linear systems. The IgA discretization basis has several specific properties different from standard FEM basis, most importantly a higher interelement continuity leading to denser matrices. We are interested in iterative solution of the resulting linear systems using a Krylov subspace method (GMRES) preconditioned with several state‐of‐the‐art block preconditioners. We compare the efficiency of the ideal versions of these preconditioners for three model problems (for both steady and unsteady flow in two and three dimensions) and investigate their properties with focus on the IgA specifics, that is, various degree and continuity of the discretization basis. Our experiments show that the block preconditioners can be successfully applied to the systems arising from high continuity IgA, moreover, that the high continuity can bring some benefits in this context. For example, some of the preconditioners, whose convergence is h‐dependent in the steady case, seem to be less sensitive to the mesh refinement for higher continuity discretizations. In the unsteady case, we generally get faster convergence for higher continuity than for C0 continuous discretizations of the same degree for most of the preconditioners.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.