Abo Bibliothek: Guest

ISBN: 978-1-56700-537-0

ISBN Online: 978-1-56700-538-7

ISSN Online: 2377-424X

International Heat Transfer Conference 17
August, 14-18, 2023, Cape Town, South Africa

Stability analysis of coupled Navier-Stokes solver in transient problems of incompressible flow

Get access (open in a dialog) DOI: 10.1615/IHTC17.50-80
11 pages

Abstrakt

Coupled solver is widely used because of its high converging efficiency and robustness. The solver solves momentum equation and continuity equation simultaneously. Generally, it is assumed that the algorithm only affects the convergence of the calculation rather than the correctness. However, we found the couple solver result in no vortex street after the cylinder when the flow passing through it. It is not consistent with the result form segregated algorithm (SIMPLE). In order to find the reason behind, the stability of coupled solver for transient problems of incompressible flow is studied. The coupled solver and SIMPLE algorithm are used to calculate the lid-driven cavity flow when Reynolds numbers are 200 and 1000 respectively. It is highly consistency for the two algorithms in solving both transient and steady-state lid-driven cavity flow while the coupled solver has even 1/10 number of converged steps than that of SIMPLE algorithm. Then we simulated the flow around a cylinder within the range of Reynolds number 40-1000, providing an asymmetric disturbance for the velocity field at the initialization stage. Different boundary conditions are used for reliability consideration. The results are compared with ANSYS Fluent CFD platform. The results show that the flow field solved by the coupled solver cannot form Karman vortex street, and tend to be a steady-state result at low Reynolds number. On the contrary, the flow field solved by SIMPLE algorithm shows an obvious Karman vortex street at the same Reynolds number after a certain time step. In conclusion, the solution stability of coupled solver is significantly higher than that of SIMPLE algorithm. Hence, we guess the different vortex shapes between the two algorithms may be caused by the existence of both stable and unstable solutions for the flow around a cylinder which is very sensitive to disturbances. The coupled solver might be resisted to such disturbance due to its tight coupling of velocity and pressure and it also has the characteristic of fully implicit while segregated algorithm has not.