Optimal vorticity accuracy in an efficient velocity-vorticity method for the 2D Navier-Stokes equations (Q1742970)

A 2D velocity-vorticity (VV) scheme is analyzed and tested for incompressible Navier-Stokes simulations. The scheme is developed by first coupling the velocity and vorticity equations through the nonlinear terms, then decoupling the system at each timestep through a linearization of the momentum equation. This allows that vorticity equation and mass-momentum system are solved simultaneously at each time step, and thus the extra cost of the VV scheme can be removed. Convergence analyses are performed for backward Euler and backward differentiation formula (BDF2) discretizations together with finite element spatial discretizations, which deliver optimal accuracy of both velocity and vorticity. Three numerical tests are performed: the first shows the predicted convergence rates hold on a test problem with a known analytical solution, the second is for the Gresho standing vortex problem [\textit{P. M. Gresho}, Int. J. Numer. Methods Fluids 11, No. 5, 587--620 (1990; Zbl 0712.76035)], and the third is for channel flow past a flat plate. The VV solutions are compared with standard velocity-pressure results for the Gresho vortex and channel flow tests. It was shown that in both tests, VV velocity and vorticity are more accurate.