Accuracy evaluation of unsteady CFD numerical schemes by vortex preservation (Q2564675)

Evaluation of vortex preservation during convection is a reasonable tool to examinate the accuracy of numerical schemes in computer fluid dynamics: during convection of a vortex in freestream, pressure is at a minimum in the vortex core and increases concentrically and asymptotically to the freestream value with radial distance from the vortex center. But a numerically dissipative scheme cannot maintain a pressure minimum at the center. Similarly, the concentric structure of the vortex remains unaltered with downstream convection; however, certain numerical schemes distort the pattern, and solution oscillation appears. To explore the quality of some numerical schemes, the papers considers a vortex of finite core, see the classical description in the book by \textit{H. Lamb} [e.g. Hydrodynamics, 6th ed., Cambridge Univ. Press (1993; Zbl 0828.01012)], and follows up its path in travelling downstream over a distance of 45 radii of the core. The basis of the numerical method used herein is the Navier-Stokes code of \textit{W. R. Briley} and the second author [J. Comp. Phys. 24, 372-397 (1977; Zbl 0363.76018)]. These are a central difference scheme, first-order accurate in time, and an improved version of second-order accuracy in time and with multiple iterations at each time step, both for three difference grids. Effects of non-uniformity of grids or of cell aspect ratio are not considered. Altogether, such a study of physically relevant effects is always worthwhile.