An efficient surface algorithm for random-particle simulation of vorticity and heat transport

A new surface algorithm has been incorporated into the random-vortex method for the simulation of 2-dimensional laminar flow, in which vortex particles are deleted rather than reflected as they cross a solid surface. This involves a modification to the strength and random walk of newly created vortex particles. Computations of the early stages of symmetric, impulsively started flow around a circular cylinder for a wide range of Reynolds numbers demonstrate that the number of vortices required for convergence is substantially reduced. The method has been further extended to accommodate forced convective heat transfer where temperature particles are created at a surface to satisfy the condition of constant surface temperature. Vortex and temperature particles are handled together throughout each time step. For long runs, in which a steady state is reached, comparison is made with some time-averaged experimental heat transfer data for Reynolds numbers up to a few hundred. A Karman vortex street occurs at the higher Reynolds numbers. © 1989.