ABSTRACT
Up until this point, we have discussed continuous fluid fields like velocity represented and solved on Cartesian grids. Now we turn to a very different representation, particles: a collection of points with stored positions and attached values for fluid quantities of interest. Particles have long proven their worth in computer graphics as a flexible way to represent phenomena, including fluid-like effects. While numerically it is usually most favorable to solve the fluid equations on grids, in particular the pressure and incompressibility parts, including particles in a grid-centric fluid solver can still be extremely useful. (Later we will look at some schemes which solve everything with particles, which makes some different trade-offs worth considering.) Many secondary fields that we want to be carried around by the fluid, or that weakly effect the flow, can be very profitably represented with particles: smoke (soot concentration), foam, mist, bubbles, temperature, etc. We will begin by examining the big advantage particles have over grids for advection.
