Reduced Boundary Viscosity for Smoke

Alexis Angelidis


We propose a method to reduce viscosity at the boundaries of incompressible flows. Our method allows artists to control the amount of boundary viscosity by mixing a new slip boundary solution with an existing no-slip solution. The cost of our method is an additional pre-computation and a post-computation that can run in parallel for every frame. The running cost of the main solver is not changed, and arguably faster, since reducing boundary viscosity tends to reduce the domain occupied by the fluid. Our method can produce both a velocity field for time integration, or a geometric deformation to bypass time integration.

Paper (PDF)

Available as Pixar Technical Memo #19-02