“Galilean Invariance for Fluid Simulation” by Shah, Pighin and Cohen

  • ©Maurya Shah, Frederic (Fred) Pighin, Jonathan (Jon) D. Cohen, and Penne Lee

  • ©Maurya Shah, Frederic (Fred) Pighin, Jonathan (Jon) D. Cohen, and Penne Lee




    Galilean Invariance for Fluid Simulation



    These partial differential equations are solved on a grid of voxels. This grid is a discrete representation of three-space. Unfortunately fluid simulations are notoriously difficult to predict. In particular slight variations in the initial conditions can have dramatic impact in turbulent flows. This makes the problem of deciding a grid size for the simulation extremely difficult. With open boundaries, the fluid quickly dissipates when it hits the boundaries of the simulation domain (as in figure 1 (a)). To prevent this, large simulation domains have to be used to keep the fluid away from the boundaries of the simulation. In this work, we describe a novel technique for implementing an adaptive grid for fluid simulation. The technique is largely based on the principle of Galilean Invariance.


    Stam, J. 1999. Stable fluids. In Proceedings of SIGGRAPH 99, 181–188.


ACM Digital Library Publication:

Overview Page: