SUMMARY
Researchers at UCLA led by Joseph Teran have developed a new technique that can better simulate the motion of particles in fluid.
BACKGROUND
Simulation techniques of natural phenomena in computer graphics have uses in the areas of special effects, interactive gaming, and engineering simulations. These simulations must account for a wide range of geometric spaces and material behaviors. The latest simulation technique, Particle-in-cell (PIC), models particles and a 3D grid in which they move inside. The particles in PIC store basic information like mass and momentum while the grid points store the movement effects. To create the next frame of movement, the grid points overwrite the momentum stored in the particles, and this exchange of information happens many times to simulate movement of fluid. However, the number of grid points is less than the number of particles modeled, and when the exchange between grid and particles happens, the number of particles decrease due to this mismatch in number. Several techniques have been developed to ameliorate this “dissipation” effect, but this causes unwanted effects like noise and clumping.
INNOVATION
Researchers at UCLA led by Joseph Teran have developed a new technique that can better simulate the motion of particles in fluid. Previous methods suffered from a “dissipation” effect due to the mismatch in resolution of a 3D grid compared to the number of particles being simulated, and other methods to solve this problem result in noise and clumping effects. This new method, PolyPic, improves upon how velocity is modeled in a 3D space, which leads to improved kinetic energy conservation in particle/grid transfers. This improvement results in simulation of fluid flow that looks more realistic than the previous techniques (PIC and FLIP). An example of this can be seen below.
Figure: Simulated droplet using PolyPIC compared to competitors
POTENTIAL APPLICATIONS
ADVANTAGES