Physical Model Based Simulation Of Smoke

Fluid simulation is the focus of research in the computer graphics field. Smoke which is an important component of fluid, the simulation received more and more attention. Generally, the simulation of smoke animation have two ways:one method based on Particle systems, and another method based on physical models. The simulation based on particle system meet the real-time simulation when they under some certain conditions. The main idea of physics models based simulation is reflect the smoke fumes which based the movement equation, then constructing the movement model of smoke, and choosing the appropriate numerical method to solve the model equations. The solution of equation can be obtained the motion state of each moment. Last, combination of graphical drawing tool can draw the simulation results of smoke, the effect is more consistent with the real motion of smoke.Using the method which based on physical model to simulation the smoke has three steps:The first is to determine the physical model of smoke. In this paper, the non viscous and incompressible Navier-Stokes equations are used as physical model of simulation smoke, and add the density equation which changed with the velocity fields. And assuming the value of the viscosity term is zero, which in the incompressible equation, this can simplify the computer calculated. To reduce the numerical dissipation, in this paper using the vortex operator to add the loss which because of energy dissipation, binding to the vortex equation as a component of force equation.Then need to solution the physical models. In this paper, the finite difference method is used in solving the physical equation of smoke, discrete the region of equations which solved is used non-staggered grid, and using the splitting method to solve the model equations. Solving the items of diffusion and projection used the Gauss-Seidel iteration method which method ensures the stability of a large time step; In solving the advection term, used two method:selected the semi-Lagrangian method which unconditionally stable to solving the advection equation, semi-Lagrangian method is stable and easy solution, but it will bring a large number of numerical dissipation. Then focuses on the use of a modified CIP method for solving the equation, it can get a high order result. Original CIP method is unstable, but the modified CIP method is a third-order advection scheme and unconditionally stable. The smoke simulation is conducted in a closed environment.Finally, using the OpenGl rendering technology to drawn the smoke. In this paper, using the vortex operator to reduce the numerical dissipation. Mapping the graphics, they used two methods which is semi-Lagrangian scheme and USCIP scheme.