Real-Time Fluid Simulation Based On Physical Models On GPU

The simulation of fluid is more and more important in the field of computer game,movie and advertisement. But their movement is random, simulation of them is difficulty. So it is the hotspot and difficult point of computer graphics all of these days.Parametric modeling is the main method in simulation of fluid in the early days because of the limitation of computation ability, but this method did badly in reality. So people started to explore the root cause of the fluid movement. Only if using the physical description of it can we reappear the visualization of the fluid. The physically based method just can do it. But the computational complexity of this method is very high, so it can not meet the requirement of real-time. With the development of the computer hardware especially the computation ability of the graphics processing unit, the based physically method is more and more important.This paper focus on the following respects: the acceleration of the GPU; the study of the Navier-Stokes equations, especially the solution on GPU; the study on physically method and particle system method of flame simulation, and this paper combine the two methods effectively; the study of the display of the flame.1. The physically model that we use is Navier-Stokes equations. We simulate fluid dynamics on a regular Cartesian grid with spatial coordinates x = (x, y) and time variable t. The fluid is represented by its velocity field u(x, t), as described earlier, and a scalar pressure field p(x, t). These fields vary in both space and time. If the velocity and pressure are known for the initial time t = 0, then the state of the fluid over time can be described by the Navier-Stokes equations for incompressible flow: Whereρis the (constant) fluid density,νis the kinematic viscosity, and F = ( fx, fy ) represents any external forces that act on the fluid.. The"."denotes a dot product between vectors, while the symbol ? is the vector of spatial partial derivatives. We simplified the Navier-Stokes equations by Stable Fluids of Stam1999, and compute it on GPU.This paper modify the advection term by introduce the thought of particle system. By this modify our simulation both verisimilitude like physically based method and simple like particle system method.The similarities and differences of this method and the particle method are as 2. With the development of the GPU, texture is becoming more and more flexible. The natural representation for this grid on the CPU is an array. The analog of an array on the GPU is a texture. Textures on current GPUs support all the basic operations necessary to implement a fluid simulation. Because textures usually have three or four color channels, they provide a natural data structure for vector data types with two to four components. Alternatively, multiple scalar fields can be stored in a single texture. By scaled and biased, we displayed the textures.3. We introduce random function to produce the horizontal velocity of the source of the fluid by drawing lessons from particle system, and we import Gaussian function as smooth function. To decrease computational complexity we add a vertical velocity to represent the influence of the gravity and buoyancy. f ( x)=ae?(x?b)2 /c24. In this paper we provide a preliminary probe into fluid control. By setting specific velocity on specific mesh point, we simulated the wind effects flame and the berried flame. Believe after more fluid control research, we can obtain more varied control effect and make the animator design more flexible.5. We use density field and color principle to control flame color. The solving of the velocity field,vorticity field and pressure field all come down to get the variations of the density field, so that control the display of the flame. Compare the spectral radiant energy method our method is simple and effective. The two points that we propose in this papar are as follows: 1. We modify the advection term by introduce the thought of particle system. By this modify our simulation both verisimilitude like physically based method and simple like particle system method.2. We use density field and color principle to control flame color. Compare the spectral radiant energy method our method is simple and effective.By study of Navier-Stokes equations, introduce the thought of the particle system, and realize on GPU, we get a flame simulation both reality and real-time.