Cloud Simulation Based On Physical Model

In Computer Graphics area, people gradually reproduce the real world around us in the computers. Therefore the simulation and rendering of the natural phenomenon is becoming more and more important in recent decades, while the fluid simulation based on physical models is becoming a hot research topic. Building a similar model of the fluid flow is a very important and challenging question, a good solution can play a positive and significant role in other related fields.Cloud as an important element of outdoor scenes is a very complex natural phenomena, which is the atmospheric flow visualization with the irregular shape, complex and randomly movement. As a result, the three-dimensional simulation of the cloud has become the most challenging problem in computer graphics field. At present, as for the static and dynamic research on the cloud simulation, the researchers have done enough research about the simulation of the static clouds, but lacked of researching for the simulation of the dynamic clouds. Yet in reality, the fast simulation technology of the dynamic clouds has important application value in the computer game animation, television advertising and so on. Roaming in three-dimensional virtual scene and flight simulation, real-time simulation of the dynamic cloud becomes increasingly important.In this thesis, Firstly, the current research and results about the cloud have been reviewed; and then the fluid simulation based on physical models has been analysised. Secondly, the Navier-Stokes equation, described the phenomenon of the fluid, has been discussed in detail, for solving which, semi-Lagrangian method and the finite difference method as a core ideology are introduced, through the powerful GPU computing power and programmability of parallelism to speed up the calculation of Navier-Stokes equations, based on the physical characteristics of cumulus clouds, a new real-time simulation algorithms about dynamic cumulus clouds is proposed, the use of physical models based on fluid dynamics and particle systems thinking approach combines the modeling of the cloud, so that this simulation method is both realistic based on the advantages of physical simulation, another method of simple particle system advantage, and ultimately generate real-time, realistic simulation of the dynamic effects of cumulus clouds.