Visualisation of particle-based simulations with implicit surfaces

OData support
Supervisor:
Dr. Szécsi László
Department of Control Engineering and Information Technology

The simulation and visualization of natural phenomenons play an important role in the virtual reality systems and in computer games. Real-time and realistic visualization as well as interaction with objects of fluids and liquids are part of the simulation. The physical formulas that describe the fluid dynamics are usually complex differential equations, which cannot be solved in real-time with the current state of information technology. Therefore approximation algorithms should be used, which solve a simplified discretized version of the formulas. The Lagrangian particle-based approach is the most widely used because of its speed and closeness to the object-oriented aspect, but there are alternatives, such as the Eulerian grid-based approach.

Due to the discrete approximation of the simulations it is not acceptable to display directly the evolved surface. Thus the construction of an implicit surface from the generated surface is necessary, which is also not trivial and requires high performance. The metaball method provides a solution for the creation of fluid-like implicit volume, however there are several ways of application. One approach is the traditional raytracing, but there is a newer, screen space based approximation, too.

This thesis aims to offer solutions for both simulation and visualization that can achieve a real-time and spectacular fluid simulation with the computing capacities of current hardwares. First it presents the main types of simulations and gives a detailed description of the Lagrangian particle-based approach and its implementation. To test the simulation it describes the construction of a simple scene, where the fluid movement can be interactively observed. For the visualization of particles it explains in details the metaball method and its two specific, but different realizations. Whilst it enumerates the advantages and disadvantages of the alternatives. Finally, it compares and evaluates the techniques by the aspects of complexity, development time and performance, with emphasis on their individual characteristics that affect the performance.

Downloads

Please sign in to download the files of this thesis.