Shallow water layer In this layer, water flows across the paper in the areas indicated by the wet-area mask. A wash is composed of three layers, each of which is responsible for a different part of the simulation. The next chapter will present selected applications of point-based representations in computer graphics.ДЇIGURE 4.5. Depending on the application at hand, one should carefully choose which simulation method is most appropriate. Still, large simulations with photorealistic results have for the most part been left to Eulerian approaches, although recent work has produced results of visual quality comparable to Eulerian simulations. The surface tracking method discussed above, although not suitable for real-time environments, offers tangible advantages over level-set methods. Particle methods are relatively easy to implement, and the simple interaction with objects simulated using different simulation methods is an advantage for example in computer games. Until now, it has been used mainly in interactive or real-time settings with relatively few particles. In practice, methods that use an auxiliary grid to solve for a divergence-free velocity field are more common.ДЇluid simulation using particle methods is a topic of ongoing research. An analogous method exists for SPH simulation. In Eulerian simulations, it is relatively easy to enforce incompressibility by solving a global linear system. In contrast, in an SPH simulation, mass is carried by the particles, thus mass preservation is guarantueed.ДЄ major disadvantage of SPH simulation, especially for fluids like water, is the inherent compressibility of the resulting material. In practice, this leads to mass loss, especially in thin sheets that cannot be resolved by the simulation grid. However, due to integration errors, the total volume of the fluid can change. Usually, the interface is tracked in the velocity field by integration. ДЄnother practical difficulty in Eulerian fluid simulations is the surface representation. For two-way interaction between fluids and other objects, the forces or impulses used to enforce the boundary conditions on the particles can be applied to the boundary. In particle-based simulations, the boundary condition can be applied to individual particles.

The same technique could also be used to represent boundary conditions. developed a method that couples a Eulerian fluid simulation to thin shells. solve the problem by using adaptive meshes to discretize the simulation domain. For performance reasons, these meshes are usually regular grids, thus leading to artifacts when representing boundaries. Due to the discretization of space, boundary conditions in Eulerian simulation have to be aligned with the grid used to represent the simulation domain. In general, boundary conditions are easier to enforce in particle-based methods than in Eulerian simulations. Markus Gross, in Point-Based Graphics, 2007 7.3.6 CONCLUSIONДЄ particle-based fluid simulation has both advantages and disadvantages over alternative approaches, most prominently Eulerian fluid simulation methods.

Fluid simulation