This paper was presented at the NAFEMS DACH seminar "New Methods in CFD - Alternatives To Finite Volume Methods?"

The numerical flow simulation has developed enormously in recent years, both in number and width of applications as well as the complexity of the tasks. More and more complex physical models can be solved in reasonable computing times in large computing areas. In recent years, mostly classical finite-volume or finite element or finite difference methods have been used to solve the equations. A disadvantage of these methods is that calculation grids must be elaborately created in preprocessing.

For quite some time, alternative methods have been available. In addition to grid-free methods, eg. For example, the Smoothed Particle Hydrodynamics (SPH) or the Finite Pointset Method (FPM), Lattice-Boltzmann methods are used successfully. Spectral methods as well as boundary element methods complement the extensive range of options in flow calculation. This also opens up new possibilities with regard to the use of modern computer architectures (eg GPU), which constantly expand the limits of what is feasible in numerical flow calculation.

Many of these solution processes are at the beginning of their industrial application or are already in productive use. In particular for special applications (eg movement of free surfaces, large deformations of the computing area) interesting options arise for the calculation engineer.

This seminar explored new methods for Computational Fluid Dynamics and their industrial application.