3D shock wave visualization on unstructured grids

Abstract

A critical issue in understanding high speed flows is the study of shock waves. The paper summarizes our research on techniques for the detection and visualization of shock waves occuring in simulations of three dimensional flows on unstructured grids. Detection algorithms based on Mach number, density gradient and directional derivatives are compared using a data set from calculations of a transonic flow with a weak double shock around an airfoil. Both surface and volume rendering techniques are used to display the shocks. The issues in this research area are very much like those occurring in medical imaging. Since the data themselves (in this case the results of the fluid dynamics simulation) are intrinsically low resolution and noisy, properly extracting and visualizing the shock is very difficult. In this environment, blurry, low resolution techniques, like the splatting volume rendering, seem to do rather well. More complex schemes, using sophisticated numerical shock detectors coupled with polygon rendering, produce visually sharper shocks, but also introduce "graphics artifacts," which complicate understanding of the flow physics. On the other hand, visualization results produced with techniques like splatting are, in effect, relying more on the human visual system to compensate for limited resolution in the simulation.