Principal Investigator John Williams
High quality rendering of complex three dimensional simulation results is crucial for evaluation and interpretation of the results. Viewing results within the body of a three dimensional simulation is difficult, particularly for particle methods such as SPH. Solution to this challenge is generally achieved by surface identification, iso-surface calculation and sectioning.
We have developed algorithms based on Marching Cubes to identify surfaces, iso-surfaces etc for SPH simulation results including multi-phase results. The triangular mesh and surface normals which are returned by this algorithm are then rendered using POV-Ray, a freely available ray tracing software. Some of these rendered results are shown in what follows.
Falling Drop Simulation -- The images below demonstrate the tracking of free surfaces in the example of a droplet of water falling on a flat surface of water. The algorithm recognizes the interfaces between air and water and renders them via a curved triangular mesh. The simulation results were derived using the developed SPH simulator.
Three-Dimensional Rayleigh-Taylor Instability -- A more advanced demonstration of the capabilities of the rendering package is in visualizing the 3D Reyleigh-Taylor Instability problem. Here complex surfaces between 3 phases; air, oil and water (and additionally the outer boundaries), must be established based on particle density and tracked in 3D. A curved triangular mesh surface is generated around fluid volumes of each phase and can then be rendered using ray tracing methods. The developed surface tracking method is robust and accurate.
Three-Dimensional Wetting Water Invasion -- Once again, the multi-phase problem of 3D water invasion poses significant challenges for rendering. Here the 3 phases of oil, water and grain (and additionally the outer boundaries) are all present and must be distinguished from one another. The images below demonstrate the developed rendering packages performance.
Model Geometry From Segmented Images -- A key advantage of Smooth Particle Hydrodynamics is in the methods ability to easily accommodate complex boundary definitions. Below are surface plots of the rock (left) and inner pore (right) developed from a segmented image of actual Berea Sandstone. The developed SPH simulator allows model geometry to be defined using such data. Numerical flow testing on such models is being carried out presently.
From the point of view of rendering, the images demonstrate how the developed methodology can generate smoothed surface images from voxelated data. Here the algorithm recognizes rock from void space.