Expert system for coherent feature detection in high-fidelity fluid dynamic simulations

Current computational fluid dynamics simulations produce large amounts of data. These data have become so large that traditional methods employing manual inspection and visualization cannot effectively process the information. Thus the value of these simulations is diminished because the information they contain is not fully utilized. We proposed to couple leading edge feature detection algorithms with an open-source visualization framework to create develop a flexible, robust, scalable and expandable framework that can be incorporated into industrial, commercial and research codes. Such a framework will enable broad collaboration, but still provide a platform for delivering proprietary modules for long term commercial potential. BENEFIT: This work will develop automated methods for identifying important features in computational fluid dynamics analysis. Such methods are critical to understanding the results of simulations, especially as the simulations become larger and more complex. Without such techniques, it is very difficult to effectively use modern computational tools.