Flow Simulation: Computational algorithms for chasing eddies in fluid flows

Tracking a hairpin packet in a Mach 3 turbulent boundary layer from O'Farrell & Martin Journal of Turbulence (2009)

Working with colleagues in the Department of Electrical Engineering and Computer Science at Rutgers University, the CRoCCo Lab is developing and evaluating algorithms for automated feature analyses from time-varying scientific data and the generation of metadata to facilitate research and knowledge discovery. This includes developing a catalogue of feature-activity and feature-evolutions to semantically characterize data in time and space. The dynamic and complex nature of unsteady fluid flow presents major challenges to identification and tracking algorithms, making such data ideal for the development of generic algorithms. In particular, we have developed a method to track dynamically coherent structures using scientifically rooted criteria, in combination with computer science object segmentation techniques and feature tracking algorithms.