Interactive Visual Analysis of Time-dependent Flows: Physics- and Statistics-based Semantics

Abstract

With the increasing use of numerical simulations in the fluid mechanics community in recent years flow visualization increasingly gains importance as an advanced analysis tool for the simulation output. Up to now, flow visualization has mainly focused on the extraction and visualization of structures that are defined by their semantic meaning. Examples for such structures are vortices or separation structures between different groups of particles that travel together. In order to deepen our understanding of structures linked to certain flow phenomena, e.g., how and why they appear, evolve, and finally are destroyed, also linking structures to semantic meaning that is not attributed to them by their very definition, is a highly promising research direction to pursue. In this thesis we provide several approaches on how to augment structures stemming from classical flow visualization techniques by additional semantic information originating from new methods based on physics and statistics. In particular, we target separation structures, the linking of structures with a local semantics to global flow phenomena, and minimal representation of particle dynamics in the context of path line attributes.