A Study On Vortex Structures In 3D Unsteady Flows

The vortex motion of fluids is a common phenomenon in engineering and nature. The formation mechanism of vortices and its stability under steady state conditions has been well studied, but the mechanism under unsteady state conditions is still being explored.Traditional analysis method for unsteady complex fluid fields is visualization technique, i.e. by means of using streamlines, tracing lines, isosurfaces, volume rendering and texture mapping. Since the unsteady flow field data is high dimensional time series data, the traditional visualization methods will encounter issues such as information missing and complexity in displaying, and it is necessary to study more intelligent and simplified methods.A new analysis method is proposed in this thesis to study vortices in 3D unsteady flows. Firstly, vortex structures are extracted from a flow field by using the critical point theory based method; secondly, to track them with the vortex core matching method to get the evolution graph of the vortices; then to calculate the geometrical and kinematical parameters; and finally to label vortex evolution graph with vortex properties. Every vertex in the vortex evolution graph represents a vortex in the corresponding time step. The attributes of vertices are the properties of vortex structures. Therefore the vortex evolution graph maps the complex time series data of the 3D field into a directed acyclic graph. Fluid mechanics experts can use them to research the vortex evolution. Lastly we could use a graph mining algorithm to find the most frequent sub-graphs in the vortex evolution graph. The frequent sub-graphs can help researchers better understand the nature of the flow field.Experiments are completed with this method, on datasets for a Rayleigh-Benard unsteady flow field and a field with high-speed trains running into tunnels, and some interesting patterns are found hereby. The analysis process and corresponding method are recognized by fluid dynamics experts.