Discrete representations of orbit structures of flows for topological data analysis

TL;DR

This paper introduces a novel discrete symbolic framework for representing the topological orbit structures of a broad class of 2D and 3D flows, enabling new topological data analysis methods for complex fluid flows.

Contribution

It establishes a one-to-one correspondence between flow structures and discrete combinatorial objects, extending existing theories to more general vector fields and practical applications.

Findings

Discrete structures accurately represent flow topologies

Method applies to both incompressible and compressible flows

Demonstrated on industrial 3D flow data

Abstract

This paper shows that the topological structures of particle orbits generated by a generic class of vector fields on spherical surfaces, called {\it the flow of finite type}, are in one-to-one correspondence with discrete structures such as trees/graphs and sequence of letters. The flow of finite type is an extension of structurally stable Hamiltonian vector fields, which appear in many theoretical and numerical investigations of 2D incompressible fluid flows. Moreover, it contains compressible 2D vector fields such as the Morse--Smale vector fields and the projection of 3D vector fields onto 2D sections. The discrete representation is not only a simple symbolic identifier for the topological structure of complex flows, but it also gives rise to a new methodology of topological data analysis for flows when applied to data brought by measurements, experiments, and numerical simulations of complex flows. As a proof of concept, we provide some applications of the representation theory to 2D compressible vector fields and a 3D vector field arising in an industrial problem.