Complexity Analysis of Unsaturated Flow in Heterogeneous Media Using a Complex Network Approach

TL;DR

This paper applies complex network theory to analyze the complexity of two-phase flow in heterogeneous soils, revealing small-world network properties and linking network metrics to flow dynamics and phase transition criteria.

Contribution

It introduces a novel approach using complex network analysis to study unsaturated flow in heterogeneous media, connecting network properties with flow behavior and phase transitions.

Findings

Networks exhibit small-world properties with low mean path length and high clustering.

Evolution of macro-states correlates with network structure complexity.

Phase transition criteria can be identified through network-based velocity profile analysis.

Abstract

In this study, we investigate the complexity of two-phase flow (air/water) in a heterogeneous soil sample by using complex network theory, where the supposed porous media is non-deformable media, under the time-dependent gas pressure. Based on the different similarity measurements (i.e., correlation, Euclidean metrics) over the emerged patterns from the evolution of saturation of non-wetting phase of a multi-heterogeneous soil sample, the emerged complex networks are recognized. Understanding of the properties of complex networks (such degree distribution, mean path length, clustering coefficient) can be supposed as a way to analysis of variation of saturation profiles structures (as the solution of finite element method on the coupled PDEs) where complexity is coming from the changeable connection and links between assumed nodes. Also, the path of evolution of the supposed system will be illustrated on the state space of networks either in correlation and Euclidean measurements. The results of analysis showed in a closed system the designed complex networks approach to small world network where the mean path length and clustering coefficient are low and high, respectively. As another result, the evolution of macro -states of system (such mean velocity of air or pressure) can be scaled with characteristics of structure complexity of saturation. In other part, we tried to find a phase transition criterion based on the variation of non-wetting phase velocity profiles over a network which had been constructed over correlation distance.