Regime Classification Methods for a Gas Jet in a Liquid Co-Flow

Abstract

The present effort evaluates various methods to classify flow regimes from computational fluid dynamics (CFD) simulations of various multiphase flows. The present effort specifically aims to understand a gas jetted into a liquid co-flow, with parametric variations of the freestream liquid velocity and jet mass flow rate. Previous CFD indicated several flow regimes occur in this interaction. However, the method of regime classification relied solely on visual inspection of simulation videos, as well as time-averaged solutions. This work compares several methods to classify gas-liquid interactions, and assesses the regime identification capabilities of each method. The methods evaluated are as follows: Proper Orthogonal Decomposition (POD), Fast Fourier Transformation (FFT), and Mixture of Gaussian (MOG) methods. The results indicate that POD may be the most efficient, effective method to classify the flow regime.