Session: 7.6 - Experimental methods for multiphase flows

Paper Number: 170242

170242 - Velocity Field Measurements and Bubble Size Distribution in Bubble-Laden Turbulent Flows: Insights Into Flow Dynamics From Piv and High-Speed Imaging 

Abstract: 

This talk presents an experimental investigation of bubble-laden turbulent flow (water + air) in a square duct, with bulk Reynolds number (Re) ranging ~ 5-10 x 10^4 and air-to-water volume fraction (void fraction) ranging ~ 0.5-2%. Studying bubbly flows at these conditions is important to optimize multiphase industrial processes, reducing energy consumption and refining predictive models for bubble-laden flows in high-intensity turbulence. Flow kinematics is studied via (1) planar two-dimensional two-component (2D-2C) particle image velocimetry (PIV), and (2) high-speed, cost-effective, planar particle shadow velocimetry (PSV). The two methods are compared with one another as well as with numerical simulations. Due to the smaller size of particle shadows seen in PSV images compared to larger ‘light-scattering’ particle images in PIV, the PSV field of view (FOV) is generally smaller than the PIV FOV. Resultantly, PSV is effective resolving smaller flow scales while PIV is effective resolving larger flow scales across the entire flow span. Both methods are used to obtain and compare mean and higher-order turbulence statistics (velocity fluctuations, Reynolds stresses, turbulent kinetic energy) across different experimental cases. Estimates of dissipation in the flow are also presented. The flow velocity field measurements presented here complement previous works on bubble size distribution in multiphase flows and provide deep insights into the dynamics of bubble coalescence and breakup in highly turbulent flows. Inside the pump driving the multiphase experimental flow loop, the intense turbulence creates a breakup-dominated regime. Further downstream in the duct, as turbulence intensity decays by over 90%, bubble coalescence becomes dominant. Consequently, the Sauter mean diameter (d_32) increases with downstream distance and toward the duct center, while the bubble count decreases despite a stable overall void fraction — indicating net coalescence (fewer but larger bubbles). Bubble size distributions deviate from classical log-normal behavior, and the typical size ratio of d_99.8/d_32 ≃ 2.2 observed in highly turbulent, coalescence-dominated regions gradually recovers as turbulence continues to decay downstream. The bubble distribution, velocity measurements, and the higher-order turbulence statistics presented here could be used to validate and develop numerical models for multiphase turbulent flows, as well as initialize numerical simulations.  

Presenting Author: Prasoon Suchandra Georgia Institute of Technology

Presenting Author Biography: Prasoon Suchandra received his PhD in Mechanical Engineering from the Georgia Institute of Technology, where he worked with Prof. Devesh Ranjan on experimental investigation of fluid instabilities and variable-density turbulent flows using particle image velocimetry (PIV) and laser induced fluorescence (LIF). For his postdoctoral research work at the Rowland Institute at Harvard University with Dr. Shabnam Raayai, Prasoon experimentally studied flow past arrays of passive and active objects, to develop efficient group motion strategies for aerial and underwater vehicles. Prasoon is currently working as a research engineer at Georgia Tech with Prof. Cyrus Aidun on experimental characterization of transition to turbulence in non-Newtonian and multi-phase flows in channels and other relevant nozzle geometries using PIV and other optical diagnostics, with the aim to enhance the efficiency of drying of paper and other fiber composite products.