A predictive model for bubble–particle collisions in turbulence

Timothy T.K. Chan, Linfeng Jiang, Dominik Krug · 2025

Read A predictive model for bubble–particle collisions in turbulence by Timothy T.K. Chan, Linfeng Jiang, Dominik Krug (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-13878.txt.

Introduction

Timothy Chan, Linfeng Jiang, and Dominik Krug build a predictive model for bubble-particle collisions in turbulent flow, a key step in froth flotation efficiency. Bubbles are typically much larger than particles, and no earlier model handled finite bubble size and particle inertia together consistently. Building on a frozen turbulence approach, they now predict collision rates a priori from bubble, particle, and turbulence properties.

Comparisons with simulations show good agreement for modest Froude numbers across a range of bubble and particle inertia values. The authors give a criterion for the model's validity beyond the tested range. At typical flotation parameters, particle inertia and settling usually matter, and smaller bubbles, larger particles, and stronger turbulence raise collision rates.

Catalog record

CREATORTimothy T.K. Chan, Linfeng Jiang, Dominik Krug
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE82 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-13878

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A predictive model for bubble–particle collisions in turbulence Timothy T.K. Chan, Linfeng Jiang, Dominik Krug Abstract The modelling of bubble–particle collisions is crucial to improving the efficiency of industrial processes such as froth flotation. Although such systems usually have turbulent flows and the bubbles are typically much larger than the particles, there currently exist no predictive models for this case which consistently include finite-size effects in the interaction with the bubbles as well as inertial effects for the particles simultaneously. As a first step, Jiang and Krug 2025 proposed a frozen turbulence approach which captures the collision rate between finite-size bubbles and inertial particles in homogeneous isotropic turbulence using the bubble slip velocity probability density function measured from simulations as an input. In this study, we further develop this approach into a model where the bubble–particle collision rate can be predicted a priori based on the bubble, particle, and turbulence properties. By comparing the predicted collision rate with simulations of bubbles with Stokes numbers of 2.8 and 6.3, and particles with Stokes numbers ranging from 0.01 to 2 in turbulence with a Taylor Reynolds number of 64, good agreement is found between model and simulations for Froude number F ​ r ≤ 0.25 Fr\leq 0.25 . Beyond this range of bubble Stokes number, we propose a criterion for using our model and discuss the model’s validity. Evaluating our model at typical flotation parameters indicates that particle inertia effects are usually important. Generally, smaller bubbles, larger particles, and stronger turbulence increase the overall collision rate. † † graphicalabstract: † † highlights: Fully predictive model of collision rate in homogeneous isotropic turbulence Includes effects of particle inertia and bubble flow field distortion Bubble, particle, and turbulence properties are the only inputs required Potential subgrid scale model for industrial scale froth flotation simulations Direct numerical simulations of finite-size bubbles and point-particles in turbulence 1 Introduction Bubble–particle collisions lie at the heart of froth flotation — an industrial process which selectively recovers particles based on their hydrophobicity with applications ranging from mineral extraction to wastewater treatment ( Rubio et al. 2002 ; Nguyen and Schulze 2004 ) . In this process, bubbles and particles are injected into a cell filled with water so they can collide with each other. As the bubbles and particles collide, the target particles (which are hydrophobic) attach to the bubbles and are floated to the surface, where they can be removed. In other words, every single particle has to collide with a bubble before it is recovered. Given the broad range of application of froth flotation, as well as the large size of flotation cells and the complex flows therein which precludes particle-resolved simulations, there is a strong incentive to model the bubble–particle collision rate. This parameter is usually reported in terms of the collision kernel \Gamma=Z/n_{b}n_{p}, (1) where Z is the collision rate per unit volume and n_{b} and n_{p} are the bubble and particle number densities, respectively, to facilitate comparison between systems with a different number of particles. 1.1 Bubble–particle collisions in still fluid A key characteristic of bubble–particle collisions in flotation of mineral particles lies in the huge size difference between bubbles and particles, with typical bubble-to-particle radius ratios r_{b}/r_{p}\sim\mathrm{\textit{O}}(10) ( Rubio et al. 2002 ; Ngo-Cong et al. 2018 ; Kostoglou et al. 2020a ) . This means that the actual collision rate is significantly affected by the distortion of the flow field due to the bubbles. Many related studies have focused exclusively on the case of a bubble rising in a still fluid with a constant slip velocity of w_{b} ( Huang et al. 2012 ; Nguyen and Schulze 2004 ; Sarrot et al. 2005 ; Schulze 1989 ; Weber and Paddock 1983 ; Zheng et al. 2024 , see e.g.) . In this limit, the problem is deterministic. As figure 1 ( a ) illustrates, only particles that are initially (i.e. in the undisturbed flow far from the bubble) located within a distance r_{g} from the centreline of the cylinder swept by the bubble eventually collide with the bubble ( r_{g} is hence known as the ‘grazing radius’). The ratio of the number of particles that actually hit the bubble to the number of particles inside the cylinder swept by the bubble is the ‘collision efficiency’ E_{c}=\bigg(\frac{r_{g}}{r_{b}}\bigg)^{2}, (2) such that when the particle number density immediately upstream of the bubble is equal to the global value, the collision kernel is given by \Gamma=\pi r_{b}^{2}w_{b}E_{c}. (3) In general, E_{c} is determined by five nondimensional parameters, i.e. E_{c}=f\bigg(Re_{b}=\frac{2r_{b}w_{b}}{\nu},\frac{r_{b}}{r_{p}},St_{p}^{\prime}=\frac{\tau_{p}w_{b}}{2r_{b}},\frac{v_{s}}{w_{b}},\theta_{cap}\bigg), (4) where Re_{b} is the bubble Reynolds number, \nu is the kinematic viscosity, St_{p}^{\prime} is the particle Stokes number based on the bubble slip velocity, \tau_{p}=r_{p}^{2}(2\rho_{p}/\rho_{f}+1)/(9\nu) is the particle response time, \rho_{p} is the particle density, \rho_{f} is the fluid density, v_{s} is the particle settling speed, and \theta_{cap} is the bubble contamination angle. In the discussion below, we restrict ourselves to the case of fully contaminated bubbles ( \theta_{cap}=\pi ) and non-settling particles ( v_{s}=0 ). The former is reasonable for bubbles in flotation ( Nguyen and Schulze 2004 ; Huang et al. 2012 ) , while the latter assumption is strictly valid when v_{s} is much smaller than the bubble velocity and is imposed here as a simplification of the real situation. In the St_{p}^{\prime}\to 0 limit particles perfectly follow the fluid flow, a collision occurs if the streamline carrying the particle comes within the collision distance r_{c} . This is termed the ‘interception mechanism’ ( Schulze 1989 ) and Sarrot et al. 2005 showed that E_{c}^{(i)}=\bigg(\frac{r_{g}}{r_{b}}\bigg)^{2}=\frac{3}{2}\bigg(\frac{r_{p}}{r_{b}}\bigg)^{2}\bigg(1+\frac{Re_{b}^{2/3}}{5}\bigg). (5) As St_{p}^{\prime} increases, the particle trajectories are less affected by the locally distorted flow field, which can cause more of the incident particles to collide with the front half of the bubble as shown in figure 1 ( b ), thus increasing the collision efficiency ( Huang et al. 2012 ) . This ‘inertial mechanism’ can be expressed as ( Schuch and Löffler 1978 ; Schulze 1989 ) E_{c}^{(in)}=\bigg(1+\frac{r_{p}}{r_{b}}\bigg)^{2}\bigg(\frac{St_{p}^{\prime}}{St_{p}^{\prime}+a}\bigg)^{b}, (6) where a and b are fitting parameters (see A ). Schulze 1989 proposed to model the overall collision efficiency as a superposition of the interception and inertial contributions, leading to E_{c}=E_{c}^{(i)}+E_{c}^{(in)}\bigg(1-\frac{E_{c}^{(i)}}{(1+r_{p}/r_{b})^{2}}\bigg). (7) Here, the term in brackets ensures that E_{c} scales as (1+r_{p}/r_{b})^{2} in the high St_{p}^{\prime} limit. Figure 1: ( a ) Sketch of a particle with St_{p}^{\prime}=0 travelling along the grazing trajectory (black solid line) to collide with a bubble in still fluid. The red dashed line is the corresponding streamline. The collision cylinder used to define the collision efficiency is also shown. ( b ) Sketch of a bubble–particle collision with finite St_{p}^{\prime} where the grazing trajectory deviates from the corresponding streamline. 1.2 Bubble–particle collisions in turbulence The above deterministic picture, however, falls short in the presence of turbulence, which is an important factor in most applications. In these, turbulence is either induced by the rising bubbles or even generated in order to increase the bubble–particle collision rate ( Jiang and Krug 2025 ) . Turbulence introduces characteristic time and length scales into the problem leading to the emergence of additional dimensionless numbers. The most relevant ones are the Taylor Reynolds number Re_{\lambda} , which indicates the turbulence intensity, the Froude number Fr characterising

Read the complete text