Original Research Paper
Numerical investigations of acoustic agglomeration of liquid droplet using a coupled CFD-DEM model

https://doi.org/10.1016/j.apt.2020.04.003Get rights and content

Highlights

  • Acoustic agglomeration of liquid droplets in acoustic field is numerically studied.

  • The CFD-DEM model is validated against experimental data.

  • Effect of droplet fragmentation on acoustic agglomeration is investigated.

  • Effect of acoustic and granular parameters on aerosol agglomeration is analyzed.

Abstract

Acoustic agglomeration is widely considered a potentially effective technology for application in artificial defogging and precipitation. A coupled three-dimensional Computational Fluid Dynamics and Discrete Element Method (CFD-DEM) model was constructed to investigate the agglomeration performance of liquid droplets in the acoustic field. The acoustic field is calculated by solving the Linearized Navier-Stokes Equations (LNSEs) in the time domain, and the background flow is initially obtained using the Reynolds-averaged Navier-Stokes (RANS) equations with a kε turbulence model. The motion of the droplet aerosol follows Newton’s second law with fluid-particle and particle-particle interactions, including collision, agglomeration, and fragmentation. The agglomeration performance of liquid droplets under high-intensity acoustic waves was numerically investigated in terms of the effects of the acoustic properties as well as the droplet characteristics.

The numerical results show that it is necessary to consider droplet fragmentation in the process of acoustic agglomeration under the action of high-speed jet. The sprayed droplets are more likely to collide and condense than those without a breakup model, which has rarely been reported in previous studies. Acoustic frequency has a significant effect on agglomeration behavior, with optimal frequencies of about 225 Hz, 150 Hz, and 125 Hz corresponding to droplets with mode diameters of 15.97 μm, 25.85 μm, and 42.88 μm, respectively. However, despite the fact that most studies favoured large acoustic intensity for agglomeration performance, the agglomeration performance of aerosol particles is not always positively correlated with acoustic intensity, especially for large droplets. The optimal intensity of droplet with dp = 42.88 μm is in the range of 120-130 dB, which is smaller than the maximum operation pressure of 150 dB used in this study. In addition, an effective approach to increase the agglomerate size is to extend the residence time that liquid droplets are exposed in the acoustic and flow field, especially because the typical acoustic intensity of actual operation is usually not that high.

Introduction

It is important to investigate the effects of acoustic waves on the agglomeration and dispersion of suspended aerosols, as this technology has been applied to industrial dust removal and airport defogging [1], [2], [3]. In recent years, acoustic agglomeration technology has been applied to enhance artificial precipitation [4], [5], [6]. As a potentially effective technology for artificial defogging and precipitation, acoustic agglomeration has received wide attention due to its low operating cost and lack of pollution. The main principle of acoustic agglomeration technology is to emit directional low-frequency acoustic waves to the target area with high-concentration aerosol, causing the collision of liquid droplet or droplet aerosol and agglomeration of these droplets into large-sized droplets, eventually leading to a decrease in the droplet concentration (Fig. 1). Therefore, the key requirement for application of this technology is to understand the mechanism of acoustic agglomeration of liquid droplets and increase acoustic precipitation.

The phenomenon of acoustic agglomeration was first observed by Patterson and Cawood [7], which led to several experimental and theoretical investigations [8], [9], [10]. Acoustic agglomeration is a spatially uneven and temporally unstable process involving a variety of mechanisms such as orthokinetic interaction [8], [11], hydrodynamic interaction [12], [13], [14], acoustically generated turbulence [15], [16], [17], and Brownian agglomeration [18], [19]. Of these, orthokinetic and hydrodynamic interactions are widely considered the main agglomeration mechanisms. Orthokinetic interaction refers to agglomeration behavior caused by different amplitudes of particles with different sizes under the entrainment of air flow. This theory successfully explains the collision and agglomeration behavior of discrete particles in a multi-dispersed system, but it fails to explain monodisperse particle groups lacking relative oscillation motion [13], [20]. Instead, hydrodynamic interactions that originate from the viscous interaction between aerosol particles and the ambient asymmetry flow are generally considered to be the dominant agglomeration mechanism of monodisperse aerosols in the acoustic field. Typical hydrodynamic interactions include the acoustic wake effect [11], [21], [22] and mutual radiation pressure effect [10], [12], [20], [23]. The acoustic wake effect refers to the influence of the wake of aerosol particles on the motion responses of sheltered particles. A particle in the wake of another particle, also known as a sheltered particle, has a tendency to accelerate toward the front particle due to the decrease of fluid resistance (Fig. 2). In another arrangement, two particles attract each other when their center line is perpendicular to the direction of sound propagation. This is due to a decrease in the gas pressure in the narrow area between the two particles due to the shrinkage of the overflow area according to Bernoulli's hydrodynamic principle. The mutual radiation pressure effect is caused by the scattering of acoustic waves on the particles, which can be considered as a refill mechanism of orthokinetic agglomeration [24]. However, this analytical theory has not been fully verified by experimental observation [13], [14].

In general, although above mechanisms reveal dynamic processes of acoustic agglomeration, understanding of these processes remains insufficient or even contradictory [22]. Most studies have focused on the motion patterns of individual and binary particles under the action of acoustic waves, while ignoring the macroscopic agglomeration behavior of particle clusters. The complex dynamic behaviors of particle clusters in the fields of acoustic wave and jet flow, such as collision, agglomeration, and fragmentation, have not been fully considered. Until now, most studies on acoustic agglomeration have concentrated on parameter optimization of the macroscopic removal effect of solid particles [3], [25], [26], [27], but related research for liquid particles is quite limited [28]. Unlike solid agglomerates, liquid agglomerates lack multiple contact structures, and several small liquid aerosols can merge into a new larger droplet after agglomeration [29]. Overall, in-depth research is required to better describe the dynamics and agglomeration performance of liquid droplets in the acoustic field.

Most studies on the acoustic agglomeration effect of aerosol particles have been theoretical analysis or semi-deterministic experimental research. However, laboratory and field tests are limited because (i) the physical parameters of droplet aerosols are difficult to monitor finely, (ii) the experimental boundary conditions are difficult to control, and (iii) the high cost of experimental research makes it difficult to comprehensively optimize acoustic agglomeration of droplet aerosols based on theoretical analysis and physical experiments. Instead, the mechanism of acoustic agglomeration of droplet can be studied by numerical methods. Numerical simulation is widely applied in the field of multiphase flow, and use of a numerical model allows analysis of discrete particle dynamics and acoustic agglomeration process with a wide operating parameter range [30], [31], [32], [33], [34], which largely compensates for the lack of theoretical analysis and experimental research. In general, numerical models in this field fall into two categories. Population balance modeling (PBM) uses statistical methods to describe discrete phase dynamic events and obtains the dynamic evolution process of the internal variables of the particle group by solving the particle internal function distribution transfer equation [35]. Because the agglomeration mechanisms of discrete system are insufficiently understood, present acoustic agglomeration models are usually simplified in terms of agglomeration kernels. Examples of these simplifications include (i) only considering orthokinetic interaction while neglecting hydrodynamic interaction [9], [36], [37], or (ii) only considering the mutual radiation pressure effect while ignoring the acoustic wake effect [10], [38]. Of course, these simplifications directly affect the accuracy of the numerical model. In addition, PBM approaches usually assume that the particle field is uniform and isotropic, which is to some extent contrary to what occurs in real agglomeration events [39]. The second main type of numerical model is the Lagrangian-based Discrete Element Method (DEM) model [40]. The DEM model can characterize the evolutionary dynamics through the detailed information of a discrete system and is widely applied in the field of particle flow [41], [42], [43]. DEM strategy directly solves the motion of discrete particles in the flow field according to Newton's second law, allowing good resolution of time-varying dynamic processes of discrete particles. Recently, some researchers applied a DEM model to investigate the agglomeration behavior of aerosol particles in acoustic waves [30], [32], [34], [44], [45], [46]. The main agglomeration mechanisms such as orthokinetic interaction [8] and hydrodynamic interaction [13], [14] are implicitly included in the DEM simulation by introduction of drag force and mutual radiation pressure force into the motion equations of individual particles, without the requirement for an agglomeration kernel function [32]. The acoustic agglomeration performance of micron-sized solid particles in the acoustic field has been fully investigated in two-dimensional [32], [46] and three-dimensional [30], [34], [45] particle systems. These studies demonstrate the advantages of using the DEM model to simulate the acoustic agglomeration process, allowing characterization of the motion response of aerosol particles. However, effects of turbulent airflow and droplet fragmentation on acoustic agglomeration have not yet been studied. The motion of the air flow under acoustic waves is often reduced to a type of sinusoidal wave, which may differ from reality.

Therefore, this work presents a three-dimensional CFD-DEM model to investigate the dynamic characteristics of liquid droplets in the field of acoustic wave and turbulent flow with consideration of droplet fragmentation. A four-way coupling strategy is utilized, allowing simulation that considers the drag effect of the fluid on the droplet, the disturbance of droplets on the fluid, movements of the air flow and droplet, and the interactions between the droplets. The airflow process is obtained by solving the Reynolds-averaged Navier-Stokes (RANS) and fluid turbulence characteristics are taken into account. The velocity and trajectory of liquid droplet are solved using Newton's second law with consideration of the drag force, gravitational force, and acoustophoretic force. The agglomeration performances of liquid droplets under high-intensity acoustic waves can then be numerically investigated in terms of the effects of the acoustic characteristics of acoustic frequency, acoustic intensity, and operating period as well as the droplet size and distribution.

The paper is organized as follows. After the introductory section, the methodology of numerical model is presented, which includes the governing equations of the CFD-DEM model, the boundary treatment method, and the time stepping technique. In Section 3, the numerical model is validated against the acoustic field structure data and liquid particle distribution. In Section 4, influences of acoustic and droplet parameters on aerosol agglomeration are discussed. Additionally, relationships between agglomeration performance and operating parameters are described. In Section 5, conclusions are summarized.

Section snippets

Governing equations for air flow

Continuous airflow is solved by the Reynolds-averaged Navier-Stokes (RANS) equations as follows [47]ρ·u=0ρut+ρu·u=·σ+Fpl+gwhere ρ denotes fluid density, u denotes time-averaged fluid velocity, Fpl denotes the volume force induced by liquid droplets, g denotes the gravity force, and σ denotes the stress tensor, which can be expressed asσ=-pI+μu+uT-23μ·uIwhere p denotes the fluid pressure and μ denotes the total fluid viscosity, which is the sum of the molecular and dynamic viscosities.

To

Experimental setup

The experimental device includes an agglomeration chamber, an industrial atomizing nozzle, an acoustic generator and its measuring device, and a laser particle size analyzer. The measuring range of the acoustimeter (Model 308 from BSWA Tech, Inc.) is 22-154 dB and its precision is 0.1 dB. A laser particle size analyzer (Model 319 from Winner Particle, Inc.) is used to observe the size and distribution of micro-droplets in the agglomeration chamber in real time, with a range of 1-500 μm and

Influence of droplet fragmentation

Fig. 13(a) shows that the flow field in the agglomeration chamber is obviously uneven in space, which is different from what was reported in previous studies [22], [32], [45]. There is a conical high-speed airflow area above the atomizing nozzle, and the liquid droplets inside this region are carried by the ambient air-flow and thereby have a relatively higher velocity. Liquid droplets with higher velocities also tend to have more kinetic energy and move more violently. It can be seen from

Conclusions

This paper investigated the acoustic agglomeration of liquid droplets using a three-dimensional CFD‐DEM coupled model. The numerical model solves the flow field based on the Reynolds-averaged Navier-Stokes equations coupled with a kε turbulence model, and describes the acoustic field by Linearized Navier-Stokes Equations (LNSEs) in the time domain. The motion of the liquid droplet follows Newton’s second law with consideration of complex aerosol dynamics process such as collision,

CRediT authorship contribution statement

Yang Shi: Formal analysis, Methodology, Writing - original draft. Jiahua Wei: Methodology, Data curation, Writing - review & editing, Supervision. Wenwen Bai: Formal analysis. Guangqian Wang: Supervision.

Acknowledgements

This research is supported by National Key Research and Development project, grant number 2017YFC0403600, 2016YFE0201900; the China Postdoctoral Science Foundation, grant number 2018M641372; State Key Laboratory of Hydroscience and Engineering of China, grant number 2017-KY-04.

References (78)

  • R. Kacianauskas et al.

    Comparative DEM-CFD study of binary interaction and acoustic agglomeration of aerosol microparticles at low frequencies

    Chem. Eng. Res. Des.

    (2018)
  • D. Markauskas et al.

    Numerical particle-based analysis of the effects responsible for acoustic particle agglomeration

    Adv. Powder Technol.

    (2015)
  • G. Zhang et al.

    A new multiple-time-step three-dimensional discrete element modeling of aerosol acoustic agglomeration

    Powder Technol.

    (2018)
  • S. Kumar et al.

    On the solution of population balance equations by discretization. -I. A fixed pivot technique

    Chem. Eng. Sci.

    (1996)
  • O. Ezekoye et al.

    Simulation of acoustic agglomeration processes using a sectional algorithm

    J. Aerosol Sci.

    (1999)
  • A. Reinhold et al.

    Discrete element modeling for multiscale simulation of aggregation processes

  • Z.X. Li et al.

    Nanofluids as secondary fluid in the refrigeration system: Experimental data, regression, ANFIS, and NN modeling

    Int. J. Heat Mass Transfer

    (2019)
  • A. Holmberg et al.

    A frequency domain linearized Navier-Stokes method including acoustic damping by eddy viscosity using RANS

    J. Sound Vib.

    (2015)
  • M.-M. Kim et al.

    Effect of electrostatic, hydrodynamic, and Brownian forces on particle trajectories and sieving in normal flow filtration

    J. Colloid Interface Sci.

    (2004)
  • X. Wang et al.

    A multi-sphere based modelling method for maize grain assemblies

    Adv. Powder Technol.

    (2017)
  • Y. He et al.

    Discrete modelling of the compaction of non-spherical particles using a multi-sphere approach

    Miner. Eng.

    (2018)
  • P. Capéran et al.

    Acoustic agglomeration of a glycol fog aerosol: Influence of particle concentration and intensity of the sound field at two frequencies

    J. Aerosol Sci.

    (1995)
  • J. Yan et al.

    Removal of fine particles from coal combustion in the combined effect of acoustic agglomeration and seed droplets with wetting agent

    Fuel

    (2016)
  • B.F. Ng et al.

    Application of acoustic agglomeration to enhance air filtration efficiency in air-conditioning and mechanical ventilation (ACMV) systems

    PLoS One

    (2017)
  • M.S. Lu et al.

    Insights into agglomeration and separation of fly-ash particles in a sound wave field

    RSC Adv.

    (2019)
  • T. Tulaikova et al.

    Acoustic Rains

    (2010)
  • O. Nalbandyan

    The clouds microstructure and the rain stimulation by acoustic waves

    Atmos. Clim. Sci.

    (2011)
  • G.A. Galechyan

    On acoustic stimulation of atmospheric precipitation

    Tech. Phys.

    (2005)
  • H.S. Patterson et al.

    Phenomena in a sounding tube

    Nature

    (1931)
  • E.P. Mednikov

    Acoustic Coagulation and Precipitation of Aerosols

    (1965)
  • R. Tiwary

    Acoustic Agglomeration of Micron and Submicron Fly-ash Aerosols

    (1985)
  • L. Song

    Modeling of Acoustic Agglomeration of Fine Aerosol Particles

    (1990)
  • S.D. Danilov

    Average force acting on a small sphere in a traveling-wave field in a viscous-fluid

    Sov. Phys. Acoust.

    (1985)
  • T.L. Hoffmann et al.

    Visualization of acoustic particle interaction and agglomeration: Theory and experiments

    J. Acoust. Soc. Am.

    (1996)
  • T.L. Hoffmann et al.

    Visualization of acoustic particle interaction and agglomeration: Theory evaluation

    J. Acoust. Soc. Am.

    (1997)
  • R. Tiwary et al.

    Acoustically generated turbulence and its effect on acoustic agglomeration

    J. Acoust. Soc. Am.

    (1984)
  • P.S. Lee et al.

    The influence of hydrodynamic turbulence on acoustic turbulent agglomeration

    Aerosol Sci. Technol.

    (1981)
  • G. Zhang et al.

    Numerical simulation of acoustic wake effect in acoustic agglomeration under Oseen flow condition

    Chin. Sci. Bull.

    (2012)
  • H. Bruus

    Acoustofluidics 7: The acoustic radiation force on small particles

    Lab Chip

    (2012)
  • Cited by (19)

    • Investigation of strong acoustic interference on clouds and precipitation in the source region of the Yellow River using KaKu radar

      2022, Atmospheric Research
      Citation Excerpt :

      Previous studies have shown that the SPL value of the air-modulated speaker decays exponentially with the traveling distance (L, in m) of sound waves following the expression of SPL = 148.6 L-0.06339. Since air-suspended droplets can agglomerate in the sound field of 80– 90 dB (Shi et al., 2020a), thus the effective range of a single speaker for acoustic precipitation is in the range of 5– 8 km (Cao et al., 2019), which exceeded the precipitating cloud base height in the Darlag region, at ~2 km (Shi et al., 2021a). Note that the acoustic interference in clouds may cause potential risk on local precipitation balance and ecological environment.

    View all citing articles on Scopus
    View full text