Stochastic DSMC model for large scale dense bubbly flows

S.S. Kamath; J.T. Padding; K.A. Buist; J.A.M. Kuipers

Stochastic DSMC model for large scale dense bubbly flows

S.S. Kamath
, J.T. Padding
, K.A. Buist
, J.A.M. Kuipers

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

Bubble columns are widely used in the chemical industry because
of their simple design and high efficiency. The scale-up of these
kinds of columns is challenging and time-consuming. Since high
throughput is targeted, they are operated in the heterogeneous bub-
bling regime where the flow is complex and turbulent. Large-scale
bubble columns can in principle be simulated using continuum
models (TFM/MFM) with closures from more detailed models
such as Front Tracking (FT) or Volume of Fluid (VOF). Multi-fluid
models are capable of predicting the flow field, but to accurately
describe mass transfer rates, an accurate interfacial area of the
bubbles is required as well as mass transfer coefficients for dense
bubble swarms. This requires the MFM to be coupled with models
that can predict bubble size distributions. The Discrete Bubble
Model (DBM) can be scaled up but the bubble-bubble interactions
make it computationally very intensive.
Stochastic Direct Simulation Monte Carlo (DSMC) methods treat
the bubbles in a discrete manner while more efficiently handling
the collisions compared to the DBM. The DSMC model has earlier
been used for very small particles in the size range of Angstroms
to microns where the particles are purely inertial at high Stokes
numbers. In the work of Pawar et al. (2014) this was used for
micrometer sized particles/droplets where this method proved to be
60 to 70 times faster than more classical methods like the Discrete
Particle Model (DPM).
In this work the DSMC method has been extended to finite sized
bubbles/particles in the order of millimeters. A 4-way cou-
pling (liquid-bubble-bubble) is achieved using the volume-averaged
Navier Stokes equations. The model is verified first for mono-
disperse impinging particle streams without gas. Then the model
is verified with the DBM of a 3D periodic bubble driven system.
The collision frequencies are all within 10 percent accuracy and the
speed up achieved per DEM time step is nearly 10 times compared
to the DBM, which facilitates simulation of large systems.

Original language	English
Title of host publication	Progress in Applied CFD – CFD2017
Subtitle of host publication	Proceedings of the 12th International Conference on Computational Fluid Dynamics in the Oil & Gas, Metallurgical and Process Industries, 30 May -1 June 2017, Trondheim, Norway
Place of Publication	Blindern
Publisher	SINTEF Academic Press
ISBN (Electronic)	978-82-536-1544-8
Publication status	Published - 2017

Cite this

Kamath, S. S., Padding, J. T., Buist, K. A., & Kuipers, J. A. M. (2017). Stochastic DSMC model for large scale dense bubbly flows. In Progress in Applied CFD – CFD2017: Proceedings of the 12th International Conference on Computational Fluid Dynamics in the Oil & Gas, Metallurgical and Process Industries, 30 May -1 June 2017, Trondheim, Norway SINTEF Academic Press.

@inproceedings{0ac53438d8c2491ea2aaae83e950219b,

title = "Stochastic DSMC model for large scale dense bubbly flows",

abstract = "Bubble columns are widely used in the chemical industry because of their simple design and high efficiency. The scale-up of these kinds of columns is challenging and time-consuming. Since high throughput is targeted, they are operated in the heterogeneous bub- bling regime where the flow is complex and turbulent. Large-scale bubble columns can in principle be simulated using continuum models (TFM/MFM) with closures from more detailed models such as Front Tracking (FT) or Volume of Fluid (VOF). Multi-fluid models are capable of predicting the flow field, but to accurately describe mass transfer rates, an accurate interfacial area of the bubbles is required as well as mass transfer coefficients for dense bubble swarms. This requires the MFM to be coupled with models that can predict bubble size distributions. The Discrete Bubble Model (DBM) can be scaled up but the bubble-bubble interactions make it computationally very intensive. Stochastic Direct Simulation Monte Carlo (DSMC) methods treat the bubbles in a discrete manner while more efficiently handling the collisions compared to the DBM. The DSMC model has earlier been used for very small particles in the size range of Angstroms to microns where the particles are purely inertial at high Stokes numbers. In the work of Pawar et al. (2014) this was used for micrometer sized particles/droplets where this method proved to be 60 to 70 times faster than more classical methods like the Discrete Particle Model (DPM). In this work the DSMC method has been extended to finite sized bubbles/particles in the order of millimeters. A 4-way cou- pling (liquid-bubble-bubble) is achieved using the volume-averaged Navier Stokes equations. The model is verified first for mono- disperse impinging particle streams without gas. Then the model is verified with the DBM of a 3D periodic bubble driven system. The collision frequencies are all within 10 percent accuracy and the speed up achieved per DEM time step is nearly 10 times compared to the DBM, which facilitates simulation of large systems.",

author = "S.S. Kamath and J.T. Padding and K.A. Buist and J.A.M. Kuipers",

year = "2017",

language = "English",

booktitle = "Progress in Applied CFD – CFD2017",

publisher = "SINTEF Academic Press",

}

Stochastic DSMC model for large scale dense bubbly flows. / Kamath, S.S.; Padding, J.T.; Buist, K.A. et al.
Progress in Applied CFD – CFD2017: Proceedings of the 12th International Conference on Computational Fluid Dynamics in the Oil & Gas, Metallurgical and Process Industries, 30 May -1 June 2017, Trondheim, Norway. Blindern: SINTEF Academic Press, 2017.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Stochastic DSMC model for large scale dense bubbly flows

AU - Kamath, S.S.

AU - Padding, J.T.

AU - Buist, K.A.

AU - Kuipers, J.A.M.

PY - 2017

Y1 - 2017

N2 - Bubble columns are widely used in the chemical industry because of their simple design and high efficiency. The scale-up of these kinds of columns is challenging and time-consuming. Since high throughput is targeted, they are operated in the heterogeneous bub- bling regime where the flow is complex and turbulent. Large-scale bubble columns can in principle be simulated using continuum models (TFM/MFM) with closures from more detailed models such as Front Tracking (FT) or Volume of Fluid (VOF). Multi-fluid models are capable of predicting the flow field, but to accurately describe mass transfer rates, an accurate interfacial area of the bubbles is required as well as mass transfer coefficients for dense bubble swarms. This requires the MFM to be coupled with models that can predict bubble size distributions. The Discrete Bubble Model (DBM) can be scaled up but the bubble-bubble interactions make it computationally very intensive. Stochastic Direct Simulation Monte Carlo (DSMC) methods treat the bubbles in a discrete manner while more efficiently handling the collisions compared to the DBM. The DSMC model has earlier been used for very small particles in the size range of Angstroms to microns where the particles are purely inertial at high Stokes numbers. In the work of Pawar et al. (2014) this was used for micrometer sized particles/droplets where this method proved to be 60 to 70 times faster than more classical methods like the Discrete Particle Model (DPM). In this work the DSMC method has been extended to finite sized bubbles/particles in the order of millimeters. A 4-way cou- pling (liquid-bubble-bubble) is achieved using the volume-averaged Navier Stokes equations. The model is verified first for mono- disperse impinging particle streams without gas. Then the model is verified with the DBM of a 3D periodic bubble driven system. The collision frequencies are all within 10 percent accuracy and the speed up achieved per DEM time step is nearly 10 times compared to the DBM, which facilitates simulation of large systems.

AB - Bubble columns are widely used in the chemical industry because of their simple design and high efficiency. The scale-up of these kinds of columns is challenging and time-consuming. Since high throughput is targeted, they are operated in the heterogeneous bub- bling regime where the flow is complex and turbulent. Large-scale bubble columns can in principle be simulated using continuum models (TFM/MFM) with closures from more detailed models such as Front Tracking (FT) or Volume of Fluid (VOF). Multi-fluid models are capable of predicting the flow field, but to accurately describe mass transfer rates, an accurate interfacial area of the bubbles is required as well as mass transfer coefficients for dense bubble swarms. This requires the MFM to be coupled with models that can predict bubble size distributions. The Discrete Bubble Model (DBM) can be scaled up but the bubble-bubble interactions make it computationally very intensive. Stochastic Direct Simulation Monte Carlo (DSMC) methods treat the bubbles in a discrete manner while more efficiently handling the collisions compared to the DBM. The DSMC model has earlier been used for very small particles in the size range of Angstroms to microns where the particles are purely inertial at high Stokes numbers. In the work of Pawar et al. (2014) this was used for micrometer sized particles/droplets where this method proved to be 60 to 70 times faster than more classical methods like the Discrete Particle Model (DPM). In this work the DSMC method has been extended to finite sized bubbles/particles in the order of millimeters. A 4-way cou- pling (liquid-bubble-bubble) is achieved using the volume-averaged Navier Stokes equations. The model is verified first for mono- disperse impinging particle streams without gas. Then the model is verified with the DBM of a 3D periodic bubble driven system. The collision frequencies are all within 10 percent accuracy and the speed up achieved per DEM time step is nearly 10 times compared to the DBM, which facilitates simulation of large systems.

M3 - Conference contribution

BT - Progress in Applied CFD – CFD2017

PB - SINTEF Academic Press

CY - Blindern

ER -

Stochastic DSMC model for large scale dense bubbly flows

Abstract

Fingerprint

Cite this