You are on page 1of 4

Multiphase Flow Modeling

Reliable and accurate modeling of many industrial processes


requires accounting for interactions between two or more phases.
The simultaneous ow of phases in a domain is called multiphase
ow; examples include conversion of crude to high-value petroleum
products in a renery, combustion of coal to generate energy in a
power plant, and combustion of gasoline droplets in an internal
combustion engine.
The successful design of equipment and processes that involve
multiphase ows crucially depends on accurately predicting the
interactions (mechanical, thermal and chemical) between the
phases. Since most of these processes are impossible to observe,
engineers rely on models and experiments to gain insight into
improving efciency, throughput, safety and reliability factors
important to a companys bottom line. This knowledge can aid in
scaling up and retrotting existing equipment as well as improving
processes. Computational uid dynamics (CFD) software from ANSYS
incorporates most physical models required to describe multiphase
ows, offering a virtual view inside these processes.
Multiphase ows are broadly classied as either segregated or
dispersed. In segregated ows, the phases owing through the
domain are separated by macroscopic interfaces, often comparable to
the size of the domain. For dispersed ows, at least one of the phases
is a small droplet, bubble or particle.
Segregated Flows
The sloshing of fuel in an automobile, the ow of oil and gas in
a pipeline, the impact of waves on seafaring vessels and offshore
platforms, and the movement of ink in a printer are a few examples
of segregated ows. The volume of uids (VOF) model and the
level-set method are typically used to simulate problems that
involve the evolution of interfaces. In these algorithms a single
momentum equation is solved for the mixture along with the explicit
reconstruction of the interphase between individual phases. ANSYS
CFD software has an extensive array of sophisticated interface
tracking schemes that can be reliably used to model a wide variety of
industrial problems. These models require virtually no empirical input,
thereby enhancing the accuracy and reliability of model predictions.
Multiphase simulation of feed nozzles in
distillation columns for a petroleum renery
Courtesy Petrobras.
Industry Solutions
Comparison of droplet size observed in
experiment and simulations based on
experiments of Nisisako et al. [1] for
production of emulsions in a microuidic
T-junction
Comparison of experiment (grey) and simu-
lation (color) at different times following
a dam break; white regions correspond to
breaking waves in the experiment
Courtesy Technical University of Berlin.
Microuidics
Ink jet printers, lab-on-a-chip devices and cooling circuits for
microelectronics involve multiphase ow in sub-millimeter-length
channels. This type of ow is dominated by surface tension and wall-
adhesion forces. The geometric reconstruction scheme available in
ANSYS CFD allows for the accurate treatment of surface tension in
these ows.
Coupled VOF for Steady-State Flows
The ow of water around a ship hull and the stratied ow of water in
a pipe allow for steady-state solutions. For such problems, ANSYS CFD
has pioneered numerical algorithms that allow for coupled calculation
of volume fraction along with pressure and momentum, providing a
speedy and accurate solution. When coupled with boundary conditions,
such as those that generate waves or represent boundaries at pressure
outlets, these coupled calculations enhance the ability to dependably
simulate real-life problems.
Dispersed Flows
The motion of bubbles in a slurry bubble
column reactor, gasoline droplets from
spray in an internal combustion engine,
and catalyst particles in a uid catalytic
cracker are all examples of dispersed
multiphase ows. Due to the unresolved
subgrid nature of these particles, empirical
models such as drag, virtual mass forces
or lift forces are used to describe the
interaction between phases. ANSYS CFD
pioneered numerical and physical models
for simulating these problems using an
unstructured grid [2].
Granular Flows
The dispersed phase takes the form of
solid particulates in examples such as a
uid catalytic cracker, regenerator, or any
application that includes a solid suspension.
These particles exhibit a uid-like character
due to the emergent behavior of many
individual particles. The rheology of these
particles is a function of the ow condition
and must be addressed as part of the
solution. ANSYS CFD uses the kinetic theory
of granular ow, as well as frictional models
at high volume fractions, to accurately
determine the ow behavior of interacting
solid particles.
Population Balance
In unit operations in which granulation and crystallization occur and in
reactors such as bubble columns, or for oilwater separators or sprays,
the size of the dispersed phase changes due to physical phenomena
such as breakage and coalescence. This size change, in turn, affects the
Industry Solutions
Calculations using the coupled VOF
option converge faster for problems
that require time-independent solutions
The effect of inlet piping
on overall performance of
a horizontal three-phase
separator can be studied
using the Eulerian multi-
phase model.
Courtesy Swift TG Solutions.
The Euler-granular model
is used to understand the
mixing of particles during
suction phase in a mixer for
processing nuclear waste.
Courtesy Bechtel National Inc.
The effect of hardware on distribution of
droplets and catalyst is simulated to under-
stand the under-par performance of a riser
reactor using the Eulergranular model.
Courtesy KBC Advanced Technologies.
Predictions of holdup proles of gas in a
gas lift system using experiments of Guet
[3] and ANSYS CFD simulations at two
different heights from the sparger
4 m Experiment
12 m Experiment
4 m Simulation
12 m Simulation
G
a
s

H
o
l
d
u
p
ow dynamics. ANSYS CFD provides a wealth of models to account
for size change processes using the framework of population balance
models. Sizes can be calculated by tracking mass fraction in either
discrete sizes or in the form of statistical moments.
Multi-Fluid VOF Model
When simulating multiphase ows, situations can arise in which neither
the fully segregated VOF model nor the interpenetrating Eulerian multi-
uid model is adequate to describe ow dynamics. The frothy nose of
an airwater slug or the motion of an air bubble through a slurry of ne
particles illustrate applications in which regions of mixed ow coexist
with large regions of segregated ow. For solving such problems, a
hybrid model that describes the physics between the mixed dispersed
phases as well as the physics at interfaces for segregated phases is built
into ANSYS CFD.
Multiphase Turbulence Models
The effect that turbulence has on multiphase problems is as profound
as its impact on single-phase problems. Extensions to the single-
phase turbulence models, such as Tchens model for secondary phase
turbulence and the particle-induced turbulence model, are available
to better model multiphase turbulence. The effect of turbulence
uctuations on particle dispersion is available in both Favre-averaged
and volume fraction-weighted formulations. For segregated ows using
the VOF model, out-of-the-box ANSYS CFD accounts for turbulence
damping at interfaces, which leads to better predictions for many
turbulent gasliquid ows (for example, in a slug catcher).
Multiphase Reactions
One of the most important applications
for multiphase ow is chemical reactions
or thermodynamic changes to the phases.
ANSYS CFD provides templates to easily model
homogeneous and heterogeneous reactions as
well as phase change processes. The software
also includes the ability to customize kinetics
for these processes. Out-of-the-box models
are available for some common phase change
processes such as cavitation and boiling
models.
Dense Lagrangian Models
Lagrangian tracking of particles in multiphase
systems provides an economical model for
handling the wide particle size distribution
that naturally occurs in uidized beds and
bubble columns, for example. The dense
discrete phase (DDPM) model extends the
Eulerian multiphase models in ANSYS CFD
by allowing for phases to be represented by
a Lagrangian description for all volume fractions up to the maximum
packing. This feature is unique to ANSYS CFD and provides solution to
problems that involve the need to model particle tracks as well as those
involving a wide range of particle sizes.
Industry Solutions
The motion of a swarm of bubbles
through a slurry of ne particles is simu-
lated using the multi-uid VOF model.
The ne solids (red) are dispersed within
the water phase, while air bubbles (cyan)
are segregated from the water phase.
The effect of an increase in throughput of
a slug-catcher is simulated. Appropriate
treatment of turbulence at the interface
is crucial for these simulations.
Courtesy BG Group.
The Euler-granular model helps in under-
standing the complex chemical interactions
between the formation of clusters and
the overall effectiveness of the reactor.
Studies suggest that increasing reactivity
of the catalysts does not always result in a
proportional increase in the effectiveness
of the reactor.
Courtesy SINTEF Material and Chemistry.
Nucleate boiling down-
stream of spacers in fuel
rod bundle assembly
Courtesy Dr. E. Krepper, FZ Dresden.
The ANSYS Multiphase Team
The multiphase development team is engaged in R&D projects and
actively collaborates with academic experts and leaders in the research
community, such as the Research Center Dresden-Rossendorf, Institut of
Safety Research (FZD) in Dresden, Germany; University of Technology,
Munich (TUM), Germany; University of Applied Sciences Zittau-Grlitz,
ETHZ, Zrich, Switzerland; and KTH Royal Institute of Technology,
Stockholm, Sweden.
ANSYS has also collaborated with well known researchers in the United
States including Professor Mike Dudukovic of the Chemical Reactions
Engineering Laboratory (CREL) at Washington University in St. Louis
and Professor Rodney Fox of Iowa State University. This work has not
only led to the enhancement of physical models within the ANSYS
multiphase solver, but has resulted in widely cited publications
on bubble column validation [4], stirred tanks, liquidsolid risers
and implementation of moment methods in the population
balance model [5].
ANSYS multiphase application specialists play an important role
in bridging the gap between customers and developers as well as
encouraging customers to use simulation best practices. This interaction
enables ANSYS to transfer new technologies into the latest software
releases. Successful application of multiphase technology depends on
a wide range of factors from understanding numerics to physical
model implementation. The ANSYS multiphase team makes every effort
to ensure that the quality of the implementation meets the highest
standard, providing customers with reliable, proven results.
ANSYS offers a robust, advanced and scalable set of multiphase ow
models to effectively address a broad spectrum of needs.
References
[1] Nisisako, T.; Torri, T.; Higuchi, T. Droplet Formation in a
Micro-Channel Network. Lab Chip, 2002, Vol 2, No 1, pp. 24-26.
[2] Vasquez, S.; Ivanov, V. A Phase Coupled Method for Solving Multiphase Problems
on Unstructured Meshes. Proc. of ASME FEDSM 2000: ASME 2000 Fluids
Engineering Division Summer Meeting, Boston, Massachusetts, June 1115, 2000.
[3] Guet, S. Bubble Size Effect on the Gas Lift Technique. Ph.D. Thesis, Delft University
of Technology.
[4] Sanyal, J.; Sergio Vsquez, Roy, S:, Dudukovic, M.P. Numerical Simulation of Gas
Liquid Dynamics in Cylindrical Bubble Column Reactors. Chemical Engineering
Science, 1999, Vol. 54, Issue 21, pp. 50715083.
[5] Sanyal, J.; Marchisio,D.L.; Fox, R.O.; Dhanasekharan, K. On the Comparison
Between Population Balance Models for CFD Simulation of Bubble Columns.
Ind. Eng. Chem. Res., 2005, Vol. 44, Issue 14, pp. 50635072.
Industry Solutions
ANSYS, Inc.
Southpointe
275 Technology Drive
Canonsburg, PA 15317
U.S.A.
724.746.3304
ansysinfo@ansys.com
Toll Free U.S.A./Canada:
1.866.267.9724
Toll Free Mexico:
001.866.267.9724
Europe:
44.870.010.4456
eu.sales@ansys.com
ANSYS, ANSYS Workbench, Ansoft, AUTODYN, CFX, FLUENT and any and all ANSYS, Inc. brand, product,
service and feature names, logos and slogans are registered trademarks or trademarks of ANSYS, Inc. or its
subsidiaries in the United States or other countries. ICEM CFD is a trademark used under license. All other
brand, product, service and feature names or trademarks are the property of their respective owners.
2011 ANSYS, Inc. All Rights Reserved. Printed in U.S.A. MKT0000075 04-11
ansys.com
Simulation of cavitation around a rotating
propeller compared with actual conditions
Courtesy University of Tokyo.
Flow of catalyst particles in a uid catalytic
cracker as predicted by the DDPM model
in ANSYS FLUENT. This model can capture
the presence of clusters and streaks in a
relatively coarse mesh.
Pressure drop in a uid catalytic cracker riser
is predicted by the DDPM model and
compared with experiment. The pressure
drop in a uidized bed is directly correlated
to the holdup of particles.

You might also like