SESSION 12
MEASUREMENT AND INSTRUMENTATION
Chairmen: R. Semiat, M. Yamaguchi
CHARACTERIZATION OF DROPLETS IN LIQUID/LIQUIDEXTRACTION BY LASEROPTICAL MEASUREMENT
E. Hanno Schombacher, Thomas Wriedt, Klaus Bauckhage
University of Bremen
Department of Chemical Engineering/FBO4
Badgasteinerstr. 3
D28359 Bremen
Abstract
In the recent years increasing activities in the field of liquid/liquidExtraction (LLE)
Process modelization can be found in literature [1]. One prerequisite for further
success in this field of research is the determination of process data like droplet
size and velocity or mass transfer rates as a result of singledroplet experiments
or determined from technical apparatus. Within this study improved laseroptical
measurement techniques, PhaseDopplerAnemometry (PDA) and Rainbow
Refractometry (RBR), offering the ability of fast and reliable data acquisition will
be proposed.
PDA can be regarded as a well established device for the determination of size
and velocity of particles in twophase or multiphase flows such as liquid/gas,
solid/gas, solidlliquid and gas/liquid [2]. In standard applications the frequency f
of a Dopplerburst received at a photodetector can be related to droplet velocity
whereas the phasedifference measured between two Doppler bursts is
proportional to the droplet diameter. One basic step for the adaption of the
measurement technique to the specific conditions occuring in liquid/liquidsystems
is the optimization of the PhaseDoppler setup by light scattering considerations.
The systems under investigation were the recommended test systems
Toluene/water und Butyl Acetate/water [3].
Using the so found optimum setup, experiments at singledroplets of the
recommended test systems Toluene/water und Butyl Acetat/water [3] were
performed. PDA exhibited satisfactory capabilities in terms of determination of
singledroplet size and velocity. Additional simulations performed by an extended
theory of Geometrical Optics show sensibility of the measurement technique to
phenomena like droplet deformation, oscillation and acceleration. The theoretically
obtained conclusions could be consolidated by additional experiments.
A second laseroptical technique, RBR [4], allows the determination of droplet
composition within ternary twophase systems. Concentrations can be concluded
from the angular position of the primary monochromatic rainbow, which is sensitive to droplet refractive index.
In a first step, the performance of RBR in terms of determination of homogeneous
concentrations was investigated. For different equilibrium concentrations of the
recommended ternary testsystem Toluene/water/Acetone droplets of well defined
diameter were produced in order to detect from a linear CCDscan.
The instationary mass transfer process of a third component from or to droplets is
coming with the establishment of more or less distinctive concentration gradients
in both the droplet phase and bulkphase. These gradients are of influence on the
light scattering properties of the droplet and thus on the rainbow position
. To investigate this effect, a combined diffusional masstransfer and light scattering model is developed. From simulations can be concluded that the determination of concentration gradients by RBR under certain conditions is possible. This is underlined by masstransfer experiments to pendant singledroplets.
 Steiner, L.:
Drop Population Balance  Werkzeug oder Flop.
Chem.Ing.Tech. 68 (1996) 1 +2, p.141146
 Bauckhage, K.:
The PhaseDopplerDifferenceMethod, a New LaserDoppler Technique for
Simultaneous Size and Velocity Measurement  Part 1: Description of the Method.
Part. Part. Sys. Charact. 5 (1988), p.1622
 Misek, T.; Berger, R.; Schröter, J.:
Standard Test Systems for Liquid Extraction.
European Federation of Chemical Engineering GB, Rugby 1985
 Roth, N.; Frohn, A.:
Theoretical and experimental studies on the influence of internal temperature
gradients on rainbow refractometry.
PARTEC 95  4th International Congress Optical Particle Sizing, Nürnberg 1995
INTERFACIAL INSTABILITIES IN LIQUIDLIQUID SYSTEMS
S.Wolf, J.Stichlmair
Technical University of Munich
Dep. Chemical Engineering
A, Arcisstrasse 21, D80290
Muenchen, Germany
Tel. +49 89 28916505
Email: stefan.wolf@www.vta.mw.tumuenchen.de
Fax: +49 89 28916510
During mass transfer in liquidliquid systems, the transport of components across the interface
is accompanied by local variations of composition and temperature. If these small
impairments are enhanced, macroscopic interfacial convection, known as the Marangoni
effect, can occur. Depending on system properties and fluiddynamics, interfacial phenomena
in form of roll cells, eruptions or interfacial turbulences are observed. Measurements of mass
transfer rates show that these effects lead to higher mass transfer coefficients. Reasons of
these enhancements are a faster transport of the transferred components between the bulk
phases and the interface and a higher renewal rate of the interface.
The objective of the project is to quantify the influence of the Marangonieffect on mass
transfer rate and to include the shear stress at the interface for the development of criteria for
the prediction of the onset of the interfacial convection. The dependence of the observed
interfacial phenomena on parameters as mass transfer rate and bulk phase fluiddynamics will
be investigated.
The first part of the experiments was carried out in a diffusion cell where the stability
behavior of several liquidliquid systems was investigated. Marangoniconvection was
detected by using a holographic interferometer. Holographic interferometry is an optical
method based on the recording and reconstiuction of the distribution of phase and amplitude
of coherent lightwaves on a holographic plate. The experiments were carried out with drops
and with flat interfaces. Parameters, e.g. as investigated systems, direction of mass transfer,
concentration difference and flow rate of the phases, were varied during the investigations. It
has been shown that these parameters have a strong influence on the occurence and on the
form of the Marangoniconvection. Many systems showed interfacial instabilities in both
directions of mass transfer.
Additional concentration gradients can be established in the bulk phases by interfacial
instabilities. If density instabilities and Marangoniconvection occure at the same time, the
result will be chaotic concentration profiles across the whole measuring cell.
To obtain more data about interfacial convection, studies are made in a flowing channel with
the system water / acetone / toluene now. Measurements of the velocity profiles of the flowing
phases are accomplished by laserdoppleranemometer (LDA). The concentration gradients at
the inlet and the outlet are determined by a holographic interferometer.
During cocurrent flow of the phases, the intensity of the Marangoniconvection is the
strongest at the first contact of the phases since the concentration difference is the highest. At
the outlet, the instabilities are less violent because of the smaller concentration gradient and
may even disappear at low concentrations.
The analysis of the holographic interference pattern by a picture evaluation systems reveals
that interfacial instabilities lead to locally different mass transfer rates and to concentration
profiles that are totally different from those profiles predicted by the penetration model.
LASER GRATING VELOCIMETRY FOR LIQUIDLIQUID SYSTEMS
R. Semiat
Chemical Engineering Faculty,
Technion, Institute of Technology
Haifa 32000
The Laser Grating Technique (LGV) was developed for the measurement of large particles velocity and size in flow. The technique uses a single laser beam passing through a single or a double rotating optical grating. These allow a subdivision of the beam into a net of subbeams. An interaction with a particle in motion generates a set of refracted or reflected light beams that are turned into a burst type electronic signal with the use of a simple photodetector. The frequency of the photodetector signal, required for the evaluation of the local velocity, is determined by Fast Fourier Transform analysis. The rotation of two disk gratings at different speeds allows after electronic filtration, the detection of two directional velocity, using a single beam. The size of the particle may be calculated by the detected passage time of the particle through the main beam.
Measurements of the local rise velocity of large drops of 37 mm in diameter and the local interfacial velocity of a liquid jet in liquidliquid systems were successfully performed using the Laser Grating Velocimeter technique. The LGV was applied to the determination of the drop rise velocity distribution of a population of 300 drops and axial distribution of the jet interfacial velocity in order to demonstrate the capability and effectiveness of the technique.
FINITE ELEMENT ANALYSIS OF FULLYDEVELOPED TURBULENT TWOPHASE FLOW DISTRIBUTION IN TRIANGULAR CONDUITS
S. H. Hwang and G. C. Park
Department of Nuclear Engineering
Seoul National University
ShillimDong, GwanakGu,
Seoul 151742, Korea
S. K. Sim
Korea Atomic Energy Research Institute
P.O. Box 105, Yuseong, Taejon, 305600, Korea
Tel) +8228807210
Fax) +8228892688
Email) parkgc@plaza.snu.ac.kr
A threedimensional finite element computer code, FEMOTH, has been previously
developed in order to study the multidimensional flow patterns and the turbulent
structure of fully developed turbulent twophase flows. This code, which is a
timedependent, three dimensional, fully developed turbulent twophase flow analysis
code, was designed to investigate multidimensional phase distribution phenomena in
flow channels. In this code, however, due to a homogeneous secondary flow
assumption, the FEMOTH code predicted same streamlines for both phases. The
FEMOTH, thus, calculates flow patterns based on the initial void distribution
specified as initial guess rather than predicting void distribution in flow channels.
In order to analyze phase distribution mechanisms in flow channels, the
mathematical model has been upgraded by relaxing the homogeneous secondary flow
assumption. We derived the analytical model of turbulent twophase flow in
triangular ducts using the twofluid conservation equations of continuity and
momentum for fully developed, adiabatic, incompressible twocomponent twophase
flows. Therefore, we used two continuity equations for vapor and liquid, and six
momentum equations for x, y, zdirection for vapor and liquid; respectively. We
assumed "no slip" boundary conditions on all solid surface. We used the Prandtl
mixing length model of Drew derived for triangular conduit, to specify the turbulent
stress tensors. The variables that are associated with each triangular element are the
axial liquid and vapor velocities, the lateral liquid and vapor secondary velocities, the
void fraction and the static pressure.
Standard finite element approximations, based on the Galerkin formulation, were
used. Therefore no upwind finite element schemes are used. Rather the phasic
viscosities were artificially increased to ensure numerical stability. This is equivalent
to upwinding and allows the test functions to be identical to the shape functions.
FEMOTH code was tested for fully developed turbulent singlephase flows in
equilateral triangular duct. The FEMOTH code predicted three identical secondary
velocities in a half triangular duct due to the geometrical symmetry. It was also
found, after sensitivity studies, that turbulence is an important parameter for the
secondary flow fields.
FEMOTH code was then tested for fully developed turbulent twophase flow in
equilateral and isosceles triangular ducts. The FEMOTH code predicted qualitatively
similar secondary flow patterns for the turbulent twophase flow in the equilateral
triangular duct as in the corresponding singlephase flows. By varying parameters in
twophase turbulent model, it was demonstrated that the ratio of shearinduced to
bubbleinduced turbulence determines the structure of the secondary flow fields for
fully developed turbulent twophase flows in triangular duct. It was found that the
slip ratios between the phasic axial velocities are strongly dependent on the void
distribution.
FEMOTH code also predicted the phase distribution in the equilateral and isosceles
triangular ducts. It was found that turbulence and secondary flow patterns are
important flow parameters for the phase distribution in the channel. However, due to
the lack of twophase turbulence data on triangular ducts in the literature, the
FEMOTH predictions of turbulent twophase flow were qualitative than quantitative.
The FEMOTH code is undergoing further development for the upwinding schemes
and turbulent models.
