Group Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Title: 9.3.2 - Ultrafast Multiphase Flow Imaging by Electron Beam X-ray Computed Tomography
ALL VOLUMES CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102023/00223
 Material Information
Title: 9.3.2 - Ultrafast Multiphase Flow Imaging by Electron Beam X-ray Computed Tomography Experimental Methods for Multiphase Flows
Series Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Physical Description: Conference Papers
Creator: Bieberle, M.
Fischer, F.
Schleicher, E.
Franke, M.
Menz, H.-J.
Mayer, H.-G.
Laurien, E.
Hampel, U.
Publisher: International Conference on Multiphase Flow (ICMF)
Publication Date: June 4, 2010
 Subjects
Subject: X-ray
computed tomography
ultrafast
multiphase flow
imaging
 Notes
Abstract: Ultrafast electron beam X-ray computed tomography (CT) is an imaging technique (Bieberle & Hampel., 2006), which is able to recover the cross-sectional density distribution of multiphase flows with a frame rate of up to 10,000 fps and a spatial resolution of about 1 mm. Originally, electron beam CT was developed for cardiac imaging by Boyd (1983) and modern medical systems reach frame rates of about 20 fps. During the last years, this measurement technique has been advanced and qualified for flow imaging by the Forschungszentrum Dresden-Rossendorf and the University of Stuttgart. Its applicability to different two-phase flows has been demonstrated in a number of experiments (Bieberle et al., 2007, Bieberle et al., 2009). Specifically adapted image reconstruction methods allow to extract the phase boundaries and thus to determine the phase fractions within the tomography plane. The latest developments in ultrafast electron beam X-ray CT include the extension towards two or more tomography planes which enable furthermore to measure phase velocities by using cross correlation techniques. This in turn is essential for determining bubble volumes and volumetric flow rates, which are important parameters for the validation of CFD codes.
General Note: The International Conference on Multiphase Flow (ICMF) first was held in Tsukuba, Japan in 1991 and the second ICMF took place in Kyoto, Japan in 1995. During this conference, it was decided to establish an International Governing Board which oversees the major aspects of the conference and makes decisions about future conference locations. Due to the great importance of the field, it was furthermore decided to hold the conference every three years successively in Asia including Australia, Europe including Africa, Russia and the Near East and America. Hence, ICMF 1998 was held in Lyon, France, ICMF 2001 in New Orleans, USA, ICMF 2004 in Yokohama, Japan, and ICMF 2007 in Leipzig, Germany. ICMF-2010 is devoted to all aspects of Multiphase Flow. Researchers from all over the world gathered in order to introduce their recent advances in the field and thereby promote the exchange of new ideas, results and techniques. The conference is a key event in Multiphase Flow and supports the advancement of science in this very important field. The major research topics relevant for the conference are as follows: Bio-Fluid Dynamics; Boiling; Bubbly Flows; Cavitation; Colloidal and Suspension Dynamics; Collision, Agglomeration and Breakup; Computational Techniques for Multiphase Flows; Droplet Flows; Environmental and Geophysical Flows; Experimental Methods for Multiphase Flows; Fluidized and Circulating Fluidized Beds; Fluid Structure Interactions; Granular Media; Industrial Applications; Instabilities; Interfacial Flows; Micro and Nano-Scale Multiphase Flows; Microgravity in Two-Phase Flow; Multiphase Flows with Heat and Mass Transfer; Non-Newtonian Multiphase Flows; Particle-Laden Flows; Particle, Bubble and Drop Dynamics; Reactive Multiphase Flows
 Record Information
Bibliographic ID: UF00102023
Volume ID: VID00223
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: 932-Bieberle-ICMF2010.pdf

Full Text

Paper No 7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010


Ultrafast Multiphase Flow Imaging by Electron Beam X-ray Computed Tomography

Martina Bieberle, Frank Fischer, Eckhard Schleicher, Martin Franke, Hans-Jorgen Menz,
Hans-Georg Mayer, Eckart Laurien and Uwe Hampel

Forschungszentrum Dresden-Rossendorf e.V., Institute of Safety Research
01328 Dresden, Germany
M.Bieberle@fzd.de

Stuttgart University, Institute of Nuclear Technology and Energy Systems
70569 Stuttgart, Germany



Keywords: X-ray, computed tomography, ultrafast, multiphase flow, imaging




Abstract

Ultrafast electron beam X-ray computed tomography (CT) is an imaging technique (Bieberle & Hampel., 2006), which is able
to recover the cross-sectional density distribution of multiphase flows with a frame rate of up to 10,000 fps and a spatial
resolution of about 1 mm. Originally, electron beam CT was developed for cardiac imaging by Boyd (1983) and modern
medical systems reach frame rates of about 20 fps. During the last years, this measurement technique has been advanced and
qualified for flow imaging by the Forschungszentrum Dresden-Rossendorf and the University of Stuttgart. Its applicability to
different two-phase flows has been demonstrated in a number of experiments (Bieberle et al., 2007, Bieberle et al., 2009).
Specifically adapted image reconstruction methods allow to extract the phase boundaries and thus to determine the phase
fractions within the tomography plane. The latest developments in ultrafast electron beam X-ray CT include the extension
towards two or more tomography planes which enable furthermore to measure phase velocities by using cross correlation
techniques. This in turn is essential for determining bubble volumes and volumetric flow rates, which are important parameters
for the validation of CFD codes.


Introduction

Multiphase flows are present in many industrial plants and
they mainly influence the safety and efficiency of the
underlying physical and chemical processes. In order to
understand and optimize those processes considerable effort
is being spent in the modelling of multiphase flows.
Contrary to single phase flows, computational fluid
dynamics (CFD) codes for multiphase flows are still not
mature for many applications. However, the development of
such CFD codes is fostered by the increasing computational
power available today. Hence, there is also a great interest in
multiphase flow measurement technologies, which can
provide detailed and accurate validation data for CFD codes.
Although many imaging modalities, such as optical and
acoustic methods, are well developed for single phase flow,
they are hardly applicable to multiphase flows. Problems
here are refracting phase boundaries and opaque media.
Many tomographic techniques have been proposed as an
alternative. However, many of the classical tomographic
imaging modalities, as known from medicine and
nondestructive testing, cannot reach high spatial and
temporal resolution at the same time.
Computed tomography (CT) methods based on either


ionising radiation or magnetic resonance have long been
considered as incapable to capture fast flows. However, in
recent years some attempts have been made to adapt such
techniques to the needs of flow measurement. For example,
a fast gamma ray modality was introduced by Johansen et al.
(1996), who built a scanner with five stationary Am-241
sources, which achieves frame rates of up to 100 Hz.
Misawa et al. (2003) and Hori et al. (2000) built fast
multi-tube X-ray scanners and achieved frame rates of up to
2 kHz. An alternative is electron beam tomography, which
was first introduced by Boyd and Lipton (1983) for cardiac
imaging. An electron beam scanner does not employ the
classical CT principle of a rotating source-detector setup but
rather uses an inertia-free electron beam, which is focused
onto a tungsten target and swept across this target very
rapidly by means of an electromagnetic deflection system.
At the position of beam impingement, X-ray radiation is
produced within the focal spot. While medical devices are
limited to frame rates of about 20 Hz, which is sufficient to
image the beating heart, the principle can also be adapted
for flow imaging. This has successfully been demonstrated
by our group who developed and applied a much faster
electron beam X-ray CT modality (Hampel et al., 2005) for
this purpose.






Paper No


Nomenclature

a pixel weight
d distance
I X-ray intensity (Wmn2)
J superficial velocity
N number of pixels (in each direction)
p X-ray attenuation
t time (s)
v velocity

Greek letters
A relaxation factor
y linear X-ray attenuation coefficient (mn1)
r correlation coefficient


Subsripts
d detector element index
D detector
G gas
L liquid
m, n pixel indices
ref reference
s source index
S source
TP two-phase


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

object, which in turn is necessary for accurate image
reconstruction. However, with focal spot path and detector
lying in one plane, radiographic projection cannot be
acquired from all angles, which leads to the so called
limited-angle-problem. Nevertheless, adequate image
reconstruction algorithms adapted from the algebraic
reconstruction technique (ART) (Gordon et al. 1970)
provide an acceptable image quality for recovering the
phase distribution of two-phase flow experiments.


electron gun

electron beam

focussing coil


deflection coils


detector arc.


focal spot path
., tungsten target


Electron beam X-ray CT

Setup
The principle of electron beam X-ray CT is illustrated in
Figure 1. The electron beam is generated by an electron gun
with an acceleration voltage of 150 kV and a beam current
between 5 and 10 mA. The gun is followed by an
electromagnetic lens system for beam centring, focussing
and deflection. The electron beam is focused on a tungsten
target to a focal spot with an effective diameter of 300 pm.
The scan is performed linearly on the target with a
deflection width of 160 mm and a frequency of 2.5 kHz,
giving 5000 sweeps and the same number of images per
second. Opposite to the target a semicircular X-ray detector
is mounted with the detector pixels at the same height as the
path of the focal spot. The detector consists of 256
cadmium-zinc-telluride (CZT) semiconductor detector
elements, which are operated in current mode. The detector
is sampled at 1 MHz, which gives 200 detector readings per
electron beam sweep across the target.
The whole assembly is arranged in a commercial electron
beam processing box, which provides a flexible setup, but
requires components capable for operation in vacuum and
vacuum-tight experiments. Another limitation is the
maximum available height in the box which is about 50 cm
and does not enable to study larger vertical objects, such as
pipes.
The idea to perform a linear scan rather than a circular scan
originated from the intention to prevent an axial offset
between focal spot path and detector. This offset arises from
the angular overlap of a static massive target and a static
detector when trying to fully sample the Radon space of an


Figure 1: Principle of ultrafast electron beam X-ray CT.

Modifications of the basic electron beam X-ray CT setup
are used to gain further information about the two-phase
flow. In order to also determine phase velocities, a
two-plane arrangement has been set up. It is realized by
alternately scanning two focal spot paths in different heights.
Thus, the gas phase velocity can be extracted by correlating
the reconstructed phase distributions from the two resulting
tomography planes.

Data processing
In each electron beam sweep the focal spot passes once
across the target. On its continuous path from one turning
point to the other Ns data samples are taken synchronously
at ND = 256 detector elements. The number of temporal
samples per sweep depends on the scanning frequency and
was Ns = 200 in this study. Beside the X-ray intensities
IT~P = ITP)j for the two-phase flow of interest, a
,d (dark) (ref)
so-called dark scan I and a reference scan I are
acquired. The dark scan data represents the detector
readings without X-ray exposure, i. e. the baseline
detector output. The reference scan data is taken from the
flow channel completely filled with one of the phases as a
reference state. This way, for each sweep at time t a
projection data set Pt = {P,,d is generated according to


1 If the denser phase is used as a reference, the sign
changes in equation (1).






Paper No


(TP) (dark)
I -Id
P d = log (re) (dark) (1)
s,d s,d

which represents the integral attenuation values of all
possible X-ray paths from a source to a detector position.
The forward problem of the limited-angle electron beam
tomography is given by the equation system


P,,d .m, s,d,m,n '
m=l n=l


wherein NxN is the number of pixels, p is the
average X-ray attenuation coefficient for pixel (m, n) and
a,dm is the so called pixel weight, i.e. the share of pixel
(m, n) with the area of ray (s, d).
The problem of image reconstruction is solved using the
ART algorithm, which is defined as an iterative updating
scheme according to

N N
P ^-Via ^.,nZn^.n
Psd as ,d,m,n n
1 a=y +2...... ad ,n (3)
,,d Y^ adm,,
Inn m I


wherein A is a relaxation factor that regulates the speed of
convergence. In our case it has been set to 0.1 at 10
iterations per reconstruction, which was determined as an
optimum value for this problem. As a convergence stabilizer,
the positivity constraint was applied, i.e. only physically
plausible positive attenuation coefficient values p were
allowed during reconstruction. Image reconstruction was
followed by extraction of the phase distribution including a
Gaussian filter operation and the binarization using a
threshold value at half of the attenuation coefficient of the
liquid or solid phase respectively. The reconstructed
cross-sectional images from all time steps are then


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

combined to three-dimensional data sets p = {p }.

Velocity evaluation
From experiments with two tomography planes and thus
two reconstructed phase distribution data sets up""p) and
(lor), velocity information can be extracted by applying
cross-correlation techniques. The temporal offset between
the two sequences can be determined as the maximum of the
cross correlation function


S= max 1- T (Iower) (upper)
t= 10


In combination with the known distance d, between the
tomography planes, gas phase velocities are calculated
according to


Experiments

Water-air two-phase flow
The performance of the ultrafast limited-angle
electron-beam X-ray CT has been demonstrated with an
experimental water-air two-phase flow loop. Figure 2
illustrates the experimental setup with the specially
designed flow loop operated inside the vacuum box. Water
is continuously pumped at a defined flow rate through a
short vertical test section of a 40 mm diameter pipe. After
passage of a flow straightener, gas is injected via a
single-hole cannula. After 22 mm this two-phase flow
passes the X-ray tomography plane and goes off the vacuum
box into a separator. Due to the small distance between the
gas inlet and the tomography plane, the two-phase flow is
not fully developed in the imaged cross-section.


gas


flow straightened


electron
beam









focal
spot path


tungsten
target


Figure 2: Experimental two-phase flow loop setup.


X-ray
detector T






Paper No


Different flow regimes were generated by varying the liquid
and the gas flow rate. For electron beam X-ray CT a beam
current of 6 mA was used and scan sequences of 0.4 s
duration were acquired for each of the parameter
combinations. Selected axial cuts of the observed flow
patterns with constant liquid flow rate on one hand and
constant gas flow rate on the other hand are given in Figure
3. The results show, that the ultrafast X-ray CT is able to
image different structures of gas-liquid two-phase flows
with high temporal and spatial resolution and without
significant motion or limited-angle artefacts. Figure 4
illustrates the progress of the phase boundary of one of the


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010


flow patterns in two different tomography planes in 3D. As
can be seen, the shape of the flow slightly changes between
the two planes which are 10 mm apart from each other. The
differences are due to the undeveloped flow pattern at this
short distance from the gas inlet. Nevertheless, most of the
bubbles can clearly be identified in both planes. The
three-dimensional illustration furthermore substantiates the
fact, that electron beam X-ray CT reveals contrary to
high-speed videometry non-superimposed phase
distributions of transient two-phase flows at comparable
frame rates.


A











AL


0.4 s si-
superficial gas 0.07 m/s 0.13 m/s 0.27 m/s 0.40 m/s 0.13 m/s
velocity
superficial liquid 0.53 m/s 0.13 m/s 0.27 m/s 0.53 m/s
velocity
Figure 3: Axial cuts of the reconstructed phase distributions at different gas and liquid flow rates.


0.00, 0.00.

0.05s 0.05-

0.10, 0.10

0.15 0.15

S0.20, 0.20

0.25, 0.25

0.30 0.30

0.35, 0.35

0.40 1 0.40
25 25 25 25
-2525 5 -25 -25
y/mm x/mm y/mm x/mm
Figure 4: Three-dimensional plots of the development of
the phase boundary within a lower (left) and an upper (right)
cross-section of a water-air two-phase flow
( J, = 0.13m/s, J, = 0.27m/s).

From the two-plane measurements, velocity fields of the gas
phase have been determined. The results in Figure 5 show


the temporal offset between the tomography planes as well
as the velocity map. Regions, where no gas phase was
present, are marked in black. The mean velocity within the
gas phase region is v = 1.1 m/s, which matches the
expected magnitude. Although the variance of the
determined values is quite high, which is caused by the
undeveloped flow and the small amount of data, the results
still demonstrate the applicability of the evaluation
technique. Further measurements of fully developed
two-phase flows are necessary to quantify the accuracy of
this method.

2 20 2,0
15
1 0 10* 1,5

0 1,0

-1 *5 -lO0 0,5
*ms 1 M/
-20 0 -20 0,0
-20 -10 0 10 20 -20 -10 0 10 20
a) z / mm b) / mm
Figure 5: Temporal offset (a) and velocity map (b) of the
gas phase of a water-air two-phase flow (J, = 0.13m/s,
JL = 0.27m/s).


-A









Ah





A
41111111,
P


-9


i





7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010


Fluidized bed
As an example for gas-solid two-phase flow an
experimental fluidized bed with tube diameter 40 mm was
investigated. The particle bed consisted of monodisperse
glass beads of diameter 0.9 mm, 1.7 mm and 3.0 mm,
respectively. As shown in Figure 6, the particle bed was
fluidized through feeding gas in from the bottom of the bed.
The volumetric gas flow rate was varied between 250 1/min
and 2000 1/min. Electron beam X-ray CT scanning was
performed at a beam current of 8 mA. In the scanning time
of one second, projection data for 5000 cross-sectional
images were gathered for each configuration.


gas bubbles







gas inlet
Figure 6: Schematic of fluidized bed setup.


0.00 s ,,




0.25 s. |










0.75 gass
3.0 mm ii-




1.0 s r, '




a) solid b) 250 1/min 500 1/min 750 1/min 1000 1/min
Figure 7: Reconstructed phase distributions of the gas-solid two-phase flow in a fluidized bed: a) cross-sectional images for
different particle sizes, b) axial cuts through the sequences with 1.7 mm beads at different gas flow rates.


In Figure 7a selected cross-sectional phase distributions for
different particle sizes are shown. Evidently, even the
smallest particles are resolved by the ultrafast electron beam
X-ray CT at full flow dynamics when they are loosely
distributed. This manifests the expected spatial resolution of
the system of about 1 mm. It gives the chance to study the
particle dynamics in detail which is to our knowledge
unprecedented. Figure 7b illustrates the progress of the
phase distribution along the diameter line through the
fluidized bed perpendicular to the target for the 1.7 mm
particles at different gas flow rates. Ultrafast electron beam
CT clearly reveals periods of quasi static distributions inside
the imaging plane from periods of higher particle dynamics
before bubble or slug formation, which cannot be seen with
imaging techniques of lower temporal or spatial resolution.
It can also be observed, that the size of bubbles increases
with higher gas flow rate whereas the bubble frequency
remains nearly constant.
Figure 8 finally shows the three-dimensional structure of the
phase boundary changes in the sequence of the 1.7 mm
spheres at a gas flow rate of 750 1/min.


0,2-

0,4

0,6 .





-20
-10 20
10 0 0
x/mm 20 -20 10y / mm


Figure 8: Three-dimensional plot of the fluctuating phase
distribution within a cross-section of a fluidized bed.


Paper No






7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010


Conclusions


Vol. 47(3), 369-78 (2009).


It has been shown in a series of experiments that ultrafast
electron beam X-ray CT is a suitable measurement
technique for imaging gas-liquid and gas-solid two-phase
flows with high temporal and high spatial resolution. The
frame rate of 5000 images per second in combination with
the spatial resolution of about 1 mm enables detailed studies
of phase structures and flow dynamics. Especially in the
example of a fluidized bed, a so far unprecedented insight
into the particle dynamics of an opaque particle bed has
been achieved, which opens up new possibilities for
studying and understanding gas-solid flows.

Acknowledgements

This work was supported by Deutsche
Forschungsgemeinschaft (DFG) grant no. HA 3088/3-1 and
grant no. KO 2942/3-1.

References

Bieberle, M. & Hampel, U. Evaluation of a limited angle
scanned electron beam X-ray approach for two-phase pipe
flows. Meas. Sci. Technol., Vol. 17, 2057-65 (2006).

Bieberle, M., Fischer, F., Schleicher, E., Hampel, U., Koch,
D., do Couto Aktay, K.S., Menz, H.-J. & Mayer, H.-G.
Ultrafast limited-angle-type x-ray tomography. Appl. Phys.
Lett., Vol. 91(12), 123516 (2007).

Bieberle, M., Fischer, F., Schleicher, E., Koch, D., Menz,
H.-J., Mayer, H.-G. & Hampel, U. Experimental two-phase
flow measurement using ultra fast limited-angle-type
electron beam X-ray computed tomography. Exp. Fluids.,


Bieberle, M., Fischer, F., Schleicher, E., Menz, H.-J., Mayer,
H.-G. & Hampel, U. Ultrafast cross-sectional imaging of
gas-particle flow in a fluidized bed. AIChE Journal (DOI
10.1002/aic.12121, online), (2010).

Boyd, D. P. & Lipton, M. J. Cardiac computed tomography.
Proc. IEEE, Vol. 71, 298-307 (1983).

Gordon, R., Bender, R. & Herman, G. T. Algebraic
reconstruction techniques (ART) for three-dimensional
electron microscopy and X-ray photography. J. Theor. Biol.,
Vol. 29, 471-481 (1970).

Hampel, U., Speck, M., Koch, D., Menz, H.-J., Mayer,
H.-G., Fietz, J., Hoppe, D., Schleicher, E., Zippe, C. &
Prasser, H.-M. Ultrafast X-ray computed tomography with a
linearly scanned electron beam source. Flow Meas. Instrum.,
Vol. 16, 65-72 (2005).

Hori, K., Fujimoto, T., Kawanashi, K. & Nishikawa, H.
Development of an ultra fast X-ray computed tomography
scanner system: application for measurement of
instantaneous void distribution of gas-liquid two-phase flow.
Heat Trans Asian Res, Vol. 29(3), 155-165 (2000).

Johansen, G. A., Fr0ystein, T., Hjertaker, B. T. & Olsen, 0.
A dual sensor flow imaging tomographic system. Meas. Sci.
Technol., Vol. 7, 297-307 (1996).

Misawa, M., Tiseanu, I., Prasser, H.-M., Ichikawa, N. &
Akai, M. Ultra-fast X-ray tomography for multi-phase flow
interface dynamic studies. Kerntechnik, Vol. 68, 85-90
(2003).


Paper No




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs