(602f) Using Complimentary Imaging Techniques: X-Ray and MRI Studies of a Bubbling Gas-Solid Fluidized Bed

Introduction

Gas-solid
fluidized systems are difficult to measure non-invasively, due to their optical
opacity. Advances in instrumentation and computing power have lead to the rise
of tomographic techniques, such as Magnetic Resonance Imaging (MRI) (Holland et al., 2008),
Electrical Capacitance Tomography (ECT) (Du et al.,
2005) and X-ray Computed Tomography (XCT) (Mudde, 2010), for observing fluidization phenomena in 3D systems. An
ideal measurement technique would allow multiscale, non-invasive imaging to a
high spatial and temporal resolution, and be suitable for on-line use in
industrial applications. Since no single technique has all of these
capabilities, it is necessary to use complimentary techniques. In order to use
multiple measurement methods, the two techniques must first be cross-validated.
The complimentary characteristics of the techniques can be used to gain more
information about the system than would have been possible if each technique
was used in isolation. This study aims to cross-validate MRI and high-speed
X-ray imaging and investigate the use of these complimentary techniques for
measurements in a bubbling gas-solid fluidised bed.

Experimental

Fluidization Set-Up: The bed
was 50 mm I.D., provided with either a) a single orifice (orifice
diameter, d_o, of 2 mm) or
b) a 10-hole or 14-hole multi-hole (orifice diameter, d_o, of 1 mm) distributor. The distributors were designed
to ensure that there were no jet-wall interactions. The particles were poppy
seeds (0.5 mm diameter, 200 mm unfluidised bed height) fluidised with air
at a superficial gas velocity of 60 - 300 mm/s, as measured at 293 K and 1 bar.
The particles were classified as Group B using the Geldart particle groupings (Geldart, 1973).

MR Imaging: Experiments were conducted in a 7.1 T (300 MHz ¹H frequency) magnet equipped with a Bruker AV3 spectrometer. The
actively shielded gradient system was capable of producing a maximum gradient
strength of 77.7 G/cm. A birdcage radiofrequency (r.f.) coil (66 mm i.d.) was
used to excite and detect the ¹H nuclei within the
poppy seeds. The region above the distributor was imaged using 3D time-averaged
and 2D ultra-fast imaging sequences.

X-ray Imaging:
Experiments were conducted using the high power pulsed X-ray facility at
University College London (capable of producing X-rays at 150kV and 450mA, with
a pulse length of 0.2ms, and a frame capture rate of 72 fps). For this work,
X-rays were generated with 38 kV for the single jet system and 41kV for the
multiorifice system, and a current of 400 mA. The X-ray pulses were 500 µs long
and images were acquired at 72 frames per second.

Results and Discussion

The two
techniques (fast X-ray imaging and MRI) were cross-validated by measuring the
length of a single jet formed at a 2mm orifice. Figure 1 shows the length of a single jet with
increasing superficial gas velocity for X-ray and MRI measurements. Good
agreement was seen between the jet lengths measured using the two techniques.

Figure 1 Jet length as a function of
superficial gas velocity for a single jet from a 2 mm orifice in a bed of Group
B particles.

Figure
2 shows consecutive
images of the bubbling bed fitted with the 10-hole distributor acquired using
X-ray and MRI. The bed was fluidised at a superficial gas flow rate of 0.21m/s.
X-ray imaging is advantageous as it can image larger regions and therefore it
is possible to image the gas entering as jets and the subsequent formation,
coalescence and rising of bubbles in the bed. MRI cannot image as large a
region, but it does allow slices of the bed to be imaged non-invasively,
thereby providing more detailed information on the gas jets formed at the
distributor and the bubble formation at the top of the jets. It is not possible
to observe bubbles detaching from the tips of the jets in the X-ray images as
the images are a projection through the bed and due to the small size of the
bubbles formed, the contrast in the images is low in this region. By combining
measurements from both X-ray and MRI measurements it is possible to resolve
fluidization phenomena in both the grid region and higher in the bed.

Figure 2 Fast imaging of a bubbling bed
fitted with a 10-hole distributor (superficial gas flow rate, U_s, of
0.21 m/s) using a) X-ray and b) MRI imaging. The separation time between frames
was 13.8 and 25 ms respectively for the X-ray and MR images. The X-ray images
are of the bed projected along the y-direction. The MR images are of the centre
slice of the bed. The MR images show clearly the individual jets and bubble
formations from the tips of the jets, which cannot be seen from the projected
X-ray images. The gray-scale is inverted between the two sets of images as the
particles absorb X-rays, but give an MR signal. Therefore dark regions in the
X-ray images and white regions in the MR images indicate a high particle
density.

A time series of
projected X-ray data is shown in Figure
3. Images were
projected along the y-direction to
more clearly identify the trajectory of bubbles. High signal intensity
indicates the presence of a high voidage region i.e. a bubble or a jet. Bubble
rise velocities, bubble frequencies and bubble coalescence (labelled in Figure 3
as a, b and c respectively) can be observed.

Figure 3: Time series of X-ray data projected along the
x-direction for a bed fitted with a 10-hole distributor and fluidized at a superficial
gas flowrate of 0.21 m/s. Several features of a bubbling bed can be extracted:
a) bubble rise velocity can be calculated using the gradient of the trajectory
of the bubble, b) bubble frequencies can be calculated using Fourier Transform
analysis and c) bubble coalescence events can be observed. White regions
indicate trajectories of bubbles passing through the bed.

Bubble frequencies can
be calculated from X-ray data using a Fourier transform analysis of the
projected time series. Figure
4 shows bubble
frequencies as a function of height above the distributor for a 10-hole
distributor with a superficial gas flow rate of 0.21 m/s. A bubble frequency of
9.2 Hz occurred at approximately 50 mm above the distributor with a smaller
secondary peak at 11.8 Hz. The bubble frequency dropped to 3.5 Hz at 75 mm
above the distributor with a smaller secondary peak at 7.0 Hz, indicating that
either bubble coalescence or dispersion occurred in the region between 50 and
75 mm above the distributor, both of which can be observed visually in the
shaded region in Figure 3. Below a height of 50
mm above the distributor, it is not possible to distinguish bubble frequencies
due to the poor contrast of small bubbles in the X-ray images. However, a
similar analysis can be conducted on the MR images to calculate bubble
frequencies at lower heights and thereby obtain a full description of bubble
frequencies in the bed using both techniques in conjunction. This paper will
consider quantitative methods for calculating bubble rise velocities, bubble
frequencies, bubble coalescence rates and permanent jet heights from MRI and
X-ray data. The X-ray analysis will also be extended in a scale-up study of a
140 mm diameter bed.

Figure 4 Bubble frequency as a function of height above the
distributor for a bed fitted with a 10-hole distributor and fluidised with a
superficial gas flow rate of 0.21 m/s. Peaks are seen at 9.2 Hz at 50 mm (with
a secondary peak at 11.8 Hz) and at 3.5 Hz a at 75 mm and higher (with a
secondary peak at 7.0 Hz). Due to the poor contrast of small bubbles in X-ray
images, it is not possible to resolve bubble frequencies at heights below 50 mm
using X-ray data.

Conclusions

This
paper demonstrates the use of two complimentary imaging techniques in
measurements of fluidized beds. MRI provides high spatial and temporal
resolution data, but is limited in the height of the bed that can be measured.
Conversely, fast X-ray imaging can image larger beds to a high-temporal
resolution, but is limited in spatial resolution near the distributor. The two
techniques can therefore be combined to observe a wider range of length scales
than could be observed when using both techniques in isolation.

Acknowledgements

This work
is funded by the EPSRC, Grant EP/F041772/1.

References

Du, B.,
Warsito, W., Fan, L.-S., 2005. ECT Studies of Gas-Solid Fluidized Beds of
Different Diameters. Ind. Eng. Chem. Res. 44, 5020–5030.

Geldart,
D., 1973. Types of gas fluidization. Powder Technol. 7, 285–292.

Holland,
D., M?ller, C., Sederman, A., Mantle, M., Gladden, L., Davidson, J., 2008.
Magnetic resonance imaging of fluidized beds: Recent advances. Theor. Found.
Chem. Eng. 42, 469–478.

Mudde,
R.F., 2010. Time-resolved X-ray tomography of a fluidized bed. Powder Technol.
199, 55–59.