Owing to excellent
mixing, heat and mass transfer characteristics, gas-solid fluidized beds are
widely used to carry out gas-solid operations/reactions in power generation
(e.g. coal combustion/gasification), chemical process industry (e.g. fluid catalytic
cracking), etc. Several applications involve fluidization of more than one type
of solid material having different properties (size, shape, density, etc.) e.g.
co-gasification of coal and biomass, transport of coal and ash, etc.
Considerable research efforts have been made to develop CFD models to predict
the flow behavior and performance of fluidized bed reactors for unary gas-solid
flows. However, past experimental and numerical investigations (using continuum
models) of gas-solid flows have pre-dominantly comprised of a particular
(unary) type of solid material/particles. Though the fluidization of
laboratory-/large-scale binary gas-solid flow is important, it received less
attention (notable exceptions being the works of Goldschmidt et al., 2003; Rao
et al., 2011; Gidaspow et
al., 2013). The present work is carried out to quantify the flow regimes and
dynamics of unary and binary gas-solid flows using time-resolved electrical
capacitance tomography (ECT) measurements, and to simulate these flows using
open source CFD code OpenFOAM and to validate the
predictions.

Figure 1: Electrical Capacitance Tomography facility at IIT Delhi.

In the present work, the
ECT measurements were performed under cold flow conditions in a cylindrical
column of 9 cm ID, 1.5 m height fitted with a perforated sieve plate gas
distributor (with 7.9% free area), as shown in Figure 1. Experiments were
performed under ambient condition using air as the gas phase and glass beads
(GB) (d_p=530 µm, ρ_p=2600
kg/m³) and mustard seeds (MS) (d_p=1500
µm, ρ_p=720 kg/m³), or
mixtures thereof as the solid phase. Further details of the experimental setup, gas distributors, material
preparation, ECT sensors, sensor calibration and data processing will be
provided in the full length manuscript. Continuum (Euler-Euler) simulations of
unary and binary gas-solid flows were performed using the open source CFD code OpenFOAM. The ?twoPhaseEulerFoam?
model of the OpenFOAM was modified to implement
appropriate closure laws. Further details of solver modification, boundary
conditions and other numerics will be discussed in the full length manuscript.

The ECT measurements were performed for gas
superficial velocities (U_G) in the range of 0.2-3.0 m/s (that
exhibits single/multiple bubbling and slugging regimes) with a frame
acquisition speed of 33 frames/s. The time-averaged solids volume fraction
profiles along the diameter of the column for unary beds of GB and MS at U_G
of 0.87 and 2.62 m/s are shown in Figure 2. At U_G of 0.87 m/s
(bubbling regime), a considerable difference in the time-averaged solids volume
fraction profiles for GB and MS was observed. Also, at low gas velocities, the
time-averaged solids volume fraction profile for MS was not symmetric. From
instantaneous tomograms of solid volume fraction for GB and MS (not shown
here), it was seen that air was bubbling only through one half of the cross
section at U_G= 0.87 m/s and this led to asymmetric time-averaged
solid hold-up profile. However, for MS at U_G= 2.62 m/s and GB at U_G=
0.87 and 2.62 m/s, air was seen to bubble throughout the entire cross section
(tomograms not shown here) leading to symmetric profiles of time-averaged solid
volume fraction (see Figure 2). A comparison of solids volume fraction
fluctuations recorded at the center of the column for GB and MS is shown in
Figure 3. At U_G= 0.87 m/s, the bubbling frequency observed for GB at
the center of the fluidized bed was ~ 6.0 Hz (±0.2 Hz) and that for MS was ~
4.0 Hz (± 0.5 Hz)  (see Figure 3(a)). At U_G= 2.62 m/s, the
bubbling frequency of the GB and MS became ~ 4.0 Hz (± 0.4 Hz) and ~ 3.0 Hz (±
0.6 Hz) respectively (see Figure 3(b)).

Spatially averaged (on
the column cross section at 89 mm from sparger) volume fraction of the GB and
MS was plotted against the superficial velocity (U_G) and the flow
regimes were identified based on the changes in the slope of the profiles. For
GB, single/multiple bubbling regimes were observed for U_G in the
range of 0.25 to 0.87 m/s and slugging regime was observed for U_G >
0.87 m/s. For MS, the single/multiple bubbling regimes were observed between
0.5 and 1.5 m/s and slugging regime for U_G > 1.5 m/s. However,
the transition between single bubbling and multiple bubbling regime could not
be identified through the aforementioned plot. A plot of standard deviation of
the spatially averaged volume fraction plotted against the U_G showed
a sudden decrease in standard deviation at U_G ~ 0.6 m/s and 1.1 m/s
for GB and MS, respectively. This sudden decrease in standard deviation was
mainly due to the formation of multiple bubbles. A detailed comparison of flow
regimes and their transitional characteristics for unary and binary mixture
will be presented in the full manuscript.

Euler-Euler two-fluid
simulations were performed using OpenFOAM for GB and
MS. The measured and predicted time-averaged solid volume fraction profiles
were in a very good agreement for GB in the slugging regime (U_G=
1.75 m/s) (see Figure 4). However, in multiple bubbling
regime (U_G=0.87 m/s), the time-averaged solids volume
fraction was over-predicted, in particular at the column center (by ~11.6 %) in
comparison with the measurements. A qualitative comparison of measured and
simulated instantaneous solids volume fraction tomograms at different times
(marked by black points in Figure 6(a) and 6(b)) at 89 mm from the distributor
plate is shown in Figure 5. While the measured and simulated tomograms were in
a qualitative agreement, the predicted spatially averaged solid volume fraction
were ±6.3% and ±9.6% lesser than that obtained from the measured tomograms at U_G
of 0.87 and 1.75 m/s, respectively.

A comparison of predicted
and measured local (recorded at the center of the column) solids volume
fraction fluctuations for GB at U_G of 0.87 and 1.75 m/s is shown in
Figure 6 (a) and (b), respectively. For U_G =0.87 m/s, the
experimentally observed multiple bubbling regime is well predicted in the simulations
and that the measured ~ 6.0 Hz (± 0.2 Hz) and predicted ~ 5.0 Hz (± 0.8 Hz)
bubble frequencies were in a quantitative agreement. Also, the experimentally
observed slugging behavior at U_G= 1.75 m/s and the corresponding
measured (~ 5.0 Hz (± 0.4 Hz)) and predicted (~ 5.0 Hz (± 0.2 Hz)) frequencies
were in a good agreement (see Figure 6(b)). The existing ?twoPhaseEulerFoam? of OpenFOAM is being modified further to implement multi-fluid model that will allow the simulations of
binary mixtures with size dependent closure laws. Simulations of binary
mixtures and their comparisons with experiments and unary systems in terms of
time-averaged solid volume fraction distribution, instantaneous solid volume
fraction fluctuations and dynamics characteristics will be presented in the
full length manuscript. An analysis in frequency domain will be presented for
establishing the dominant frequencies of the volume fraction fluctuations as
one changes the composition of the bed inventory, and flow conditions. 

The time-resolved ECT measurements of solid volume
fraction performed in the present work for unary and binary gas-solid flows,
experimentally verified OpenFOAM based CFD code will
be important to understand the dynamics of binary gas-solid flows. This work
will be a step forward in moving towards binary fluidization of coal-biomass,
coal-ash mixtures relevant to gasification and power generation processes.

References:  

Gidaspow, D., Chaiwang, P.,
(2013). Bubble free fluidization of a binary mixture of large particles.
Chemical Engineering Science 97, 152?161.

Goldschmidt, M.J.V., Link, J.M., Mellema,
S., Kuipers, J.A.M., (2003). Digital image analysis
measurements of bed expansion and segregation dynamics in dense gas-fluidized
beds. Powder Technology 138, 135 ? 159.

Rao, A., Curtis, J.S., Hancock, B.C., Wassgren, C., (2011). Classifying the fluidization and
segregation behavior of binary mixtures using particle size and density ratios.
Particle Technology and Fluidization 57, 1446-1458.