SLURRY BUBBLE COLUMN HYDRODYNAMICS

Tenth Quarterly Report

Budget Year 2: July 1 - September 30, 1997

Submitted to

Air Products and Chemicals

Contract #DOE-FC 22 95 PC 95051

Chemical Reaction Engineering Laboratory

Chemical Engineering Department

Washington University

1

SLURRY BUBBLE COLUMN HYDRODYNAMICS

Tenth Quarterly Report

Chemical Reaction Engineering Laboratory

Contract #DOE-FC 22 95 PC 95051

Budget Year 2: July 1 -September 30, 1997

TABLE OF CONTENTS

Page No.

Executive Summary / i
1. / Objective for the Second Budget Year / 1
2. / Liquid Recirculation Velocity Profile Measurements by
Computer Automated Radioactive Particle Tracking (CARPT)
for Air-Drakeoil System in an 18” Diameter Bubble Column
with Internals / 2
2.1 Outline of the Experiment / 2
2.2 Experimental Results and Discussion / 3
2.2.1 The Two-Dimensional Vector Velocity Profile / 3
2.2.2 The Radial Profiles of the Liquid Recirculation Velocity / 3
2.2.3 Comparison of the Results Obtained in Columns
with and without Internals / 4
2.3 References / 4
3. / Evaluation of the Dominant Terms in the Radial Momentum Balance Equation Using Gas Holdup Profiles Measured by CT and
Reynolds Stresses Measured by CARPT
3.1 Turbulent Drag Model
3.2 Experimental Validation
3.3 Discussion
3.4 Nomenclature
3.5 References

SLURRY BUBBLE COLUMN HYDRODYNAMICS

Tenth Quarterly Report

Chemical Reaction Engineering Laboratory

Contract #DOE-FC 22 95 PC 95051

Budget Year 2: July 1 -September 30, 19

EXECUTIVE SUMMARY

The main purpose of this subcontract from the Department of Energy via Air Products and Chemicals to the Chemical Reaction Engineering Laboratory (CREL) at Washington University is to study the fluid dynamics of bubble/slurry bubble columns and address issues related to scale-up and design.

The second budget year ended on September 30, 1997. All the objectives set for budget year 2 were completed and exceeded.

In this report, we summarize the research progress made during the tenth quarter (July 1-September 30, 1997), according to the objectives set for the second budget year. The accomplishments of this past quarter are as follows:

Computer Automated Radioactive Particle Tracking (CARPT) experiments, data filtering and data processing were completed for the time averaged velocity profiles. These experiments were conducted with the air-drakeoil system in a 44.0 cm (18”) diameter column with internals (16 tubes of 1 inch diameter) and with a perforated plate distributor (0.7 mm hole size and 0.076% open area) at superficial gas velocities of 2, 5 and 10 cm/s. The experimental conditions used are similar to the conditions employed in the same column without internals.

The time averaged axial liquid velocity profiles in the column with internals exhibit similar trends to those obtained in the column without internals. However, as the superficial gas velocity increases, the velocity inversion point seems to move radially inward (further from the wall), whereas in the column without internals the velocity inversion point occurs almost at the same radial position (i.e. same r/R). In the center of the column, the time averaged axial velocity in the column with internals is somewhat larger than that in the column without internals. Since the middle region of the column is represented by the smallest compartments for counting of the CARPT particle visits and, therefore, has the poorest statistics due to the small number of particle occurrences which increases the error in estimated velocity. Hence in the first approximation, in the range of the superficial gas velocities used, the internals do not seem to affect liquid recirculation much. However, in the column with internals, a larger distance from the distributor is needed for the flow to be developed. Turbulence parameters in columns with and without internals will be evaluated and reported.

Gas holdup measured by Computer Tomography (CT) and Reynolds stresses measured by CARPT are used to determine the dominant terms in the radial momentum equation. When the overall drag force is neglected, the comparison between the left hand side (LHS) and right hand side (RHS) of the radial momentum balance equation exhibits a systematic deviation close to the wall. As the superficial gas velocity increases, the discrepancy between the LHS and the RHS of the radial momentum equation becomes very large. When the turbulent drag force is included, by using the form proposed by Sannaes (1997) and Jakobsen (1991) and by assuming a gas velocity profile based on the measured time averaged liquid velocity profile and constant slip, the difference between the LHS and RHS of the radial momentum balance equation in the region between the mid-radius of the column and the wall is still large but is smaller compared to the difference obtained without accounting for the turbulent drag force. A low sensitivity of the radial momentum balance has been shown with respect to the shape and magnitude of the gas velocity profile. Investigations are in process in determining all the needed terms for the radial momentum balance.

1

SLURRY BUBBLE COLUMN HYDRODYNAMICS

Tenth Quarterly Report

Chemical Reaction Engineering Laboratory

Contract #DOE-FC 22 95 PC 95051

Budget Year 2: July 1 -September 30, 1997

1. Objectives for the Second Budget Year

The main goal of this subcontract from the Department of Energy via Air Products to the Chemical Reaction Engineering Laboratory (CREL) at Washington University is to study the fluid dynamics of slurry bubble columns and address issues related to scale-up and design. The objectives for the second budget year were set as follows:

1. Complete review of gamma ray tomography and densitometry for obtaining density profiles with emphasis on applications in the La Porte AFDU reactor.

2. Develop phenomenological models in interpretation of tracer runs at La Porte.

3. Extend the Computer Aided Radioactive Particle Tracking/Computed Tomography (CARPT/CT) data base.

4. Continue the evaluation of closure schemes for Computational Fluid Dynamic (CFD) codes. (It should be noted that this objective and objective No. 3 are coupled).

The second budget year ended on September 30, 1997. The goals set for the second budget year, and other accomplishments that exceed these goals, were completed successfully. The results and achievements were reported in the previous quarterly and topical reports.

In the following sections, the research progress and achievements accomplished in the tenth quarter (July 1 -September 30, 1997) are discussed.

2. Liquid Recirculation Velocity Profile Measurements Obtained by Computer Automated Radioactive Particle Tracking (CARPT) for Air-Drakeoil System in an 18” Diameter Bubble Column With Internals

The knowledge of liquid recirculation velocities and turbulent parameters in bubble columns are important for design and performance of bubble column reactors. Many studies have been reported on this subject (Chen and Fan, 1992; Menzel et al., 1990; Franz et al., 1984; Groen et al., 1996; Mudde et al, 1997; Devanathan, 1991; Devanathan et al., 1990; Degaleesan, 1997). However, most studies were limited to measurements at several single points or in a small part of the column. Only recently, the Computer Automated Radioactive Particle Tracking (CARPT) technique implemented in the Chemical Reaction Engineering Laboratory (CREL) makes it possible to measure the time averaged flow patterns in the whole column (Devanathan, 1991; Devanathan et al., 1990; Moslemian et al., 1992; Yang et al 1993; Limtrakul, 1996; Roy et al., 1997; Degaleesan, 1997). Even with CARPT, previous studies for liquid recirculation velocity were limited to small diameter columns with air-water systems and to columns without internals. Very few studies have been reported for large diameter bubble columns with more viscous liquids than water (Menzel et al., 1990). Moreover, time averaged liquid recirculation velocities in bubble columns with internals similar to those used in industrial scale units have not been reported. Hence, it is important to investigate the effect of internals on the liquid recirculation velocity. In the ninth quarterly report we discussed the results for the time averaged gas holdup distribution obtained using the air-drakeoil system in an 18” diameter column with internals. A comparison between the gas holdup measured in an 18” diameter column with internals and those measured in the same size column without internals were reported as well. The gas holdup in the column with internals has the same trend as that in the column without internals. The magnitude of the gas holdup in the column with internals is slightly higher compared to that without internals at low superficial gas velocities and almost the same at high gas superficial velocity.

In this work, we used our unique facility, Computer Automated Radioactive Particle Tracking (CARPT), to measure the time averaged liquid recirculation velocity in the 18” diameter column with internals for the air-drakeoil system. The internals were designed to simulate the heat exchanger tubes used in the AFDU reactor in La Porte, Texas which is also an 18” diameter vessel.

2.1 Outline of the Experiments

The experimental set-up is shown in Figure 2.1(a). The column is made of Plexiglas with a diameter of 18 inches and a height of 8 feet (L/D=5.3). The distributor used is a perforated plate with holes 0.7 mm in diameter and an open area of 0.076% as shown in Figure 2.1(b). The internals are composed of two bundles of 1 inch aluminum tubes. Each bundle has 8 equally distributed tubes. The configuration of the internals, shown in Figure 2.1(c), simulates the heat exchanger tubes in the 18” diameter slurry bubble column reactor used in La Porte, Texas. The details of the CARPT facility used in the investigation can be found elsewhere (Devanathan, 1991). The radioactive particle used as tracer was Sc 46 with activity of about 350 Ci. Thirty NaI scintillation detectors were employed in CARPT experiments. Data collection, at each gas superficial velocity, lasted about forty hours in order to get good statistical results. For comparison, the experiments were conducted at the same superficial gas velocities of 2, 5, 10 cm/s as used in the CARPT experiments for the same size column without internals. Figures 2.2 to 2.6 display selected time averaged liquid recirculation velocity for the air-drakeoil system in a column with internals and also show a comparison between the results and those obtained in a column without internals.

2.2 Experimental Results and Discussion

2.2.1 The Two-Dimensional Vector Velocity Profile

Kumar et al. (1995, 1997), among others, have shown that gas holdup in a bubble column is high in the center and low at the wall and this leads to gross liquid circulation throughout the column with liquid flowing up in the center and down near the wall. Figures 2.2(a) and (b) display typical two dimensional time averaged velocity vector plots at superficial gas velocities of 2.0 and 10.0 cm/s, respectively, for the column with internals, while Figures 2.3(a) and (b) display the liquid velocity profiles for the column without internals at the same gas velocity.

Figure 2.3 for a column without internals indicates an asymmetric flow pattern but basically a single liquid recirculation cell ,at both gas velocities, with liquid rising at the center and falling down by the walls. The asymmetric flow pattern, possibly caused by a non-uniformity of the distributor, is such more pronounced at low gas velocity (Figure 2.3(a)) and tends to become more uniform in churn turbulent flow (Figure 2.3(b)). The asymmetry of the liquid flow is caused by the asymmetry in the gas holdup distribution which was observed in CT scans. The cross sectional gas holdup distribution corresponding to Figure 2.3a and b are shown in Figures 2.5(a) and (b), respectively. Similar conclusions can be drawn regarding the column with internals for which the corresponding holdup distribution are shown in Figures 2.4 (a) and (b). A closer comparison of the liquid time averaged velocity vector plots for the columns with internals (Figure 2.2) and without internals (Figure 2.3) reveals that the non-uniformity of flow is more pronounced in the column with internals and that in the presence internals a larger distance is needed for establishing a region of, what looks like, fully developed flow. At high gas superficial velocity of 10 cm/s, the liquid velocity vectors in the column with and without internals become similar although the entrance effect is still much more pronounced in the column with internals. For that reason, for comparison purposes we will consider the data only in the section between 90 cm and 190 cm, i.e. roughly the top half of the column minus the very top disengagement zone.

The maldistribution of gas holdup and the asymmetry in recirculation velocity pattern of the liquid may also be caused by the large diameter of the column and small holes in the distributor, which are prone to plugging. This needs further investigation and interpretation and the findings will be reported when available.

2.2.2 The Radial Profiles of the Liquid Recirculation Velocity

Figures 2.6(a), (b) and (c) show the azimuthally and time averaged radial profiles of the axial and radial liquid velocity components, respectively, at different gas superficial velocities for the column with internals. Note that at 2.0 cm/s gas superficial velocity, it is meaningless to do azimuthal averaging for the axial velocity since the flow near one side of the column is upward and downward near the other side. Therefore, only the axial velocities at gas superficial velocities of 5.0 and 10.0 cm/s are plotted in Figure 2.6(a). It should be also mentioned that the axial and radial velocities were averaged only in the fully developed flow regions for each case ( about 90 cm to 190 cm of the column height). For comparison, the axial and radial velocity profiles in the column without internals are also shown in Figures 2.7(a) and (b). Comparison of Figure 2.6(a) and 2.7(a) indicates that the time averaged axial velocity profiles in the column with internals exhibit similar trends to those obtained in the column without internals. As the superficial gas velocity increases, the velocity inversion moves radially inward (further from the wall) for the column with internals, whereas in the column without internals the velocity inversion occurs almost at the same radial position (same r/R). The time averaged radial velocities in the column with internals (obtained in the region of fully developed flow (90 cm to 190 cm of the column height) are of the same order of magnitude as those obtained in the column without internals as shown in Figure 2.6(b ) and 2.7(b), whereas, Figure 2.7 (c) shows little larger magnitude of time averaged radial liquid velocities obtained for the whole height of the column. This confirms that the column with internals the flow is not fully developed in the bottom part of the column and, hence when averaging over the whole column height, which gets reflected in finite (but small) time averaged radial liquid velocities. It seems that in both of the columns with and without internals, the radial liquid flow is towards to the center of the column at low gas superficial velocity while it is towards to the wall at high gas superficial velocity. Nevertheless, the magnitude of radial velocities is very small compared to the axial ones indicating that the flow is close to fully developed in the top part of the column.

Note that in Figures 2.6(a), (b) and (c), the dashed lines indicate the radial positions of the tubes.

2.2.3 Comparison of the Results Obtained in Columns with and without Internals

Figures 2.8(a) and (b) illustrate the comparison between the velocity profiles in the column with internals and those obtained in the same column without internals at superficial gas velocities of 5 and 10 cm/s, respectively. The trend of the time averaged axial liquid velocity profiles in the column with and without internals are very similar. For the superficial gas velocity of 5 cm/s most of the differences between the two profiles are within the experimental error except in the center of the column where larger velocity is observed in the column with internals. However, this discrepancy may be caused by the poor statistics in the central compartments of the column (i.e inadequate number of tracer particle occurrences in the central compartments) which may cause larger error bars on the velocity profile at that location. Hence one cannot say with great certainty that the liquid time averaged velocity close to the center of the column is larger in the column with internals at 5 cm/s superficial gas velocity. At the higher gas superficial velocity of 10 cm/s the time azimuthally and axially averaged liquid velocity in the center of the column seems a little higher in the column with internals, while the inversion point seems to have moved inwards. Additional data processing and additional experiments are needed to determine whether the observed difference in the velocity profiles in the column with and without internals are reproducible and real.

Data processing and analysis to estimate the turbulence parameters in the columns with and without internals are in progress and these results will be reported later.

2.3 References

  1. Chen, R. C. and L.-S. Fan, “ Particle Image Velocimertry for Characterizing the Flow Structure in Three-Dimensional Gas-Liquid-Solid Fluidized Beds”, Chem. Eng.. Sci., 47, 3615(1992)
  1. Degaleesan, S., “Fluid Dynamic Measurements and Modeling of Liquid Mixing in Bubble Columns”, D. Sc. Thesis, Washington University in St. Louis, 1997
  1. Devanathan, N., “Investigation of Liquid Hydrodynamics in Bubble Columns via Computer Automated Radioactive Particle Tracking”, D.Sc. Thesis, Washington University in St. Louis, 1990.
  1. Devanathan, N., Moslemian, D., and Dudukovic´, M. P., “Flow Mapping in Bubble Columns Using CARPT”, Chem. Eng. Sci.,45, 2285, (1990).
  1. Franz, K., T. Borner, H. J. Kantorek, and R. Buchholz, “ Flow Structures in Bubble Columns”, Ger. Chem. Eng., 7, 365(1984)
  1. Groen, J. S., R. G. C. Oldeman, R. F. Mudde, and H. E. A. Van Den Akker, “ Coherent Structure and Axial Dispersion in Bubble Column Reactors”, Chem. Eng. Sci., 51, 2511(1996)
  2. Kumar, S. B., D. Moslemian, and M. P. Dudukovic, “A Gamma Ray Tomographic Scanner for Imaging Void Fraction Distribution in Bubble Columns”, Flow Meas. Instrum. 6(1) (1995) 61-73
  1. Kumar, S. B., D. Moslemian, and M. P. Dudukovic, “Gas Holdup Measurements in Bubble Columns Using Computed Tomography”, AIChE. J. 43(6) (1997) 1414-1425
  1. Limtrakul, S., “Hydrodynamics of Liquid Fluidized Beds and Gas-Liquid Fluidized Beds”, Ph.D. Thesis, Washington University in St. Louis, 1996
  1. Menzel, T., T. in der Weide, O. Staudacher, O. Wein, and U. Onken, “ Reynolds Shear Stress for Modeling of Bubble Column Reactors”, Ind. Eng. Chem. Res., 29, 988(1990)
  1. Moslemian, D., N. Devanathan and M. P. Dudukovic´, “Radioactive Particle Tracking Technique for Investigation of Phase Recirculation and Turbulence in Multiphase Systems”, Rev. Sci. Instrum. 63(10), 4361 (1992).
  1. Mudde, R. F., J. S. Groen and H. E. A. Van Den Akker, “ Liquid Velocity Field in a Bubble Column: LDA Experiments”, Personal Communication, 1997
  1. Roy, S., J. Chen, S. Kumar, M. H. Al-Dahhan, and M. P. dudukovic, “Tomography and Particle Tracking Studies in a Liquid-Solid Riser”, I & E C, in press, (1997).
  1. Yang, Y. B., N. Devanathan, and M. P. Dudukovic, “Liquid Backmixing in Bubble Columns via Computer Automated Radioactive Particle Tracking (CARPT)”, Exp. Fluids, 16, 1(1993)