I.  Packed Towers Hydraulics


 


Packed Towers Hydraulics

Rowan University – Department of Chemical Engineering

 

INTRODUCTION

Objective

The purpose of this experiment is to familiarize students with the flooding characteristics of a packed column.  Since packed columns are used in gas absorption applications and to an extent in distillation, the operation of a laboratory packed column exposes the student to a widely used operational unit.  Prior to operating the packed column, the student should be familiar with the appearance of a flooding point curve and should have an idea of the magnitude of the pressure drops at which flooding occurs.

 

Theory

Packed towers are vertical columns filled with a suitable packing and normally operate countercurrently.  Liquid enters the top of the column and is distributed over the top of the column packing via nozzles or distributor plates.  Liquid flows downward while contacting with the vapor phase.  Internal packing provides a large surface area for two-phase contact and facilitates transfer of materials between phases.

 
Pressure drop, loading, and flooding in a packed column.

Refer to the literature [1-3] for a discussion of pressure drop, loading, and flooding.  Liquid entering at the top of a packed column flows downward through the column due to gravitational force.  Gas entering the bottom of the column flows upward due to pressure.  Thus, the operating pressure at the top of the column must be less than the pressure at the column bottom in order for the gas to flow upwards, a consideration that is important for packed tower design.  However, the pressure drop is a function of both the liquid and the gas flow rates since the liquid occupies the same areas in the column as does the gas.  A packed distillation column operates the same way except that the vapor or gas phase is produced by energy input into the reboiler.

Typical flooding curves are shown in McCabe et al., Figure 22.4 [1], and Zenz, Figures 4, 13 and 60 [3].  Note that the Zenz graphs should all have an abscissa of log and an ordinate of log DP/Z (typographical errors on two figures).  Gas flow through a packed tower normally is turbulent; therefore, the slope of the curve for dry packing on logarithmic coordinates is usually in the range of 1.8 to 2.0 [1-2].  If the gas flow rate remains constant, the pressure drop will increase with an increase in the liquid flow rate.  This occurs because as the liquid fills the voids in the column, the cross-sectional area available for gas flow is reduced.  Up to the loading point, the pressure drop graphs at constant liquid rate show slopes that are fairly linear and approximately equal to that of the dry run.  Above the loading point, an increase in slope occurs, indicating a more rapidly increasing pressure drop with increasing gas velocity.  Between the loading and flooding points, the students should visually observe a greater amount of liquid holdup in the column.  Finally, at the flooding point, the slope becomes exponential; water will build up on top of the packing and/or spray out of the column.  At this point, the drag force of the gas bubbling through the downcoming liquid is the predominate cause of the pressure drop  (DPflood ) across the column [1-3].

 For the pressure drop of a single-phase fluid (air or water) flowing through a packed bed of solids, the Ergun equation can be used.  The following equation applies only to the dry air runs in this experiment:

 

                                               (1)

 

Where:    DP = pressure drop across column        [=] lbf/ft2

   Z   = packed bed height                        [=] ft

               µg  = viscosity (air)                        [=] lbm/ft h

   = fractional void volume                     [=] dimensionless

   gc  = Newton’s law proportionality factor = 32.174 ft-lbm/lbf-s2

               ap = square feet of packing surface in a cubic foot of packed volume

   rg = fluid (air) density                           [=]  lbm/ft3

   [=] ft

                 [=]  lbm/ft2 h.

 Typical flooding curves plot DP/Z versus  on logarithmic coordinates (McCabe et al., Figure 22.4 [1]), which should give a curved line with a slope approaching 1.0 at low Reynolds numbers and a slope approaching 2.0 at high Reynolds numbers.

For two-phase (liquid/gas) flow through a packed column, correlations based on papers by Eckert (and Sherwood, Shipley and Holloway) are graphed in various forms in McCabe et al., Figure 22.6 or 22.7 [1], Wankat, Figure 13-4 [2], and Zenz, Figure 2 [3].  The flooding line on these graphs can be calculated from the following equation [2]:

            (2)

Where:        = superficial gas mass velocity per tower cross-sectional area

               

   = superficial liquid mass velocity per tower cross-sectional area

                [=] lbm/ft2-s

   Fp    = packing factor (dimensionless)

   rL    = liquid density                                       [=] lbm/ft3

   Y    = rwater /rL

   µL    = absolute viscosity of liquid          [=] cp (centipoise)

   X     = abscissa of Wankat’s graph or   =

          

Also:

 

 

If the packing factor is known, an Eckert-type graph can be used to estimate pressure drops and design packed columns.  For packing factors above 60, the pressure drop at flooding, , is approximately 2.0 in. H2O/(ft packing) [1].  For packing factors from 10 to 60, the following equation can be used [1]:

                                                                                                            (3)

 

Packing Characteristics

           Packing materials must have the following characteristics: a large wetted surface area per unit volume of packed space (this allows a large interfacial area to exist), a large void volume (needed for low pressure drops), good wetting characteristics, corrosion resistance, low bulk density, and low cost.  No one type of packing will possess all these characteristics; therefore, compromises normally are made.

The tower packing method is also an important consideration for effective phase contact.  Channeling often occurs in a packed tower.  This phenomenon takes place when the fluid moving down the column moves towards the region of greatest void space; this occurs at the region near the wall where the packing is not tightly packed.  Thus, liquid redistributors are used to redirect the fluid flow towards the column center.

 

EXPERIMENTAL SET-UP

The equipment as shown in Figure 1 includes three Glass columns, two rotameters for the water and air flow rates, a manometer for measuring the pressure drop across the column during operation and a pressure transducer for measuring the pressure drop.  Each column has the same overall dimensions but contain a different type of packing.  The columns are arranged so they can be operated independently.  Carefully examine the system, trace the process flow directions for air and water, and identify the function of all valves:

1.      Valve FCV 1 controls the air flow.

2.      Valve FCV 4 controls the water flow.

3.      Valves V8, V9, and V10 control the outlet water flow and must be adjusted to prevent air flow through the liquid drainage pipe.

4.      Valve SV1 switches between columns for pressure drop measurements.

5.      Valves V1, V2 and V3 are for the air inlet to each column.

6.      Valves V4, V5 and V6 are for the water inlet to each column.

Figure 1: Equipment Set-up

Column 1 contains 7-mm glass Raschig Rings. Column 2 contains 10-mm glass Raschig Rings. Column 3 contains Ľ” Intalox Saddles.  The approximate packed height of each column is 1.3 m and the diameter of each column is 80 mm.  Pressure Drop across each column is measure using the manometer or the pressure transducer.  Temperature of the water inlet, air inlet and air outlet may be measure by selecting the desired thermocouple on the temperature sensor.

When one column is in use the other columns may be isolated using valves V8, V9, V10, V4, V5, V6, and V1, V2, V3.

The columns may be run in two liquid flow modes: recirculation and once through.  In the recirculation mode the liquid reservoir is charged with water; the pump is turned on and water is pumped through the flowmeter to the top of the absorber.  The water flows down the column and back into the liquid reservoir where it is recirculated.  In once through mode the unit is connected to an external water supply.  With the valve to the recirculation tank closed the water will flow up through the flowmeter to the absorber.  The water then flows down through the column and since valve V11 is open the water exits the unit through the drain.  In this experiment the recirulating liquid flow mode will be used.

 

Manipulated Variables

·        Gas flowrate - FCV1:           20 ~ 120 SLPM

·        Liquid flowrate - FCV4:       1.4 ~ 4 LPM

 

Ranges of Dependent Variables

The ranges are the following:

Pressure drop: 0 ~ 15psi

 

PROCEDURE

Dry Air Runs

      If the packing in the 10mm Raschig ring column is wet, blow air through the column by opening valve V5 and opening valve FCV1.

1.         The water rotameter should be closed (0 SLPM).

2.                  Set the air flowrate at 20 SLPM.

3.         Record the rotameter reading and pressure drop across the column. The pressure drop is measured in inches of water.

4.         Record the air temperature and air pressure in line with the air rotameter.

5.         Repeat the procedure for 30, 40, 50, 60, 70, 80, 90, 100, 110 and 120 SLPM air flowrate. 

            The air rotameter measures 12.6-126 SLPM at 14.7 psia and 70 oF.  The air flow readings must be corrected for operating pressure Po (psia) and operating temperature To (R):

                                                (4)

                                         Ps = Standard pressure, 14.7 psia

                                         Ts = Standard temperature, 70oF or 530oR

6.         Repeat this procedure using the 1/4" intalox saddels column by opening valve V4 and closing valve V5.

 

Liquid/Gas Counterflow Runs

Startup

1.                  Before starting up, make sure that V7 is open and that it the liquid has a place to go.  You should not need to move any black handled valves in this lab.  Make sure that either V8, V9, or V10 is open.  After this open the corresponding colored valves from V1 – V6.  The valves colors are grouped V1, V4, V8 for yellow, V2, V5, V9 for green, and V3, V5, V10 for blue. 

2.                  Open FCV1, the black needle valve just below the air inlet rotameter, to the desired flowrate to begin experiments.  With all of the valves of one color open and the rest of the color valves closed it is now safe to turn on the pump.  

3.                  Immediately after starting the pump, slowly open FCV4, the black needle valve just below the liquid inlet rotameter, until the desired flowrate is reached.  Be careful not to increase the liquid flowrate too rapidly in order to avoid raising the water level in the column above the air inlet. 

4.                  With a steady flow of air and water, next adjust the liquid level to safe operating conditions.  This is done by partially closing V8-10.  The desired operating conditions are between the black lines on the metal collar.  It is ideal to be level with the white, Teflon disc between the two collars.  It is okay to fall below provided that you increase the level and wait for the system to reach steady state.  If the level goes too high and reaches the air inlet open the outlet valve V8-10 to drop the level.  On occasion this may not suffice and you will have to immediately drop the liquid flowrate and possibly increase air flowrate.  It is imperative that you keep the level below the air inlet of the column and important to have a high enough level of water to have a seal in the column to prevent air flow to the bottom of the column.

5.                 Once a steady level in the operating zone is achieved it is safe to start experiments.

 

Operating Procedure

1.                  Choose five liquid flowrates in increasing order in the ranges of 1.4 – 4 LPM as points to collect data.  For each flowrate of liquid, record the pressure drop at each 20 SLPM increment in the ranges of 20 – 120 SLPM or until the column floods.  Note that at higher liquid flowrates the air flowrate range will decrease as seen at 4 LPM it may flood at less than 40 SLPM. 

2.                  Once liquid holdup in the column is observed, take data at very tiny increments in gas flow rate in order to record flooding data.  Patience is needed to obtain accurate steady-state readings especially in the holdup/flooding region.  At a given liquid and gas flowrate in the vicinity of flooding, the pressure across the packed bed should rise slowly until the liquid holdup in the column reaches its maximum, but the pressure readings will probably oscillate. 

3.                  Record data when the pressure drop across the column is more or less constant.  Three group members are necessary at this point in order to keep the two rotameters and the level in the bottom portion of the column stable and also to watch the manometers. 

4.                  When visual flooding is observed for each run, take extra data points at air flowrates just prior to flooding and slightly higher if possible.  For example, if flooding occurs at approximately an air flowrate of 80 SLPM, then take data points at 72, 74, 76, 80, 82, 84 SLPM, etc. This will allow for a more accurate determination of flooding and loading points from the pressure drop graphs.  Watch the manometers closely in this region so that you do not get water in them (pressure readings are worthless if water is in the manometer). Good data in the curved region between loading and flooding are difficult to obtain; however, visual flooding should be observed in this region. 

5.                  Record the observed value of DPflood, and include notes documenting flooding observations on your data sheet. 

6.                  After flooding, decrease the liquid flowrate and air flowrate slowly.  Be very careful not to have the operating water level increase too rapidly as the air flowrate is decreased. 

7.                  After all readings have been taken for each desired data point, it is time to switch the columns on the fly.  First open V8-10 of the column the is desired to be used next.  Open the corresponding colored inlet valves of the column that will be switched to.  Close the valves of the column that is being taken offline.  Adjust the outlet valve of the new column in order to get the proper operating level.  Lastly, set SV1 to the proper column in order to get proper pressure drop readings.  If this has not been changed you may get a negative pressure drop or a reading of 0.

 

Shutdown

Shut of the pump and turn off FCV1 and FCV4.  It may be desirable to let FCV1 run for a little while to allow the column to dry.

 

REPORT

 

1.                  Create a log-log plot of vs. Vg with liquid flow as a parameter where is the pressure drop across the top and bottom packing sections of the tower, (in H2O), Vg is the superficial gas velocity [volumetric flowrate (ft3/s)/tower cross sectional area (ft2)], H is the packing height.  The slope of the line is the exponent of the correlation:

2.             Using the dry air data, graph log DP / Z versus log (lbm/ft^2 * hr).  Use the Ergun equation (Eq. 1) to calculate DP / Z = f(G') and plot on the same graph.  Calculate the Reynolds number for each value of .  Discuss and compare the two curves on the resulting graph (slopes at low and high Re values, general shapes of the curves, numerical values, etc.).  Do the slopes of the experimental curve agree with the values expected from Eq. 1 and the Reynolds numbers (discuss)?

3.                  Graph the pressure drop data for the liquid/gas runs (all data) using the same format as the flooding curves in the literature, i.e., log DP / Z versus log (McCabe et al., Figure 22.4 [1]).  Do the slopes at lower pressures and the general shape of the curves at higher pressures follow the expected trends (discuss)?  Compare the value of the regressed slopes at lower pressures with values expected from the literature.  If the graphs curve upwards at higher pressures, indicate loading and flooding points on each curve and determine an experimental value for the pressure drop at flooding.  If the graphs do not curve upward at higher pressures but you observed flooding visually explain the discrepancy.

4.             Determine DPflood from Eq. 3 and compare with the value(s) found in Steps 2-3.

5.            Prepare a table for the visual flooding data (one point for each water flow rate) based on Zenz’s flooding correlation [3] as given below, and discuss.

                                                                              (5)

   Where:    V = flooding gas rate in acfm per ft2 of tower cross-section

                  Q = flooding liquid rate in gpm per ft2 of tower cross-section

6.            Use the flooding data to calculate a packing factor as defined by Zenz, following the example on pages 3-102 to 3-103 [3].

 

Safety

There are several safety considerations that must be taken in order to do this experiment.  Safety glasses must be worn at all times!   Be careful of equipment, tools, and other projects in the high bay lab.

 

REFERENCES

1.      W.L. McCabe, J.C. Smith, and P. Harriott, Unit Operations of Chemical Engineering, 5th ed., McGraw-Hill, New York, 1993, Chapters 7, 22.

2.      P.C. Wankat, Equilibrium Staged Separations, Prentice Hall, Englewood Cliffs, NJ, 1988, Chapter 13.

3.      F.A. Zenz, "Design of Gas Absorption Towers", Handbook of Separation Techniques for Chemical Engineers, P.A. Schweitzer (ed.), 2nd ed., McGraw-Hill, New York, 1988, pp. 3-50 to 3-66, 3-102 to 3-108.

4.      Robbins, L.A., “Improved Pressure Drop Prediction with a New Correlation,” Chemical Engineering Progress, May 1991, pp. 87-91.

5.      Phipps, D. A., Kinder, M., CTS 11 Absorption Packed Column Absorbers Student Manual, QVF Teaching Systems.

6.      Hollen, H. and C.S. Slater, EXPERIMENT 4 - Flooding Point of a Packed Column, Manhattan College.