Packedcolumn Flood And Pressure Drop

Pressure drop of a gas flowing upward through a packing countercur-rently to liquid flow is characterized graphically in Fig. 14-53. At very low liquid rates, the effective open cross section of the packing is not appreciably different from that of dry packing, and pressure drop is due to flow through a series of variable openings in the bed. Thus, pressure drop is proportional approximately to the square of the gas velocity, as indicated in the region AB. At higher liquid rates, the effective open cross section is smaller because of the presence of liquid (region A'B'). The pressure drop is higher, but still proportional to the square of the gas velocity.

At higher gas rates, a portion of the energy of the gas stream is used to support an increasing quantity of liquid in the column. For all liquid rates, a zone is reached where pressure drop is proportional to a gas flow rate to a power distinctly higher than 2; this zone is called the loading zone. The increase in pressure drop is due to the liquid accumulation in the packing voids (region BC or B'C')

As the liquid holdup increases, the effective orifice diameter may become so small that the liquid surface becomes continuous across the cross section of the column. Column instability occurs concomi-tantly with a rising continuous-phase liquid body in the column. Pressure drop shoots up with only a slight change in gas rate (condition C or C ). The phenomenon is called flooding and is analogous to entrain-ment flooding in a tray column.

Alternatively, a phase inversion occurs, and gas bubbles through the liquid. The column is not unstable and can be brought back to gasphase continuous operation by merely reducing the gas rate. A stable operating condition beyond flooding (region CD or C D ) may form with the liquid as the continuous phase and the gas as the dispersed phase [Lerner and Grove, Ind. Eng. Chem. 43, 216 (1951); Teller, Chem. Eng. 61(9), 168 (1954); Leung et al., Ind. Eng. Chem. Fund. 14 (1), 63 (1975); Buchanan, Ind. Eng. Chem. Fund. 15 (1), 87 (1976)].

For total-reflux distillation in packed columns, regions of loading and flooding are identified by their effects on mass-transfer efficiency, as shown in Fig. 14-54. Gas and liquid rate increase together, and a

Reflux Column DesignSulzer Packed ColumnsMellapak Plus Differences

FIG. 14-50 Common structured packings. (a) A small element of Mellapak™ showing embossed surface, holes, and corrugated-sheet arrangement. (b) A closeup of the surface of FlexipacTM showing grooved surface and holes. (c) Fitting structured packing elements to a large-diameter tower. (d) Mellapak Plus™, a fourth-generation structured packing, showing a 45° inclination angle in the element and near-vertical inclination at the element-to-element transition. Note that in the tower, the successive layers will be oriented 90° to each other as in part b. (Parts a, d, courtesy of Sulzer Chemtech; parts b, c, courtesy of Koch-GHtsch LP. )

FIG. 14-50 Common structured packings. (a) A small element of Mellapak™ showing embossed surface, holes, and corrugated-sheet arrangement. (b) A closeup of the surface of FlexipacTM showing grooved surface and holes. (c) Fitting structured packing elements to a large-diameter tower. (d) Mellapak Plus™, a fourth-generation structured packing, showing a 45° inclination angle in the element and near-vertical inclination at the element-to-element transition. Note that in the tower, the successive layers will be oriented 90° to each other as in part b. (Parts a, d, courtesy of Sulzer Chemtech; parts b, c, courtesy of Koch-GHtsch LP. )

point is reached at which liquid accumulates rapidly (point B) and effective surface for mass transfer decreases rapidly.

Flood-Point Definition In 1966, Silvey and Keller [Chem. Eng. Progr. 62(1), 68 (1966)] listed 10 different flood point definitions that have been used by different literature sources. A later survey (Kister and Gill, Proceedings of Chemeca 92, p. 185-2, Canberra, Australia, 1992) listed twice that many. As Silvey and Keller pointed out, the existence of so many definitions puts into question what constitutes flooding in a packed tower, and at what gas rate it occurs. Symptoms used to identify flood in these definitions include appearance of liquid on top of the bed, excessive entrainment, a sharp rise in pressure drop, a sharp rise in liquid holdup, and a sharp drop in efficiency. The survey of Kister and Gill suggests that most flood point definitions describe the point of flooding initiation (incipient flooding; point C or C' on Figs. 14-53 and 14-54). The different incipient flooding definitions gave surprisingly little scatter of flood point data (for a given packing under similar operating conditions). It follows that any definition describing flooding initiation should be satisfactory.

The author believes that due to the variations in the predominant symptom with the system and the packing, the use of multiple symptoms is most appropriate. The author prefers the following definition

Liquid Inlet

Liquid Distributor

Packed Bed (Structured Packing)

Liquid Distributor

Packed Bed (Random Packing)

Column Sump

FIG. 14-51 Illustrative cutaway of a packed tower, depicting an upper bed of structured packing and a lower bed of random packing. (Courtesy of Sulzer Chemtech.)

by Fair and Bravo [Chem. Eng. Symp. Ser. 104, A183 (1987)]: "A region of rapidly increasing pressure drop with simultaneous loss of mass transfer efficiency. Heavy entrainment is also recognized as a symptom of this region." An almost identical definition was presented earlier by Billet (Distillation Engineering, Chem. Publishing Co., New York, 1979).

The maximum operational capacity or throughput (often also referred to as maximum efficient capacity) is defined (Strigle, Packed Tower Design and Applications, 2d ed., Gulf Publishing, Houston, Tex., 1994) as the "Maximum vapor rate that provides normal efficiency of a packing" (i.e., point B in Fig. 14-54). The MOC is clear-cut in Fig. 14-54. On the other hand, locating the MOC in other cases is difficult and leaves a lot of room for subjectivity.

Triangular cross section

FIG. 14-52 Crimp geometry in structured packings. (a) Flow channel cross section. (b) Flow channel arrangement. (From J. R. Fair and J. L. Bravo, Chem. Eng. Progr., Jan. 1990, p. 19; reproduced courtesy of the American Institute of Chemical Engineers.)

In most cases, [Kister and Gill, Chem. Eng. Progr. 87(2), 32 (1991)], the velocity at which MOC is reached is related to the flood point velocity by us.MOC = 0.95 us_fl (14-139)

Flood and Pressure Drop Prediction The first generalized correlation of packed-column flood points was developed by Sherwood, Shipley, and Holloway [Ind. Eng. Chem., 30, 768 (1938)] on the basis of laboratory measurements primarily on the air-water system with random packing. Later work with air and liquids other than water led to modifications of the Sherwood correlation, first by Leva [Chem. Eng. Progr. Symp. Ser., 50(1), 51 (1954)], who also introduced the pressure drop curves, and later in a series of papers by Eckert. The generalized flooding-pressure drop chart by Eckert [Chem. Eng. Progr. 66(3), 39 (1970)], included in previous editions of this handbook, was modified and simplified by Strigle (Packed Tower Design and Applications, 2d ed., Gulf Publishing, Houston, Tex., 1994) (Fig. 14-55). It is often called the generalized pressure drop correlation (GPDC). The ordinate is a capacity parameter [Eq. (14-140)] related to the Souders-Brown coefficient used for tray columns.

CP = CsFp°-5v 0 05 = uj pG Y'50e5va05 (14-140) p V Pl - Pg / p where US = superficial gas velocity, ft/s pG, pL = gas and liquid densities, lb/ft3 or kg/m3

Superficial Gas Velocity Pressure Drop

log gas rote

FIG. 14-53 Pressure-drop characteristics of packed columns.

log gas rote

FIG. 14-53 Pressure-drop characteristics of packed columns.

Structured Packed Column Image

FIG. 14-51 Illustrative cutaway of a packed tower, depicting an upper bed of structured packing and a lower bed of random packing. (Courtesy of Sulzer Chemtech.)

Sulzer Mellapak

FIG. 14-52 Crimp geometry in structured packings. (a) Flow channel cross section. (b) Flow channel arrangement. (From J. R. Fair and J. L. Bravo, Chem. Eng. Progr., Jan. 1990, p. 19; reproduced courtesy of the American Institute of Chemical Engineers.)

Total Reflux Distillation Column
FIG. 14-54 Efficiency characteristics of packed columns (total-reflux distillation.)

Fp = packing factor, ft-1 V = kinematic viscosity of liquid, cS

C'S = C-factor, Eq. (14-77), based on tower superficial cross-

sectional area, ft/s CP = capacity factor, dimensional [units consisted with Eq. (14-140) and its symbols]

The abscissa scale term is the same flow parameter used for tray (dimensionless):

For structured packing, Kister and Gill [Chem. Eng. Symp. Ser. 128, A109 (1992)] noticed a much steeper rise of pressure drop with flow parameter than that predicted from Fig. 14-55, and presented a modified chart (Fig. 14-56).

The GPDC charts in Figs. 14-55 and 14-56 do not contain specific flood curves. Both Strigle and Kister and Gill recommend calculating the flood point from the flood pressure drop, given by the Kister and Gill equation

Equation (14-142) permits finding the pressure drop curve in Fig. 14-55 or 14-56 at which incipient flooding occurs.

For low-capacity random packings, such as the small first-generation packings and those smaller than 1-in diameter (Fp > 60 ft-1), calculated flood pressure drops are well in excess of the upper pressure drop curve in Fig. 14-55. For these packings only, the original Eckert flood correlation [Chem. Eng. Prog. 66(3), 39 (1970)] found in pre-1997 editions of this handbook and other major distillation texts is suitable.

The packing factor Fp is empirically determined for each packing type and size. Values of Fp, together with general dimensional data for individual packings, are given for random packings in Table 14-13 (to go with Fig. 14-55) and for structured packings in Table 14-14 (to go with Fig. 14-56).

Packing flood and pressure drop correlations should always be used with caution. Kister and Gill [Chem. Eng. Progr., 87(2), 32 (1991)] showed that deviations from the GPDC predictions tend to be systematic and not random. To avoid regions in which the systematic deviations lead to poor prediction, they superimposed experimental data points for each individual packing on the curves of the GPDC. Figure 14-57 is an example. This method requires a single chart for each packing type and size. It provides the highest possible accuracy as it interpolates measured data and identifies uncertain regions. A set of charts is in Chapter 10 of Kister's book (Distillation Design, McGraw-Hill, New York, 1992) with

General Pressure Drop Correlation

FIG. 14-55 Generalized pressure drop correlation of Eckert as modified by Strigle. To convert inches H2O to mm H2O, multiply by 83.31. (From Packed Tower Design and Applications by Ralph E. Strigle, Jr. Copyright © 1994 by Gulf Publishing Co., Houston, Texas. Used with permission. All rights reserved.)

FIG. 14-55 Generalized pressure drop correlation of Eckert as modified by Strigle. To convert inches H2O to mm H2O, multiply by 83.31. (From Packed Tower Design and Applications by Ralph E. Strigle, Jr. Copyright © 1994 by Gulf Publishing Co., Houston, Texas. Used with permission. All rights reserved.)

Pressure Drop Irrigated Packing Graph
FIG. 14-56 The Kister and Gill GPDC (SP) chart for structured packings only. Abscissa and ordinate same as in Fig. 14-55. (From Kister, H. Z., and D. R. Gill, IChemE Symp. Ser. 128, p. A109, 1992. Reprinted courtesy of IChemE.)

updates in Kister, Lason, and Gill, Paper presented at the AIChE Spring National Meeting, Houston, Tex., March 19-23, 1995; and in Kister, Scherffius, Afshar, and Abkar, in Distillation 2007: Topical Conference Proceedings, 2007 AIChE Spring National Meeting, Houston, Texas. The latter reference also discusses correct and incorrect applications of those interpolation charts.

There are many alternative methods for flood and pressure drop prediction. The Billet and Schultes [IChemE Symp. Ser. 104, pp. A171 and B255 (1987)] and the Mackowiak ("Fluiddynamik von Kolonnen mit Modernen Füllkorpern und Packungen für Gas/Flus-sigkeitssysteme," Otto Salle Verlag, Frankfurt am Main und Verlag Sauerländer Aarau, Frankfurt am Main, 1991) correlations are versions of the GPDC that take the liquid holdup into account. The Eiden and Bechtel correlation [IChemE Symp. Ser. 142, p. 757 (1997)] is a version of the GPDC in which accuracy is improved by using constants representative of packing shape instead of packing factors. The Lockett and Billingham correlation (IChemE Symp. Ser. 152, p. 400, London, 2006) uses a Wallis correlation where cG + mcL = cu

and was shown to work well for high-surface-area (>400 m2/m3) structured packings. Here CG is the gas C-factor, Eq. (14-77), based on the tower superficial cross-sectional area, and m and CLG are constants, available from the cited reference for some packing.

A drawback of most of these correlations (except that of Eiden and Bechtel) is the unavailability of constants for many, often most, of the modern popular packings.

The above methods apply to nonfoaming systems. Foaming systems can be handled either by applying additional derating (system) factors to the flood correlation (see Table 14-9) or by limiting the calculated pressure drop to 0.25 in of water per foot of packing (Hausch, "Distillation Tools for the Practicing Engineer," Topical Conference Proceedings, p. 119, AIChE Spring Meeting, New Orleans, March 10-14, 2002).

Pressure Drop The GPDC discussed above (Figs. 14-55 and 14-56) and the Kister and Gill interpolation charts provide popular methods for calculating packing pressure drops. An alternative popular method that is particularly suitable for lower liquid loads was presented by Robbins (below).

For gas flow through dry packings, pressure drop may be estimated by use of an orifice equation. For irrigated packings, pressure drop increases because of the presence of liquid, which effectively decreases the available cross section for gas flow (Fig. 14-53). In principle, there should be a method for correcting the dry pressure drop for the presence of liquid. This approach was used by Leva [Chem. Eng. Progr. Symp. Ser. No. 10, 50, 51 (1954)]. A more recent method by Robbins [Chem. Eng. Progr., p. 87 (May 1991)] utilizes the same approach and is described here. The total pressure drop is

where APt = total pressure drop, inches H2O per foot of packing

APl = pressure drop due to liquid presence

Gf = gas loading factor = 986Fs(Fp/20)«-5 (14-148)

Lf = liquid loading factor = L(62.4/pL)(Fp/20)<15|4L1 (14-149)

The term Fpd is a dry packing factor, specific for a given packing type and size. Values of Fpd are given in Tables 14-13 and 14-14. For operating pressures above atmospheric, and for certain packing sizes, Lf and G/ are calculated differently:

L{ = L(62.4/Pl)(V2°)°-W2 Lf = L(62.4/pL)(2°/Fpd)°-5^L1

(14-15°) Fpd > 2°° (14-151«) Fpd < 15 (14-1516)

The Robbins equations require careful attention to dimensions. However, use of the equations has been simplified through the introduction of Fig. 14-58. The terms Lf and Gf are evaluated, and the APL is obtained directly from the chart. Basic nomenclature for the Robbins method follows:

Fpd = dry packing factor, ft-1

Fs = superficial F-factor for gas, U.pf, ft/s(lb/ft3)05 G = gas mass velocity, lb/hrft2 Gf = gas loading factor, lb/hrft2 L = liquid mass velocity, lb/hrft2 Lf = liquid loading factor, lb/hrft2 AP = pressure drop, inches H2O/ft packing (x 83.3 = mm H2O/m packing)

TABLE 14-13 Characteristics of Random Packings
+5 0

Responses

  • CARRIE
    How gas is fed in packed towers?
    7 years ago
  • Raffaele
    What is the definition of ffactor in packed towers?
    6 years ago
  • dahlak
    How flooding related to pressure drop?
    3 years ago
  • faramond
    What is the meaning of chem. eng. progr. symp. ser.?
    3 years ago
  • AKI HEIKKIL
    How pressure is dropped by flooding?
    7 months ago

Post a comment