Hetp Pall Rings

FIG. 14-57 Superimposing experimental pressure-drop data for a given packing generates a GPDC interpolation chart for this packing. (a) A random packing; chart is based on Eckert's GPDC, Fig. 14-55. (b) A structured packing; chart is based on Kister and Gill's GPDC (SP), Fig. 14-56. (From Kister, H. Z., Distillation Design, copyright © McGraw-Hill, 1992; used with permission.)

Kister and Larson (in Schweitzer, Handbook of Separation Techniques for Chemical Engineers, 3d ed., McGraw-Hill, 1997) extended Eq. (14-156a) by expressing the packing diameter in terms of the more fundamental surface area per unit volume aP, m2/m3. For Pall rings, it can be shown that aP = 5.2/DP (14-157)

and Eq. (14-156a) becomes

Pall Ring Pressure Drop Graph
FIG. 14-58 The Robbins generalized pressure-drop correlation. (From L. A. Robbins. Chem Eng. Progr., May 1991. p. 87, reprinted courtesy of the American Institute of Chemical Engineers.)

Harrison and France (loc. cit.) presented the only published rule of thumb for structured packings efficiency as a function of packing crimp. Kister and Larson reexpressed it in terms of the surface area per unit volume to accommodate a wider range of packing geometries. The final expression is

Specific surface areas are listed in Tables 14-13 and 14-14.

The above rules of thumb apply to organic and hydrocarbon systems, whose surface tensions are relatively low (c < 25 mN/m). For higher surface tensions, the liquid does not adhere well to the packing surfaces (underwetting), causing higher HETPs. In a water-rich system (c = 70 mN/m or so) HETPs obtained from Eqs. (14-156), (14-158), and (14-159) need to be doubled. For intermediate surface tension systems (some amines and glycols, whose surface tension at column conditions is 40 to 50 mN/m), HETPs obtained from Eqs. (14-156), (14-158), and (14-159) need to be multiplied by 1.5.

For random packings, Eqs. (14-156) and (14-158) apply for packings of 25-mm diameter and larger. For smaller packings, use of the ap at 25-mm often gives a slightly conservative HETP estimate. For structured packing, CXY in Eq. (14-159) reflects the effect of the inclination angle (Fig. 14-600). CST = 1 for Y-type, S-type, or high-capacity packings, and CXY =1.45 for the larger (< 300 m2/m3) X-type packings. There are insufficient data to determine C'n for high specific area X-type packings, but Fig. 14-60 suggests it is somewhat lower than 1.45.

Compared to experimental data, the above rules of thumb are slightly conservative. Since packing data are usually measured with perfect distribution, a slight conservative bias is necessary to extend

Hetp For Pall Rings

FIG. 14-59 HETP values for four sizes of metal pall rings, vacuum operation. Cyclohexane/n-heptane system, total reflux, 35 kPa (5.0 psia). Column diameter = 1.2 m (4.0 ft). Bed height = 3.7 m (12 ft). Distributor = tubed drip pan, 100 streams/m2. [Adaptedfrom Shariat and Kunesh, Ind. Eng. Chem. Res., 34, 1273 (1995). Reproduced with permission. Copyright © 1995 American Chemical Society. ]

FIG. 14-59 HETP values for four sizes of metal pall rings, vacuum operation. Cyclohexane/n-heptane system, total reflux, 35 kPa (5.0 psia). Column diameter = 1.2 m (4.0 ft). Bed height = 3.7 m (12 ft). Distributor = tubed drip pan, 100 streams/m2. [Adaptedfrom Shariat and Kunesh, Ind. Eng. Chem. Res., 34, 1273 (1995). Reproduced with permission. Copyright © 1995 American Chemical Society. ]

Koch Glitsch Hetp

FIG. 14-60 Effect of structured packing surface areas, loads, and inclination angle on packing efficiency. Efficiency expressed as number of theoretical stages per meter, the reciprocal of HETP. Sulzer data, chloroben-zene-ethylbenzene, 100 mbar, at total reflux; 250-mm-diameter test column. (Reprinted courtesy of Sulzer Chemtech.)

FIG. 14-60 Effect of structured packing surface areas, loads, and inclination angle on packing efficiency. Efficiency expressed as number of theoretical stages per meter, the reciprocal of HETP. Sulzer data, chloroben-zene-ethylbenzene, 100 mbar, at total reflux; 250-mm-diameter test column. (Reprinted courtesy of Sulzer Chemtech.)

these data to the good, yet imperfect, distributors used in the real world. For poor distributors, the above rules of thumb will usually predict well below the HETPs measured in the field.

Lockett (Chem. Eng. Progr., p. 60, January 1998) simplified the fundamental Bravo-Fair-Rocha correlation [Ind. Eng. Chem. Res. 35, p. 1660 (1996)] to derive an alternative rule of thumb for structured packing efficiency. This rule of thumb predicts HETPs under perfect distribution conditions. Lockett recommends caution when applying this rule of thumb for aqueous systems as it does not predict the effects of underwetting (below).

Service-Oriented Rules of Thumb Strigle (Packed Tower Design and Applications, 2d ed., Gulf Publishing, Houston, Tex., 1994) proposed a multitude of rules of thumb as a function of the service, column pressure, and physical properties. These rules are based on the extensive experience of Strigle and the Norton Company (now merged with Koch-Glitsch LP).

Data Interpolation Interpolation of experimental HETP data is the most reliable means of obtaining design HETP values. This is hardly surprising in an area where our understanding of the theory is so poor that rules of thumb can do better than theoretical models. The author believes that it is best to derive HETP from experimental data, and to check it against a rule of thumb.

Eckert [Chem. Eng. Progr. 59(5), 76 (1963)], Chen (Chem. Eng. p. 40, March 5, 1984), and Vital et al. [Hydroc. Proc. 63(12), 75 (1984)] tabulated experimental HETP data for various random packings. Kister (Distillation Design, McGraw-Hill, 1992) extended these tabulations and included published HETP data and a detailed procedure for interpolating such HETP data. A prerequisite to any interpolation of packing data is thorough familiarity with the factors that affect HETP. Overlooking any ofthe factors listed can easily lead to poor interpolation and grossly incorrect design. In particular, it is imperative to recognize that the quality of distribution in pilot towers is generally superior to the quality of distribution in commercial towers.

Underwetting Laboratory- and pilot-scale distillation experiments with systems that exhibit large differences in surface tension along the column such as methanol-water showed a sharp drop in efficiency at the high-surface-tension end of the column [Ponter et al., Trans. Instn. Chem. Engineers [London], 45, T345 (1967)]. There appeared to be a critical methanol composition below which performance deteriorated rapidly. The poor performance at the low-methanol-concentration end appeared independent of the type and size of packing. Visual observations with disk columns attributed these effects to underwetting.

Underwetting is a packing surface phenomenon, which breaks up liquid film. The tendency of the liquid film to break (the degree of wetting) is expressed by the contact angle (Fig. 14-61). A contact angle of 0° indicates perfect wetting; an angle of 180° indicates no wetting. Mersmann and Deixler [Chem. Ing. Tech. 58(1), 19 (1986)] provide a preliminary chart for estimating contact angles. The contact angle depends on both the surface and the liquid and is a strong function of composition. In systems with large surface tension gradients, both contact angles and minimum wetting rates may vary rapidly with changes of composition or surface tension. Extensive studies by Pon-ter et al. [loc. cit.; also, Ponter and Au-Yeung, Chem. Ing. Tech., 56(9), 701 (1984)] showed that

• Underwetting is most significant in aqueous-organic systems, and tends to occur at the high-surface-tension (aqueous) end of the composition range. Liquid viscosity may also have an effect.

• Underwetting may be alleviated by changing the material and surface roughness of the packing.

• In systems susceptible to underwetting, column efficiency can sometimes (but not always) be improved by the addition of small amounts of surfactants.

Effect of Lambda Most packed-column efficiency testing has been at total reflux. Some tests for both random and structured packings [Billet, "Packed Towers Analysis and Design," Ruhr University, Bochum, Germany, 1989; Meier, Hunkeler, and Stocker, IChemE Symp. Ser. 56, 3.3/1 (1979); Eckert and Walter, Hydroc. Proc. 43(2), 107 (1964)] suggest that efficiencies at finite reflux are similar to those at total reflux when lambda (X = mGM/LM, which is the ratio of the slope of the equilibrium curve to the slope of the operating line) ranges between 0.5 and 2.0. This range is typical for most distillation systems.

Koshy and Rukovena [Hydroc. Proc., 65(5), 64 (1986)], experimenting with methanol-water and water-DMF using #25 IMTP pack-

Pressure Drop Pall Ring
FIG. 14-61 Contact angles. (a) Acute, good wetting. (b) Obtuse, poor wetting.

ing in a pilot-scale column, observed a sharp efficiency drop when the group X was greater than 2 or lower than 0.5. The efficiency loss escalated as X deviated more from this range. Koshy and Rukovena recognized that surface tension gradients and underwetting may have influenced some of their findings, but argue that the lambda effect is the major cause for the efficiency differences observed in their tests. High-relative-volatility systems are those most likely to be affected by X, because at low volatility, X ranges from 0.5 to 2. Strigle (loc. cit.) quantified the lambda effect on HETP using the following equation:

Actual HETP/standard HETP = 1 + 0.278[ABS(ln X)3] (14-160)

For 0.5 < X < 2, Eq. (14-160) gives a ratio of less than 1.1; that is, it has a little influence on HETP.

Pressure Generally, pressure has little effect on HETP of both random and structured packing, at least above 100 mbar abs. At deep vacuum (<100 mbar), there are data to suggest that efficiency decreases as pressure is lowered for random packings [Zelvinski, Titov, and Shalygin, Khim Tekhnol Topl. Masel. 12(10) (1966)], but most of these data can also be explained by poor wetting or maldistribution.

At high pressures (>15 to 20 bar), structured packing efficiency diminishes as pressure is raised (Fitz, Shariat, and Spiegel, Paper presented at the AlChE Spring National Meeting, Houston, Tex., March 1995). Zuiderweg and Nutter [IChemE Symp. 128, A481 (1992)] report the same phenomenon, but to a much smaller extent, also in random packings. They explain the efficiency reduction at higher pressure by vapor backmixing. Nooijen et al. [IChemE Symp. Ser. 142, 885 (1997)] concur, bringing a supporting case study in which high-pressure distillation efficiency improved by splitting a packed bed.

With structured packings (only) FRl's high-pressure (10 to 30 bar; flow parameters > 0.25) distillation tests measured maxima, termed humps in the HETP vs. load plot, typically at 65 to 90 percent of flood [Fitz, Shariat, and Kunesh, IChemE Symp. Ser. 142, 829 (1997); Cai et al., Trans IChemE 81, Part A, p. 85 (2003)]. These humps (Fig. 14-62) were not observed with lower pressure distillation (flow parameters < 0.2) and appeared to intensify with higher pressure. The humps did not always occur; some tests at different distributor positioning and with larger packing showed no humps. Zuiderweg et al. [IChemE Symp. Ser. 142,865 (1997); Trans. IChemE 81, Part A, p. 85 (January 2003)] and Nooijen et al. [IChemE Symp. Ser. 142, 885 (1997)] explain the humps by two-phase backmixing. At the high liquid loads in high-pressure distillation, flow unevenness prematurely floods some of the packing channels, carrying vapor bubbles downward and recirculating liquid upward.

Physical Properties Data presented by a number of workers [e.g., Vital, Grossel, and Olsen, Hydroc. Proc. 63(12), 75 (1984)] suggest that, generally, random packing HETP is relatively insensitive to system properties for nonaqueous systems. A survey of data in Chapter

Hetp Pall Rings
FIG. 14-62 HETPo data as measured in the FRI column for the iC4/nC4 system at different pressures (bara), showing efficiency humps. (From J. L. Nooijen, K. A. Kusters, and J. J. B. Pek, IChemE Symp. Ser. 142, p. 885, 1997. Reprinted courtesy of IChemE.)

11 of Kister's Distillation Design (McGraw-Hill, New York, 1992) leads to a similar conclusion for structured packings. For water-rich systems, packing HETPs tend to be much higher than for nonaqueous systems due to their high lambda or surface underwetting, as discussed above. High hydrogen concentrations (>30 percent or so in the gas) have also led to low packing efficiencies (Kister et al., Proc. 4th Ethylene Producers Conference, AIChE, New Orleans, La., p. 283, 1992), possibly due to the fast-moving hydrogen molecule dragging heavier molecules with it as it diffuses from a liquid film into the vapor.

Errors in VLE These affect packing HETP in the same way as they affect tray efficiency. The discussions and derivation earlier in this subsection apply equally to tray and packed towers.

Comparison of Various Packing Efficiencies for Absorption and Stripping In past editions of this handbook, extensive data on absorption/stripping systems were given. Emphasis was given to the following systems:

Ammonia-air-water Air-water

Sulfur dioxide-air-water Carbon dioxide-air-water

Liquid and gas phases contributing; chemical reaction contributing Gas phase controlling Liquid and gas phase controlling Liquid phase controlling

The reader may refer to the data in the 5 th edition. For the current work, emphasis will be given to one absorption system, carbon dioxide-air-caustic.

Carbon Dioxide-Air-Caustic System The vendors of packings have adopted this system as a "standard" for comparing the performance of different packing types and sizes for absorption/stripping. For tests, air containing 1.0 mol % CO2 is passed countercurrently to a circulating stream of sodium hydroxide solution. The initial concentration of NaOH in water is 1.0 N (4.0 wt %), and as the circulating NaOH is converted to sodium carbonate it is necessary to make a mass-transfer correction because of reduced mass-transfer rate in the liquid phase. The procedure has been described by Eckert et al. [Ind. Eng. Chem., 59(2), 41 (1967); Chem. Eng. Progr., 54(1), 790 (1958)]. An overall coefficient is measured using gas-phase (CO2) concentrations:

KOGae =

moles COS absorbed time-bed volume-partial pressure CO2 driving force

The coefficients are usually corrected to a hydroxide conversion of 25 percent at 24°C. For other conversions, Fig. 14-14 may be used. Reported values of KoGa for representative random packings are given in Table 14-15. The effect of liquid rate on the coefficient is shown in Fig. 14-63.

While the carbon dioxide/caustic test method has become accepted, one should use the results with caution. The chemical reaction masks

TABLE 14-15 Overall Coefficients for Representative Packings

Was this article helpful?

0 0

Responses

  • Abbondanzio
    What is hetp of pall rings?
    2 years ago

Post a comment