Schmidt's correlation is sensitive to the contact angle estimate [both in Eqs. (8.35) and (8.38)]. Table 8.5 lists the range of applicability of Schmidt's correlation. The Schmidt correlation is based on data for Raschig and Pall® rings (102), mainly for positive-surface-tension systems (58).

Prediction by rule of thumb. Popular rules of thumb for minimum wetting rates in random packings are table 8.5 Range of Applicability of Schmidt's Minimum Wetting Rate Correlation (102)

Pressure, mmHgabs 15-760

Liquid falling film number, CL 2.44 x 10®-€.77 x 10l°

Contact angle <|>, degrees 8-20

Packing material Ceramic, copper, or stainless steel

A rule of thumb cited by Ludwig (63)

This rule is conservative; a more realistic range for minimum wetting rates is 0.5 to 2 gpm/ft2. In some nonaqueous services, liquid rates sometimes as low as 0.1 to 0.25 gpm/ft2 are successfully handled. This rule applies only where underwetting (Sec. 8.2.16) is not a problem.

Glitsch's rule of thumb (Table 8.6) for CMR® random packings when the packing surface area exceeds 43 ft2/ft3. Both Ludwig (63) and the author like these rules, but they only apply to CMR® #1, #1.5, and #2. To extend these rules to other packings, the author applied Schmidt's model (102). Equation (8.38) gives

The surface area per unit volume is about 75 ft2/ft3 for CMR® #1; 57 ft2/ft3 for CMR® #1.5, and 45 for CMR® #2. It is conceivable that Glitsch's rules are based on an average value of ap of about 60. Equation (8.40) will then become

TABLE 8.6 Glitsch's Rule of Thumb for Minimum Wetting (6,63,103) (Basl*: CMR* wttha, > 43 f^/ft3)

Minimum wetting

Material rate, gpm/ft2

Unglazed ceramic (chemical stoneware) 0.2

Oxidized metal (carbon steel, copper) 0.3

Surface-treated metal (etched stainless steel) 0.4

Glazed ceramic 0.8

Glass 1.0

Bright metal (stainless steel, tantalum, other alloys) 1.2


Polypropylene 1.6

Fluoropolymers (PTFE type) 2.0

Experiments by Chuang and Miller (34) confirm good performance of #1 and #2 metallic, surface-treated CMR® and #1 Hy-Pak® at liquid rates of the order of 0.3 to 0.8 gpm/ft2 with an aqueous system. Their experiments also showed inferior performance when surface treatment was inadequate. This gives support to the metallic values listed in Table 8.6, but emphasizes that "surface-treated" and "oxidized metal" refer to proper conditioning of packing surfaces.

Structured packings. Superior wetting characteristics compared to random packings (Sec. 8.1.10) characterize structured packings. With metal packings, satisfactory performance was reported down to 0.1 gpm/ft in corrugated-sheet structured packings, and down to 0.05 gpm/ft2 for wire-mesh structured packings.

Reducing minimum wetting rates. Surface treatment of the packing can substantially reduce the MWR (34,63). Techniques of conditioning packing surfaces include oxidizing, sandblasting, and etching. The improvement depends on the technique and its effectiveness. Chuang and Miller (34) tested a metallic random packing with an aqueous system at low liquid rates (about 0.3 gpm/ft ). They used two alternative techniques for oxidizing the packing surfaces. The packings oxidized with the more effective technique gave a column efficiency twice as high as those oxidized by the less effective technique.

8.2.16 Underwettlng

Laboratory- and pilot-scale distillation experiments with systems that exhibit large differences in surface tension along the column showed a sharp drop in efficiency at the high-surface-tension end of the column. Distilling the methanol-water system, Ross, Ponter et al. (36,104) observed high efficiencies at high and moderate methanol concentrations (low surface tension) and low efficiencies at low methanol concentrations (high surface tension). There appeared to be a critical methanol composition below which performance deteriorated rapidly. This was observed with several types and sizes of random packings and one wire-mesh structured packing (Hyperfil®). The poor performance at the low-methanol-concentration end appeared independent of the type and size of packing. Visual observations with disc 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 wet ting) is expressed by the contact angle (Fig. 8.22). A contact angle of zero indicates perfect wetting; an angle of 180° indicates no wetting. Mersmann and Deixler (926) provide a preliminary chart for estimating contact angles. The contact angle depends both on the surface and the liquid (926,104,105) 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 (104,105). Liquid viscosity (106,107) and the surface tension gradient (108) may also have an effect. Reference 105 reviews most of the work reported on underwetting. Extensive studies by Ponter et al. (104,105,109) showed that

■ The underwetting effects are most significant in aqueous-organic systems, and tend to occur at the high-surface-tension (aqueous) end of the composition range.

■ Changing the material and surface roughness of the packing may significantly affect efficiency in systems susceptible to underwetting.

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

Koshy and Rukovena (110,111) extensively experimented with the methanol-water and water-DMF systems using #25 EMTP® packing in a pilot-scale column. They proposed an alternative theory to explain the low efficiencies observed at the water-rich end of the columns. Accordingly, packing efficiency varies with the group X (= mG'f L'\ which is the ratio of the slope of the equilibrium curve to the slope of the operating line. When X is close to unity, it has little effect on efficiency. As X deviates from unity (roughly, for values of X greater than 3 or lower than 0.5 for the systems studied), efficiency drops, sometimes dramatically. Koshy and Rukovena recognize that surface tension gradients may influence the differences in efficiency between widely different systems, but argue that these gradients are not the major cause for efficiency differences observed for a single system.

If the Koshy and Rukovena theory is correct, then only high-relative-volatility systems are likely to be affected by X, because M

Flgur« 8.22 Contact angle», (a) Acute, good wetting; (fr) obtuse, poor wetting.

low volatility, X ranges from Vz to 2. For high-volatility systems, the drop in efficiency will be greatest when a high reflux ratio is used, and is likely to occur near the high-purity end of the column.

The underwetting theory appears to be supported by more evidence compared to the X theory. The author analyzed Koshy and Rukovena's data (110,111) and believes that all their observations can be explained in terms of the underwetting theory. Further, the analysis showed that Koshy and Rukovena's results were in striking agreement with those of Ross, Ponter et al. (36,104), and even the critical methanol concentration appeared the same.

8.2.17 Minimum vapor rate

When vapor velocity through a packed bed is excessively low, the following adverse effects may be encountered,

1. Vapor maldistribution: Packing pressure drop places a resistance in the vapor path that helps spread the vapor radially. If pressure drop is too low, vapor will tend to channel through the bed, leading to poor mass transfer.

2. Laminar vapor flow: This will tend to reduce mass transfer.

For these reasons, it is a common practice (15,52,63) to design random packed beds for a pressure drop not smaller than 0.1 in of water per foot of packing. In practice, there are many columns that operate efficiently at a lower pressure drop.

8.3 Comparing Trays and Packings

Almost every separation can be performed either with trays or with packings. The factors discussed below influence the choice between trays and packings. These factors only represent economic pros and cons, and each may be overridden. For instance, column complexity is a factor favoring trays, but gas plant demethanizers that often use one or more interreboilers are traditionally packed.

A couple of trays-versus-packing comparisons have appeared in the recent literature (31o,lllo). These comparisons emphasize capacity, efficiency, and costs. Both are based on several assumptions that the author would strongly challenge. For this reason, the author would not recommend applying conclusions from these comparisons in any design decision. Instead, the author strongly recommends analyzing each design case on its own merits. An excellent discussion of the pros and cons of trays and packings was presented by Thibodeaux and Murill (1116). This discussion is updated and expanded on below.

8.3.1 Factors favoring packed columns

Vacuum systems. Packing pressure drop is much lower than trays. In a tray tower, the open area of each tray is typically 5 to 10 percent of the column cross-section area. Each tray is analogous to a restriction orifice of an area ratio of 10 to 20 to 1. This gives a high velocity and a high pressure drop. Further, each tray typically holds 1 to 2 inches of clear liquid, through which the vapor must pass, and which incurs further pressure drop. Tray pressure drop is typically of the order of 0.15 psi per theoretical stage. On the other hand, the open area in a packed tower is usually greater than 50 percent of the column cross-section area, and liquid resistance to gas flow is relatively small. This leads to a typical pressure drop of 0.04 psi per theoretical stage with random packings and about half of that with structured packings.

Consider a vacuum column with 10 theoretical stages, operating at 1 psi top pressure. The bottom pressure will be 2.5 psi with trays, but only 1.4 psi with packings. The packed tower will have a much better relative volatility in the lower parts, thus reducing reflux and reboil requirements and bottom temperature. This means less product degradation, more capacity, and smaller energy consumption. This is a major advantage for packings.

Low-pressure-drop applications. By virtue of their low pressure drop compared to trays (above), packings are favored in any application where it is economical to minimize pressure drop. A typical example is an atmospheric or low-pressure column whose overheads are compressed. Every pound per square inch of pressure drop here translates into greater compression ratio requirement and higher compressor capital and energy costs.

Revamps. The pressure drop advantage that packing has over trays is invaluable in vacuum column revamps. By optimizing the revamp de- j sign pressure, the pressure drop reduction can be translated into a capacity gain, an energy gain, separation improvement, or various com* binations of these benefits.

Packings also offer easy trade-off between capacity and separation. Going to smaller packings converts spare capacity in the tower inta separation stages. Larger packings can overcome capacity bottleneck» at the expense of loss in separation. If both of these can be performed in different sections of the same column, and assuming no pinch near the feed, capacity or separation or both can be improved with little penalty. In other situations, a separation lo^s due to revamping with larger packings can be compensated by a slight, almost unnoticeable increase in reflux. In tray columns, changing tray spacing will give similar results, but is more difficult to do.

Small-diameter columns. When column diameter is less than 3 ft, it is difficult to access the column from inside in order to install and maintain the trays. "Cartridge" trays are often installed, or an oversized diameter is used. Either option is expensive. Cartridge trays also run into leakage and hold-down problems (40). Packing is normally a cheaper and more desirable alternative.

Corrosive systems. The range of packing materials is wider than that commonly available for trays. Ceramic and plastic packings are cheap and effective. Trays can be manufactured in nonmetals, but packing is usually a cheaper and more desirable alternative.

Foaming (and emulsion). The foaming tendency is greater on trays than with packings due to the higher vapor and liquid velocities and the more violent vapor-liquid contact. The advantage of packings with foaming systems is small, but often appears exaggerated due to poor downcomer sizing practices.

Low liquid holdup. Packings generally have lower liquid holdup than trays. This is often advantageous either for reducing polymerization and degradation or as a safety measure aimed at reducing the inventory of hazardous materials.

Batch distillation. Because of the smaller liquid holdup of packed columns, a higher percentage of the liquid can be recovered as top product.

6.3.2 Factors favoring tray columns

The following factors generally favor trays compared to either random or structured packings:

Solids. Trays can handle solids a lot easier than packed columns. Both gas and liquid velocities are often an order of magnitude higher on a tray than through packings. These high liquid and gas velocities provide a sweeping action that keeps tray openings and perforations clear. Solids tend to accumulate in the voids of packed columns. There are fewer locations where solids can be deposited in a tray column. Further, packed towers need liquid distributors, and plugging in these has been a common troublespot. Cleaning trays is easier than cleaning random packings, while cleaning structured packings is practically impossible.

High liquid rates. Multipass trays effectively lower the liquid load "seen" by each part of the tray. A similar trick cannot be applied with packed towers; the capacity of packings, especially structured, tends to rapidly fall off at high liquid rates. It is often more economical to handle high liquid rates in tray columns.

Large diameter. Packings are prone to severe maldistribution problems in large-diameter columns. These problems are far less severe in plate columns.

Complex columns. Interreboilers, intercondensers, cooling coils, and side drawoffs are more easily incorporated in tray than in packed columns. In packed columns, every complexity requires additional distribution and/or liquid collection equipment.

Feed composition variation. One way of allowing for design uncertainties and feedstock variation is by installing alternate feed points. In packed columns, every alternate feed point requires expensive distribution equipment.

Performance prediction. There is greater uncertainty in predicting packed column performance. Greater overdesign is often required.

Chemical reaction/absorption. By using high weirs, trays are capable of providing greater residence time for absorption or chemical reaction than packing.

Weight Tray columns usually weigh less than packed columns. This saves on the cost of foundations, supports, and column shell.

Intermittent operation. When temperature is either lower or *

than atmospheric, intermittent operation repeatedly expands and cos-tracts the shell. This may crush the packings or damage the shell in a packed column, but is easy to accommodate for in tray columns.

Trays versus random packings. The following factors generally £s*v trays compared to random packings, but usually not compared (■ structured packings.

Low liquid rate». With the aid of serrated weirs, splash baffles, reverse-flow trays (40), and bubble-cap trays, low liquid rates can be handled better in tray columns. Packed columns suffer from liquid dewetting and maldistribution at low liquid rates.

Turndown. Valve and bubble-cap trays normally give better turndown than packings. Unless very expensive distributors are used, packed tower turndown is usually limited by distributor turndown. With random packings, dewetting may also limit turndown.

Process surges. Random packings are usually more troublesome than trays in services that suffer from frequent process surges (e.g., those caused by slugs of water entering a hot oil column, relief valve lifting, compressor surges, or instability of liquid seal loops). Structured packings are considered to be less troublesome than trays in such services.

Traya versus structured packings. The following factors generally favor trays compared to structured packings, but usually not compared to random packings.

Materials of construction. Due to the thin sheets used in structured packings, their materials of construction need to have better resistance for oxidation or corrosion. For a service in which carbon steel is usually satisfactory with trays, stainless steel may be required with structured packings.

Column wall Inspection. With structured packings, it is often difficult to inspect the column wall without damaging the structured packings. Due to their snug fit, structured packings are easily damaged during removal.

Washing and purging. Thorough removal of residual liquid, wash water, air, or process gas trapped in structured packings at startup and shutdown is more difficult than with trays. Inadequate removal of these fluids may be hazardous.

0 0

Post a comment