Minimum Wetting Rate Of A Mellapax 250y

-*-

PCT of useable capacity

PCT of useable capacity

An implication of this observation is that it is important to follow good redistribution practices. If a tall bed is unavoidable, it may be beneficial to attempt using larger packings in it.

9. Liquid maldistribution lowers packing turndown (15,160,161). The two upper "standard distributor" curves in Fig. 8.166 represent a progressively lower quality of initial liquid distribution. The diagram shows that packing turndown becomes progressively poorer with greater maldistribution. Figure 9.76 provides further illustration of this turndown loss. 10. Vapor distribution can be troublesome, especially in large-diameter columns. This type of maldistribution is best tackled at the source by paying attention to the vapor inlet arrangements. Common commercial vapor distributor designs are discussed elsewhere (23,40,152).

9.3 Packed-Tower Scaleup 9.3.1 Diameter considerations

For random packings there are many reports (3,56,98,136,140a, 162-167) of an increase in HETP with column diameter, fewer in which column diameter had little effect on HETP (3,56,162,163), and an odd case in which increasing column diameter led to a decrease in HETP (56). Billet and Mackowiak's (3,62,166) scaleup chart for Pall® rings implies that efficiency decreases as column diameter increases.

For structured packings, tests with the Koch-Sulzer BX and CY wire-mesh packings and with Mellapak 250Y showed no diameter effect on HETP (3,19,32,33). Wu and Chen (167) state that the same applies to Gempak packings. On the other hand, tests with Hyperfil showed an increase in HETP with column diameter (168).

The increase in HETP with column diameter appears to be most pronounced for small column diameters (< 1 ft; Refs. 56,164) and with smaller packings (<1 in; Refs. 56,136). It has therefore been recommended that for scaleup purposes, a column at least 1 ft in diameter, and preferably larger, should be used (164).

Practically all sources explain the increase of HETP with column diameter in terms of enhanced maldistribution (3,56,62,136,148,162-164). Pilot columns seldom operate at column to packing diameters ratios (Dx/Dp) larger than 20; under these conditions, lateral mixing effectively counteracts loss of efficiency due to maldistribution pinch (Sec. 9.2.3). On the other hand, industrial-scale columns usually operate at DTlDp ratios of 30 to 100; under these conditions, lateral mixing is ineffective for counteracting maldistribution pinch.

In order to increase DjjDp, it may appear attractive to perform the pilot test using a smaller packing size than will be used in the prototype. The measured HETP can then be adjusted for the difference in packing size using a theoretical model or a rule of thumb (Sees. 9.1.4 and 9.1.5). A successful application of this technique was reported for wire-mesh packings (169). In practice, however, the technique of scaling up from data for a smaller packing is seldom used (170). The author also strongly recommends against it. There is a considerable uncertainty in predicting the effect of packing size on HETP. Further, lateral spread (Sec. 9.2.4) is a strong function of packing size, and scaling up its effects is practically impossible.

In small columns, the wetted wall contributes to mass transfer. This problem can be overcome by keeping the ratio of column to packing diameter above 10 (Sec. 9.2.4). For cases where wall effects are significant, Wu and Chen (167) recommend applying the following safety factor.

HETPpUot (1 - AwlJ/Apacki0g)

The area ratio is given by

Awau ttDt H 4

apacking tt/4 D% H ap DTap

As column diameter is increased, it becomes more difficult to maintain the number of distributor drip points per unit area (3,40). Also, the fraction of unirrigated area under the top distributor (near the column wall) may vary. These changes can lead to enhancement of maldistribution in the prototype. It was recommended (170) to use the same number of drip points per unit area and to ensure that liquid is distributed to the column wall both in the prototype and in the pilot column.

Packing a pilot column is easier than packing a much larger column. A better packing technique in the pilot column will give better HETP. Billet (56) presents an example in which the opposite occurred, and the larger column achieved a better HETP than the smaller. It is best to pack the pilot column using the same technique that is intended for the prototype.

9.3.2 Height, loading, wetting, and other considerations

Height effects. Experimental data for random packings show that HETP slightly increases with bed depth (3,56,61,98,144). This increase in HETP appears to be more important at low liquid rates

(61,98). Using the zone-stage model (Sec. 9.2.5), Zuiderweg et al. (136) predicted a slight effect of bed depth on HETP for larger packings, but a large effect for smaller packings.

For structured packing, tests with Mellapak 250Y (32) showed no effect of bed height on packing efficiency. Tests by Martin et al. (171) using several structured packings in a 10-ft-tall bed showed that in some cases, the lower part of the bed has a higher HETP than the upper part, while in other cases, HETP was uniform throughout the bed.

The effect of bed depth on packing HETP is attributed to liquid maldistribution (56,136,171). Zuiderweg et al. (136) suggest that the uneven liquid flow generates an uneven concentration profile and localized pinching near the bottom of the beds. The tests by Martin et aL (171) confirm that the problem area is near the bottom. According to the zone-stage model (Sec. 9.2.5), as well as the empirical correlation by Moore and Rukovena (Sec. 9.2.6), the greater the number of stages per bed, the greater is the rise in HETP with bed depth. The presence and extent of maldistribution plays an important role in determining the bed-depth effect (136,140).

As the bed depth increases, end effects (i.e., mass transfer in the region of liquid introduction and in the region where liquid drips from the packing supports) become less important. Such end effects tend to lower the HETP observed in short columns, such as pilot-plant columns.

In summary, bed depth may significantly influence HETP. This adds uncertainty to scaleup. Shallow test beds should be avoided. Most investigators use beds at least 5 ft tall, and often more than 10 ft tall-One suggestion (172) is to test the pilot columns at different heights of packings. The FRI sampling technique (below) can detect maldistribution along the bed height.

Loadings. For many random and corrugated-sheet structured packings, HETP is independent of vapor and liquid loadings (Fig. 8.16a). For wire-mesh and many corrugated-sheet structured packings, HETP increases with vapor and liquid loads (Fig. 8.16c). For some packings, HETP decreases as vapor and liquid loads increase (Fig. 8.16d).

Wu and Chen (167) recommend pilot testing over the entire range between the expected minimum and maximum operating rates, and taking the highest measured HETP as the basis for scaleup. The author concurs. With structured packings, the load effect may be due to liquid rather than vapor loads, and the pilot tests should cover the range of liquid loads (i.e., gpm per square foot of column cross section) that is expected in the prototype.

Wetting. At low liquid rates, the onset of minimum wetting can affect scaleup, particularly with random packings and aqueous systems. The criteria in Sec. 8.2.15 should be examined to check if operation near minimum wetting is likely. If it is, scaleup can be unreliable. The unreliability can be alleviated by pilot-testing at the composition range expected in the prototype, and by using identical packing materials and surface treatment in the pilot tests and in the prototype.

Underwetting (Sec. 8.2.16). At the aqueous end of aqueous-organic columns, underwetting is most significant. Near the onset of under-wetting, HETP becomes strongly dependent on composition, packing material and surface roughness, and the presence of surfactants. When underwetting is likely, scaleup can be unreliable. The unreliability can be alleviated by pilot-testing at the composition range expected in the prototype, and by using identical packing materials and surface treatment in the pilot tests and in the prototype.

Preflooding. For one structured packing test with an aqueous system, Billet (3) measured higher efficiency for a preflooded bed compared to a non-preflooded bed. Presumably, the preflooding improved either wetting or distribution. Section 9.2.4 addresses the effect of preflooding on distribution. Billet (3) recommends preflooding the packing both in the prototype and in the pilot column to ensure maximum efficiency.

Sampling. Fractionation Research Inc. (FBI) developed a sampling technique that eliminates the influence of "end effects" and detects a maldistributed composition profile. This technique (75,88,131,160) samples the bed at frequent intervals, typically every 2 ft or so. HETP is determined from a plot of these interbed samples rather than from the top and bottom compositions.

It is imperative that the interbed samplers catch representative samples, which are an average through the bed cross section. Caution is required when the liquid is highly aerated and turbulent (e.g., above 200 psia or above 20 gpm/ft2). The author highly recommends the FRI sampling technique for all other conditions.

Aging. Billet (3) showed that for some plastic packings in aqueous systems, the efficiency after one week's operation was almost double the efficiency of new packings. Little further change was observed after one week. Billet explains the phenomenon by improved wetting. He recommends that data for plastic packings should only be used for scaieup after being in operation for an adequately long period.

9.3.3 Packed-tower scaieup: summary and recommendations

In general, the key to a successful scaieup of packed columns is producing identical distribution conditions in the pilot column to those expected in the prototype. Sections 9.2.1 to 9.3.2 show that this is practically impossible to achieve. Packed-column scaieup is therefore often uncertain and sometimes dangerous. For best results, one can only do the best he or she can. Based on Sees. 9.3.1 and 9.3.2, this includes

■ Use the same packing type and size in the pilot column as those that will be used in the prototype.

■ Use a pilot column at least 1 ft in diameter. j

■ Use a column to packing diameter ratio of at least 10; alternatively, ] correct HETP for wall effects using Eq. (9.38). i

■ In the pilot column distributor, use the same number of drip points per unit area as in the prototype and ensure that liquid is distributed to the wall in the same manner as in the prototype.

■ Pack the pilot column using the same packing technique as in the prototype.

■ Use a bed at least 5 ft tall, preferably at least 10 ft tall, in the pilot column.

■ Use the FRI sampling technique and samplers for calculating pilot column HETP.

■ Perform pilot tests over the entire range of vapor and liquid loads between the expected minimum and maximum operating rates. Use the highest measured HETP as the basis for scaieup.

■ Be on the watch for wetting and underwetting effects. When these are likely, scaieup can be particularly unreliable.

■ When wetting or underwetting effects are likely, pilot-test at the composition range expected in the prototype. Also, use identical materials of construction and surface treatment in the pilot tests and the prototype.

■ Compare the pilot HETP under preflooded conditions to the efficiency measured under non-preflooded conditions.

■ Ensure plastic packings have been properly aged prior to the pilot test.

Safety factor. Even if the above techniques are followed, some uncertainty still remains. To allow for this uncertainty, it has been recommended (98,120) to add 6 to 12 in to the HETP measured in small-scale columns, and possibly more in vacuum columns operating at low liquid rates. This recommendation has been criticized (163) for being too conservative. An alternative recommendation (167) is to add a 10 to 15 percent safety factor.

9.4 Packed-Column Sizing 9.4.1 Strategy

Most modern packings are proprietary. Optimum column sizing is an interactive process between the user (or the user's representative) and the manufacturer, beginning at an early stage of the design. An effective user-manufacturer interactive procedure is;

1. Survey process factors that may constrain packed-tower design. These include feed solids content, fouling, corrosion, foaming, potential for pressure surges, startup and shutdown procedures, the possibility of pyrophoric deposits and hot oil remaining on packing surfaces at shutdown, and the possibility of precipitation and adverse reactions. These influence the choice of packing and the distributor design.

Several troublesome experiences resulted from insufficient attention to this guideline (40). These experiences included plugging, premature flooding, poor separation, fires, dislodging of packed beds, severe corrosion, melting of plastic packing, breakage of ceramic packing, loss of capacity, and others.

2. Perform a preliminary packed-tower sizing. It is best to base this step on a packing for which reliable information is available near the expected operating point, e,g., on Pall® rings for random packings. The preliminary design should be rigorous, not shortcut. A column built as per this design must work. The only feature distinguishing this preliminary design from the final design is that at this preliminary point the column may be somewhat oversized, i.e., nonoptimum.

The preliminary design provides a basis for deciding which manufacturers, which packings, and what packing sizes should be considered. The preliminary diameter and bed height provide a suitable basis for comparing quotes. An "apples-to-apples" comparison of different quotes is extremely difficult when column height and diameter vary.

3. Prepare an inquiry specification with a quote request. This document should communicate to the suppliers information about the service, the design duty, requirements and constraints, and solicit bids that can readily be compared.

■ Include all data, constraints, or requirements arising from step 1 above. If uncertain about how fouling, potential for pressure surges, startup and shutdown constraints, etc., affect the packing design, discuss these with the suppliers.

■ Include all relevant process information as well as the preliminary column diameter, bed heights, and spaces between beds. Request a base quote using the preliminary design. Do not specify packing type and size, but list any constraints such as "we require random packing, 1.5 to 2.5 inch nominal size or equivalent." Request the suppliers to take exception to any details that in their opinion will not work.

■ Request the supplier to prepare an alternate quote for a design that can reduce the column diameter or bed heights. Specify whether diameter reduction or bed height reduction is more desirable.

4. Once the quotes are received, compile a "short list" of the more attractive bids. Evaluate the alternate quotes by comparing to experimental data, correlation predictions, or sound design criteria-Chapters 10 and 11 contain all the published information the author could find on proprietary packings. Additional information may be available in proprietary data banks. If your data are insufficient to form a judgment, question the supplier's design basis and ask for backup data.

Often, the evaluation reveals a sound alternate quote with too tight a design. Modify the design to one that you are comfortable with, and request the supplier to revise the alternative quote. The author strongly believes that the user must be confident that the alternate design will perform adequately before accepting it. The confidence must be based on engineering rather than on sales claims alone. Any problems must be discussed with the supplier.

5. Once comfortable with the quoted designs (modified per step 4 above) those quotes should be analyzed economically to determine which one is best.

9.4.2 Column sizing example

Example 9.2 For the depropanizer in Examples 2.4 (Sec. 2.3.1), 3.4 (Sec. 3.2.51. and 6.1 {Sec. 6.5.2), would a packed tower be better than a tray tower? Proceai loads and physical properties are the same as those in Table 6.10. The service m nonfouling, the streams have a negligible solid content, the corrosive tendency is low, and pressure surges are unlikely.

solution Section 8.3 compares the advantages of packed and trayed towers. A study of the items listed in this section reveals no clear frontrunner. The high tendency to foam will give a slight edge to packings; the high liquid loads and the greater uncertainty in performance prediction will give a slight edge to trays. There is a need to size the column for packings and compare to the trayed tower design (Sec. €.5.11). The procedure in Sec. 9.4.1 will be used.

step 1 considerations There are no apparent constraints imposed by solid, corrosion, and pressure surge potential. The foaming potential can be inferred from tray column information. Table 6.7 classifies depropanizers as a low-foaming system, and suggests a derating factor of 0.9 for tray columns. Tables 6.4 and 6.5 suggest a somewhat higher foaming tendency, but these specifically address downcomers. It appears that for the current design, the system is best considered low-foaming, with a derating factor of 0.9 as per Table 6.7.

Other considerations in step 1 are the startup-shutdown procedure. The startup sequence usually involves the steps of steaming, purging, and pressuring up prior to feed introduction; the shutdown sequence is the reverse. Of these steps, steaming is a hot-commissioning operation and may constrain the choice of materials of construction. Pyrophoric deposits are unlikely, nor is hot oil on packing surfaces (the heaviest hydrocarbon is in the Cs range), precipitation and adverse reaction. In summary, the relevant step 2 considerations are foaming and startup-shutdown steaming.

step 2 considerations The next step is to look at the various types of packings. The depropanizer is a high-pressure distillation service. Section 8.1.10 recommends against using structured packings for high pressure distillation. Grids are seldom used for clean distillation services (Sec. 8.1.13). Random packings are therefore clear frontrunners for the depropanizer. With its low corrosion potential, carbon steel is suitable for the packing. Plastic is unlikely to offer any distinct advantages here, and is at a disadvantage because of the steaming step at startups and shutdowns.

As with tray towers (Sec. 6.5.1), the sizing procedure involves a trial-and-error calculation, in which a preliminary design is first set and refined by checking against performance correlations. As with tray towers, the sizing calculations are performed at the points listed in Sec, 6.5,1—i.e., the highest and lowest loaded points in each section of tower. For the depropanizer, the highest loadings are on stages 19 (bottom) and 3 (top) and the lowest are on stages 9 (bottom) and 8 (top), as shown in Table 6.10. These points will be used in the sizing calculation.

9.4.3 Column sizing example: first trial

As per step 2 (Sec. 9.4.1) the preliminary design will use carbon steel Pall® rings. This design can later be refined by working with suppliers on the use of third-generation packings (Step 4). As an initial estimate, it will be assumed that 2-in Pall® rings will be used throughout the column. This assumption will be reviewed later. For this packing, the flood point can be accurately determined by interpolation (Sec. 8.2.6), using Chart 10.1004A. Alternatively, the flood point can be determined using the Kister and Gill correlation [Eq. (8.1)]. The GPDC flood correlation (Fig. 8.17) is not suitable for 2-in Pall® rings.

Was this article helpful?

0 0

Post a comment