## Pressure drop inherent limitations and traps

Pressure drop This is often used to specify packed tower capacity. The application of this criterion, and the interpretation of packing pressure drop data are not trouble-free. Some inherent limitations and traps are

1. In small columns (<3 ft in diameter) pressure drop varies with tower diameter (3,56). With random packings, the smaller the tower diameter, the lower the pressure drop, possibly due to enhancement of wall effects. Since the bulk of the published pressure drop data were obtained in pilot-scale columns, a correlation based on these data will give optimistic predictions for commercial-scale columns. The magnitude of the diameter effect on pressure drop (at fixed operating conditions) varies with packing size and geometry, with no coherent trend. For 2-in metal Pall® rings, Billet (3,62) measured a 20 to 30 percent higher dry pressure drop in a 2 ft 7 in diameter column than in a 9-in diameter column. With Vs-in Pall® rings, the corresponding difference was only of the order of 5 percent (56). This diameter effect may intensify when column diameter diminishes below 9 in (3,11). With some corrugated structured packings, the reverse effect was observed (3,31,32), e.g., pressure drop about 10 to 20 percent lower in a 3-ft-diameter column than in a 1-ft-diameter column.

2. Dry-packed beds have higher pressure drops than wet-packed beds. Billet (56) and Kister (40) report cases where changing from dry to wet packing increased column capacity by 5 percent and reduced pressure drop by 10 percent. Ludwig (63) reports cases where this pressure drop reduction was 50 to 60 percent. The packing method is seldom reported by literature sources presenting test data.

3. Packing particles may be nonuniform. Kunesh (51) shows size variation among ceramic saddles of the same nominal size that came out of a single box. Further, for the same nominal packing type and size, there are differences in shape, size, and particle thickness among packings produced by different manufacturing processes. All these have an effect on pressure drop.

4. Pressure drop is a function of vapor and liquid velocities, densities, and physical properties. The velocities and densities vaty with changes in compositions along the packed bed. For test systems where these variations are large, pressure drop data may contain significant errors.

5. Pressure drops measured under deep vacuum (<50 mmHg) are affected by the pressure drop and the pressure gradient along the bed. Consider a 10-ft-tall packed bed operating at 20 mmHg at the top with a pressure drop of 0.5 in of water per foot of packing. The pressure at the bottom of this bed will be 29 mmHg. The velocity and pressure drop at the bottom of the bed will be 31 percent lower than at the top. The average pressure drop per foot of bed will be roughly 84 percent of the pressure drop that would have been measured for a veiy short bed with the same top pressure.

6. Pressure drop measurements in pressure towers include the static head of the vapor. To obtain the actual packing pressure drop, the static head must be subtracted from the pressure drop measurement. At high pressure, the static head is often higher than the packing pressure drop, making the pressure drop a difference between two large numbers. Most pressure drop data are not corrected for static vapor head. Data sources seldom state whether such a correction has been performed. This may lead to serious errors, especially for low-pressure-drop packings, even at pressures as low as atmospheric. For an atmospheric system with a molecular weight of 100, boiling at about 200°F, the static vapor head is about 0.04 in of water per foot of packing. This static head can lead up to a 40 percent error when the measured pressure drop is 0.1 in of water per foot of packing.

7. In some cases (65), it has been observed that nonuniform liquid distribution can lead to lower pressure drop than under uniform distribution. MacDougall (58) postulated that liquid maldistribution can generate gas maldistribution, which in turn will affect pressure drop. MacDougall extended this argument to cast suspicion on pressure drop data measured at liquid rates be low the distributor turndown limit. Recent experiments by Sohlo and Kouri (66) and Stoter et al. (67) confirmed that liquid maldistribution can generate gas maldistribution at high gas rates, but not at low gas rates. Combining Sohlo and Kouri's findings with MacDougall's arguments suggest that data obtained at high vapor rates and turned-down liquid rates are likely to give optimistic pressure drops.

8. For ceramic packings, reported pressure drops are for unchipped and unbroken pieces. In practice, breakage and chipping of ceramic packings occur in shipment, when loading packings into the column and during operation. Small pieces and chips of packings reduce the cross-section area and increase pressure drop. This may increase a pressure drop by 15 to 75 percent (63).

9. Plastic and thin-wall metal packings deflect with time and temperature and compress. The problem is aggravated in tall beds and where a plastic approaches its maximum use temperature. The compression raises pressure drops. The rise in pressure drop, however, is generally smaller than that due to ceramics breakage.

10. Pressure drops for foaming systems are higher than for non-foaming systems (1). Most published pressure drop data are for nonfoaming systems, and will give optimistic estimates when applied to foaming systems. Eckert (1) presents suggestions for extending pressure drop data to foaming systems.

## Post a comment