Strigle Packed Tower

Figure 8.18 [Continued) GPDC interpolation plots for structured packings flood points, (d) The Dolan, Hausch and Petschauer plot for Intalox* 2T Structured Packing. (Part d, from G.W. Hausch, P. Quotson, and K. Seeger, "Structured Packing Revamp of a 306 Psig Depropanizer," paper presented at the AJChB Annual meeting, Los Angeles, California, November 1991; reprinted courtesy of the Norton Company.)

vacuum and low superatmospheric pressures, and when the appropriate constants are available, the Billet and Schultes correlation is also recommended.

8.2.7 Maximum operational capacity (MOC) prediction

MOC correlations for proprietary packings are sometimes available in the manufacturer literature (e.g., 15,30) or on manufacturers' com puter disks. The only other published correlation is by Kister and Gill (60,60a). This correlation simply states

where «n ^ evaluated from Eq. (8.1). This correlation was shown (60,60a) to predict practically all published MOC data to within ± 20 percent and most published MOC data are within ±15 percent, for both random and structured packings.

For some high-capacity structured packings of unique geometry, such as the Norton Intalox® 2T and the Jaeger MaxPac®, combining Eq. (8.1) with Eq. (8.10) predicts MOCs that are consistently 10 to 15 percent low. Although this is well within the correlation accuracy, it suggests that the correlation does not distinguish the unique high-capacity features of these two structured packings.

8.2.8 Pressure drop prediction by correlation

Limitations of packed-tower pressure drop correlations are systematic rather than random. It has been demonstrated (60) that a correlation that gives excellent statistical fit to experimental data can give poor predictions for many situations commonly encountered in industrial practice. This anomaly originates from a bias in pressure drop data banks toward the air-water system, and from a data shortage for nonaqueous systems at pressure, vacuum, and high and low liquid rates. Further, a correlation may work well for the majority of packing types and poorly for only a few, but the user can rarely tell (without plotting data) whether the packing under consideration is one of the few for which the correlation does not work.

An excellent statistical fit to data is therefore insufficient to render a packing pressure drop correlation suitable for design. In addition to a good fit to data, the correlation limitations must be fully explored. Most published packing pressure drop correlations fail miserably here: their limitations are often unknown, and if known, are seldom reported.

Interpolation of actual experimental data circumvents the systematic correlation limitations, gives reliable and accurate pressure drop prediction, is difficult to computerize, and requires that suitable interpolation charts are available. This section deals with predicting pressure drop by correlation. Section 8.2.9 describes interpolating experimental pressure drop data.

Neither correlation nor interpolation can overcome the limitations inherent in pressure drop data (Sec. 8.2.5). It is therefore essential that the user be aware of these when applying either correlation or interpolation.

The Eckert generalized pressure drop correlation (GPDC) chart For many decades, the Eckert GPDC chart has been the standard of the industry for random packing pressure drop prediction. Even though it has fallen in popularity in the last couple of decades, this method is still recommended by many recent design publications (1,14,15,17,41). It is based on work by Leva (70), who introduced constant pressure drop curves on the Sherwood flooding chart (Fig. 8.17). Leva's chart was later refined by Eckert (1,53,72,74), Prahl (85), and Strigle et al. (5,15,77). Recent charts by Eckert (74) and Strigle et al. (5,15,77) omit the flooding curve (see Sec. 8.2.6). Figure 8.19a and b shows the latest (Strigle) versions of the Eckert GPDC chart (15). These two diagrams are identical, except that in Fig. 8.196, Strigle changed the scales from log-log to semilog to make readings between adjacent pressure drop curves easier to obtain.

Both agreement (75,86) and disagreement (87,88) between measured pressure drops and those predicted by the Eckert GPDC have

Gpdc Chart

Figure 8.16 The latest version of the GPDC pressure drop correlation, (a) The latest log-log version of Eckert's correlation, as presented by Strigle (15). (Part a from Ralph F. Strigle, Jr., Random Packings and Packed Towers. Copyright © 1987 by Gulf Publishing Company, Houston, Texas. Used with permission. All rights reserved.)

Figure 8.16 The latest version of the GPDC pressure drop correlation, (a) The latest log-log version of Eckert's correlation, as presented by Strigle (15). (Part a from Ralph F. Strigle, Jr., Random Packings and Packed Towers. Copyright © 1987 by Gulf Publishing Company, Houston, Texas. Used with permission. All rights reserved.)

Figura 8.19 (Continued) The latest version of the GPDC pressure drop correlation. (6) A version identical to part a, but plotted on semilog paper, produced by Strigle (15). (Part b from Ralph F. Strigle, Jr., Random Packings and Packed Towers. Copyright © 1987 by Gulf Publishing Company, Houston, Texas. Used with permission. All rights reserved.)

Figura 8.19 (Continued) The latest version of the GPDC pressure drop correlation. (6) A version identical to part a, but plotted on semilog paper, produced by Strigle (15). (Part b from Ralph F. Strigle, Jr., Random Packings and Packed Towers. Copyright © 1987 by Gulf Publishing Company, Houston, Texas. Used with permission. All rights reserved.)

been reported. Bolles and Fair (55) used thousands of published data to show that pressure drops predicted by the Eckert correlation need to be multiplied by a safety factor of 2.2 to 2.5 for 95 percent confidence of success. MacDougall repeated this analysis, but pruned out data in the loading and high-turndown regions, that may have contributed to such a high safely factor, and looked only at Pall® rings. His analysis yielded the same safety factor as Bolles and Fair.

Strigle (15) and Kister and Gill (60) compared predictions from the latest version of the Eckert correlation (Fig. 8.19a and 6) to thousands of random packing pressure drop measurements. The Eckert correlation was shown to give good predictions for most pressure drop data (15,60). It generally works well for the air-water system for flow parameters as low as 0.01 and as high as 1 (60). For nonaqueous systems, it works well for flow parameters of 0.03 to 0.3 (typical of atmospheric distillation).

The Eckert correlation was shown (60) to be optimistic for flow parameters greater than 0.3 (typical of pressure distillation and/or high liquid rate applications). Strigle (15) attributes these optimistic predictions to enhanced liquid frothiness at higher pressure. The en-

0 0

Post a comment