Figure 7.10 Some factors affecting sieve tray efficiency. FRI data, total reflux, Dr = 4 ft, S = 24 in, hw = 2 in, dH = 0.5 in. Both parts show a small efficiency rise with pressure. Both parts show little effect of vapor and liquid loads above about 40 percent of flood, (a) Showing efficiency reduction when fractional hole area is increased from 8 to 14 percent of the bubbling area; (b) emphasizing little effect of vapor and liquid loads, and an efficiency increase with pressure. Af = 0.14 (Both parts reprinted with permission from T. Yanagi and M. Sakaia, Ind. Eng. Chem. Proc. Des. Dev. 21, p. 712, copyright 1982, American Chemical Society.}

This means both vapor and liquid loads are raised and lowered simultaneously. Increasing vapor rate reduces efficiency, while increasing liquid rates raises efficiency. The two effects normally cancel each other, and efficiency is practically independent of load changes (assuming no excessive entrainment or weeping). Figure 7.106 shows a typical dependence of tray efficiency on vapor and liquid loads for a commercial-scale distillation column. Anderson et al. (97) show a similar dependence for several different valve trays.

Reflux ratio. Reflux ratio was stated to have a small effect on efficiency (5).

Viscosity. Efficiency increases as liquid viscosity diminishes (5,149, 150,186). As discussed earlier (Sec. 7.2.2), lower liquid viscosity usually implies higher liquid diffusivity, and therefore, lower resistance to mass transfer in the liquid phase. It was also argued that bubbles formed in high-viscosity liquids are larger and generate less interfa-cial area (186).

Relative volatility. Efficiency increases as relative volatility is lowered (5,149,150). As discussed earlier (Sec. 7.2.2), lower volatility reduces the significance of the liquid phase resistance, and therefore, raises efficiency.

Surface tension. There is uncertainty regarding the effect of surface tension on tray efficiency. Often, it is difficult to divorce the surface tension effects from those of other physical properties. For this reason, it is difficult to tell whether the effects described below are real or imaginative.

Measurements by Fane and Sawistowski (116,117) show little effect of surface tension on efficiency in the froth regime, and a rise of tray efficiency with lower surface tension in the spray regime. The surface tension gradient (Sec. 6.4,4) appears to have an effect. A positive gradient (surface tension increases down the column) enhances efficiency in the froth regime (108,116,117,146,189,190) while a negative gradient enhances efficiency in the spray regime (108,116,117). The magnitude of these enhancements is uncertain.

Surface tension effects have frequently been used to explain observed composition effects (184,187,189) or discrepancies between theory and experiment (146). Zuiderweg (146) and Dribika and Biddulph (189) argue that the Marangoni effect (Sec. 6.4.4) stabilizes the froth and therefore enhances efficiency. The enhancement is related to

The difference y* - x represents the mass transfer driving force, while dcr/dx represents the change in surface tension with concentration. Both Zuiderweg and Dribika and Biddulph show that with large values of M (10 to 20 dynes/cm), point efficiency can be enhanced by as much as a factor of 1.5 to 2.0. However, these enhancement factors are the ratios of measured to predicted point efficiencies, and the predictions are based on the notoriously unreliable theoretical correlations (Sec. 7.2.1). Lockett (12) points out that the above enhancement factors remain high even at the froth-spray transition, and this is inconsistent with the argument. Further, experiment by Ellis and Legg <214) have demonstrated no significant effects of surface tension gradients on efficiency.

For low-viscosity absorption systems, one set of data (186) shows that in the froth regime tray efficiency increases as surface tension is reduced, while for high-viscosity absorption systems, surface tension had little effect on mass transfer (186).

Pressure. Tray efficiency slightly increases with pressure in the froth regime (17,105,119). The apparent pressure effect could be a reflection of the rise in efficiency with a reduction in liquid viscosity and in relative volatility. (Note: As distillation pressure rises, so does the equilibrium temperature; this in turn leads to a decrease in liquid viscosity.)

At pressures higher than 150 to 300 psia, and especially at high liquid rates, vapor entrainment in the downcomer liquid becomes important, and may lead to a reduction in tray efficiency with further raises in pressure (105).

Liquid and vapor entrainment. Both represent a recycling of lower-purity material which contaminates the tray liquid or vapor, both counteract the mass transfer process and lower efficiency. Liquid and vapor entrainment are discussed in Sees. 6.2.11 and 6.4.5, respectively.

Weeping. This represents liquid short-circuiting the stage and contaminating the tray below with more volatile material. Further discussion is in Sec. 6.2.12.

Special considerations In multipass trays. In a multipass tray, vapor distribution between the passes is largely determined by the hole area, while liquid distribution is largely a function of the weir height and length. If the geometry of the passes is perfectly identical, the distribution of vapor and liquid is the same for each pass, and tray efficiency is uniform. This is readily achievable in two-pass trays, where the design of each pass is identical to the other, but not so when a larger number of passes is involved. For instance, in a four-pass tray, weir length of the center passes differs from that of the side passes. Unless allowed for in the design, the LfV ratio will vary from pass to pass, with a resulting reduction in tray efficiency, as demonstrated by Bolles (191).

Bolles (191) correlated the reduction in efficiency in terms of the distribution ratio, i.e., the maximum-pass LfV ratio divided by the minimum-pass LIV ratio. The L and V for each pass are determined from the normal pressure balance and hydraulic relationships, applied to each pass. At high distribution ratios, a substantial drop in tray efficiency occurs. Bolles shows that if this distribution ratio is kept lower than 1.2, the loss in efficiency due to maldistribution is negligible. Bolles recommends designing multipass trays for such low distribution ratios. Detailed guidelines for achieving low distribution ratios (<1.2), thus minimizing the effects of pass maldistribution on efficiency, are contained in a companion book (1) and in Bolles's paper

Was this article helpful?

## Post a comment