Liquid distribution, packing

Thom and Byers (35) discussed the design and construction issues that can lead to the inaccurate model predictions compared to field observations in that liquid distribution, packing materials, fouling, chemical reaction, and end effects were of great importance. 

E] Z = packed height, m of each section with its own liquid distribution. The original work is reported in English units. Cornell et al. (Ref. 81) review early literature. Improved fit of Cornell s ( ) values given by Belles and Fair (Refs. 74 and 75) and in Fig. 5-29. 

Bed limiters commonly are used with metal or plastic tower packings. The primary function of these devices is to prevent expansion of the packed bed, as well as to maintain the bed top surface level. In large diameter columns, the packed bed will not fluidize over the entire surface. Vapor surges fluidize random spots on the top of the bed so that after return to normal operation the bed top surface is quite irregular. Thus the liquid distribution can be effected by such an occurrence. 

Packed bed heights typically vary from 20 ft to 30ft. Many times the location of manholes to provide access to the redistributor will determine the packed depth. Whenever more than 15 theoretical stages are required in one packed bed, good liquid distribution is critical. ... 

Liquid distribution in a packed bed is a function of the internal vapoi/liquid traffic, the type of packing employed, and the quality of the liquid distributors mounted above the packed bed. Vapor distribution is controlled by the internal vapor/liquid traffic, by the type of packing employed, and by the quality of the vapor distributors located below the packed beds. 

Liquid distribution probablv phtys the most important part in the efficient o]veration of a packed tower. good packing from the process viewpoint can be rednced in effectiv eness bv poor liquid distribution across the to ) of its upper surface or the packing sections below mv feed inlet(s) oi rellitx inlets. 

Figure 9-8J. MTS-109 Multipan two-stage liquid distributor for optimum liquid distribution for uniform flow for random and structured packings for low to moderate liquid rates, less than 5 gpm/ft. Also used for redistributor. Used by permission of Nutter Engineering, Harsco Corp., Bull. TI-1, under license from The Dow Chemical Co., protected by U.S. Patents No. 4,472,325 4,808,350 5,013,491.

Ov Crall, the careful design of a distributor for liquid in the top of a packed tower, and for the redistribution of liquid flowing dow n multi-section packing in the tower, is essential to good consistent tower performance. However, the liquid flow is not alone, the uniformity of vapor distribution is likewise essential, because non-uniform vapor distribution can cause non-uniform liquid downflow. Then, there is the selection of the packing itself and its characteristics and requirements/sensitivity to the uniform distribution of the liquid and vapor. As earlier emphasized, the level of the distributor tray or trough can be critical to the consistent uniform liquid distribution. 

Liquid Distribution Patterns in Packed Colunms (Data in 6-in., 12-in., and 24-in. Dia. Towers)... 

The type of distribution to select depends on the sensitivity of the tower performance to the liquid distribution as discussed earlier. Norton s [83] data indicate that the sensitivity of tower performance to liquid distribution quality depends only on the number of theoretical stages in each bed of packing achierable at its System Base HETP [83]. Tower beds of high efficiency packing are more sensitive to liquid distribution quality than shorter beds of medium efficiency packing [83]. It is important to extend the uniformity of the distributor all the way to within one packing particle diameter of the tower wall [85]. 

Perry et al. [85] point out that packed columns are more dependent on liquid distribution than trayed columns, as can be appreciated by the differences in the way the liquid must flow down the two types of columns. Liquid distribution quality is measured or described as [85] ... 

Hoek [86] proposed a radial spreading coefficient to characterize the liquid distribution. This coefficient is a measure of how quickly a packing can spread a vertical liquid stream radially as the liquid progresses down the column [86]. Radial mixing tends to reduce the effects of... 

Structured packing requires specially designed distributors recommended by the respective manufacturers to ensure the same important uniform liquid distribution across any bed of this type of packing. 

Stacked packing is a hand operation and rather costly. It is avoided where possible except for the initial layers on supports. Liquid distributed on a stacked packing usually flows straight down through the packing immediately adjacent to the point of contact. There is very little horizontal liquid flow. Packing patterns perform differently, and are illustrated in Figure 9-16A-C. 

Kunesh [126] presents tm overview of the basis for selecting rsuidom packing for a column application. In first deciding between a trayed tower or a packed one, a comparative performance design and its mechanical interpretation should be completed, considering pressure drop, capacity limitations, performance efficiencies (HETP), material/heat balances for each alternate. For one example relating to differences in liquid distribution performance, see Reference 126. 

Importance of miiform liquid distribution across the packing... 

For good and uniform performance of any structured packing it is essential to have imiform, consistent vapor and liquid distribution therefore, much care must be given to the design details. See earlier discussion in this chapter. 

For 1-in. Berl saddles allow 2.0 ft for good liquid distribution through the packing from the top. 

If PaU rings had been used, only tw O packed sections would be considered, because the general liquid distribution pattern is better. This would require a reevaluation of the performance, and a probable reduction in total packed height. 

In Reference 14, the authors modified the equations for Hg and Hj as follows (a) eliminate column diameter correction above 24 in. and (b) columns with good liquid distribution probably can allows elimination of the packing height correction. 

Satisfactory operation must be between the upper and lower limits for both liquid and vapor flow rates. At liquid rates below 0.5 GPM per square foot of packing cross-section, liquid distribution is not uniform enough to ensure thorough wetting. At liquid rates between 25 GPM and 70 GPM per square foot of packing, the column is considered liquid-loaded and becomes very sensitive to additional liquid or vapor flow. 

Zabor et al. (Zl) have described studies of the catalytic hydration of propylene under such conditions (temperature 279°C, pressure 3675 psig) that both liquid and vapor phases are present in the packed catalyst bed. Conversions are reported for cocurrent upflow and cocurrent downflow, it being assumed in that paper that the former mode corresponds to bubble flow and the latter to trickle-flow conditions. Trickle flow resulted in the higher conversions, and conversion was influenced by changes in bed height (for unchanged space velocity), in contrast to the case for bubble-flow operation. The differences are assumed to be effects of mass transfer or liquid distribution. 

Column Operation To assure intimate contact between the counterflowing interstitial streams, the volume fraction of liquid in the foam should be kept below about 10 percent—and the lower the better. Also, rather uniform bubble sizes are desirable. The foam bubbles will thus pack together as blunted polyhedra rather than as spheres, and the suction in the capillaries (Plateau borders) so formed will promote good liquid distribution and contact. To allow for this desirable deviation from sphericity, S = 6.3/d in the equations for enriching, stripping, and combined column operation [Lemlich, Chem. Eng., 75(27), 95 (1968) 76(6), 5 (1969)]. Diameter d still refers to the sphere. 

Fig. 5.2.3 Identification of rivulets and surface wetting in a packing of 5-mm diameter glass spheres contained within a column of inner diameter 40 mm. The data were acquired in a 3D array with an isotropic voxel resolution of 328 xm x 328 pm x 328 [im. (a) The original image of trickle flow is first binary gated, so that only the liquid distribution within the image is seen (white) gas-filled pixels and pixels containing glass spheres show up as zero intensity (black), (b) The liquid distribu-...

---