The main consideration for introducing reflux or intermediate feed into a packed tower is adequately distributing the incoming stream to the packing. Unlike most tray columns, packed towers are sensitive to distribution. Mal-distribution is detrimental to packing efficiency and turndown. The main devices that set the quality of distribution in a packed column are the top (or reflux) distributor, the intermediate feed distributor, the redistributor, and sometimes the vapor distributor. Adequate hydraulics in the inlet area is also important; failure to achieve this can affect distributor performance and can also cause premature flooding.
This paper goes over common distributor and redistributor types and inlet arrangements used in packed columns, outlines the preferred practices, highlights consequences of poor practices, and supplies guidelines for troubleshooting and reviewing designs of distributors, redistributors, and feed and reflux inlets to packed towers.
Nature and Effects of Mal-distribution
A detailed discussion of packed-tower mal-distribution is far too bulky for inclusion here and is available elsewhere. Conclusions which specifically pertain to distribution equipment practices are highlighted below:
- Packing efficiency may decrease by a factor as high as 2 to 3 due to mal-distribution
- A packed column has reasonable tolerance for a uniform or smooth variation in liquid distribution and for a variation that is totally random ("small-scale mal-distribution"). However, the impact of discontinuities or zonal flow ("large-scale mal-distribution") is much more severe.
- The necessity for uniform liquid distribution sharply increases with the number of theoretical stages per packed bed. For less than five theoretical stages per bed, the column is relatively insensitive to the uniformity of liquid distribution, while with ten or more stages per bed, efficiency is extremely sensitive to liquid distribution. A corollary is that beds consisting of small packings or structured packings, which develop more theoretical stages per bed, are substantially more sensitive to mal-distribution than equal-depth beds of larger random packings.
- A packed bed appears to have a "natural distribution," which is an inherent and stable property of the packings. An initial distribution which is better than natural will rapidly degrade to it, and one that is worse will finally achieve it, but
sometimes at a very slow rate. If the rate is extremely slow, recovery from a mal-distributed pattern may not be observed in practice.
- Three factors appear to set the effect of mal-distribution on efficiency:
- Mal-distribution delivers less liquid to some areas than to others. In these areas, the liquid-to-vapor ratio is relatively low, causing a composition pinch. The pinched areas contribute little to mass transfer. Vapor leaving these areas is rich with the less volatile components, which contaminate the vapor rising from the rest of the bed. Similarly, lights-rich liquid leaving these areas contaminates the liquid descending from the rest of the bed. The pinches also create nonuniform liquid and vapor composition profiles along the cross section of the column. This is referred to as the pinching effect.
- Packing particles deflect both liquid and vapor laterally. This promotes mixing of vapor and liquid and counteracts the pinching effect in above factor. This is referred to as the lateral mixing effect.
- Liquid flow through the packing is uneven. Directly under the distributor, the column wall area is poorly irrigated (unless the distributor nozzles are directed toward the wall). In the bed, liquid tends to flow toward the wall. After some depth, the liquid flow in the wall region exceeds the average flow through the bed.
- At small tower to packing diameter ratios (<10), the effect of lateral mixing outweighs the pinching effect, and a greater degree of mal-distribution can be tolerated without a serious efficiency loss. At high ratios of column to packing diameter (>40), the lateral mixing effect becomes too small to counteract the pinching effect. This implies that the effects of mal-distribution on efficiency are most severe in large-diameter columns and with small-diameter packings.
- Either a shortage or an excess of liquid near the wall causes large-scale mal-distribution and can substantially lower packing efficiency. If the wall zone is poorly irrigated at the top of the bed, it may take several feet of packing before a reasonable amount of liquid reaches the wall region. This effect is most severe with small packings, where liquid spread toward the wall is slow. On the other hand, buildup of excessive wall flow further down in the bed is most severe with larger packings, where liquid spread toward the wall is rapid.
- In the presence of large-scale mal-distribution, packing efficiency decreases as packing height increases. This is due to the composition non-uniformity generated by pinching and to the development of wall flow. With small packings, the above may occur even in the absence of initial mal-distribution.
- Liquid mal-distribution tends to lower packing turndown. The "standard distributor" curve in Figure 1 below depicts typical variation of packing HETP (height equivalent of a theoretical plate) as a function of vapor or liquid flow rate at a constant vapor-to-liquid ratio. The two upper curves represent a progressively lower quality of initial liquid distribution. A curve similar in shape to the uppermost curve is a clear indication of poor distribution. The diagram shows that packing turndown is largely reduced with greater mal-distribution.
Figure 1 - Effect of poor distribution on HETP
- Mal-distribution tends to be a greater problem at low liquid flow rates than at high liquid flow rates.
- Vapor is easier to distribute than liquid, but vapor mal-distribution can also be troublesome. Vapor flow through packing tends to be uniform if the initial liquid and vapor distribution to the packing is uniform. A non-uniform initial vapor profile is often generated in the column vapor inlet and vapor redistribution regions, especially when inlet velocities are high. Although vapor spreads radially through the packing quite rapidly, a non-uniform profile will persist at least for some height, causing pinching similar to that described in 5 above. In a number of 15-ft-diameter absorbers, vapor mal-distribution persisted throughout a 50-ft bed; the resulting efficiency was about half that encountered during good vapor distribution. Vapor mal-distribution is most severe in large-diameter columns, shallow beds [where the ratio of bed height to column diameter is less than 0.5], and where the packing geometry resists radial spread. The latter may be particularly troublesome with those structured packings that permit substantial radial spread only parallel to their sheets. Since the orientation of structured packing sheets usually alters every 8 to 12 in, this flow non-uniformity is unlikely to persist beyond the bottom packing element. However, the disturbance this creates to the composition profile may linger for a greater vertical distance. Vapor mal-distribution may also be induced by liquid mal-distribution when vapor flows are high. Areas of high liquid holdup will impede vapor rise and will channel the vapor into the lighter-loaded regions. Since liquid tends to accumulate near the wall, vapor will tend to channel through the center.
Quantitative Definition of Liquid Irrigation Quality
Moore and Rukovena have developed a highly effective index for quantifying the quality of liquid irrigation to a packed column. This index is given by the following equation:
Dq = 0.40(100 - A) + 0.60B - 0.33(C - 7.5)
In the equation, Dq is the distribution quality rating index in percent. The higher Dq, the better the irrigation quality. Typical indexes are 10 to 70 percent for most standard commercial distributors; 75 to 90 percent for intermediate-quality distributors; and over 90 percent for high-performance distributors. Figure 2a shows efficiency improvements accomplished by improving this index in various commercial columns.
In order to determine Dq, each distributor drip point is represented by a circle. The center of the circle is located where the liquid from each drip point strikes the top of the bed. The area of each circle is proportional to the liquid flow, and the sum of all circle areas equals the tower cross-sectional area. If the liquid is evenly divided among all drip points, the area of each circle equals the tower cross-sectional area divided by the number of drip points. Terms A, B, and C in the equation are then evaluated as follows (Figure 3a).
A is the percent of the cross-sectional area at the top of the bed which is not covered by the drip point circles. This is a direct measure of the fraction of un-irrigated area at the top of the bed.
B is evaluated by selecting a continuous region at the top of the packing, occupying one-twelfth of the column cross-sectional area. This is the area in which the largest deviation from the average flow occurs. If this area is under-irrigated, B is evaluated by dividing the circle area enclosed within this region by the area of the region (i.e., by one-twelfth of the column cross-sectional area). If the area is over-irrigated, B is evaluated by dividing the area of the region by the circle area enclosed within the region. The lowest value of B anywhere in the column is used in above equation. The value thus calculated is multiplied by 100 so that it is expressed as a percent. B gives an empirical measure of large-scale mal-distribution.
C is the total area of overlap of adjacent drip point circles expressed as a percent of tower cross-sectional area.
Figure 3b to 3d are examples used by Moore and Rukovena for illustrating the application of their technique for rating distributors.
A correlation (Figure 2b) proposed by Moore and Rukovena can be used to determine the efficiency loss in a packed tower containing pall rings or Metal Intalox packing as a function of their distribution quality rating Dq and the number of stages in the bed.
Regardless of the validity and accuracy of the final correlation, the analysis proposed by Moore and Rukovena for determining irrigation quality is a valuable tool for troubleshooting distributors and examining their performance. An inspection of Figure 3b to 3d readily pinpoints regions of large-scale mal-distribution and enables visualization of irrigation troublespots. It is strongly recommended using this or a similar analysis when evaluating distributor performance.
Figure 2a - Effect of irrigation quality on packing efficiency; Case histories demonstrating efficiency enhancement with higher distribution quality rating
Figure 2b - Effect of irrigation quality on packing efficiency; Correlation of the effect of irrigation quality on packing efficiency
Figure 3a - Distribution quality rating applications, Areas considered in liquid distribution quality rating (A = cross-sectional tower area not covered by point circles, B = point circle area in 1/12 tower area, C = area of overlap of point circles)
Figure 3b - Standard quality distributor, Dq = 60 percent
Figure 3c - Standard quality distributor, Dq = 72 percent
Figure 3d - Minimum quality distributor, Dq = 84 percent
* Reproduced from Distillation Operation by Henry Kister