For sieve trays, m - 1.94, Cw = 0.79. Note that the constants (69) are a slight revision of those presented in the original paper (63). Clear liquid height hc is calculated from Colwell's correlation (Sec. 6.3.5). The Hsieh and McNulty correlation applies to trays with 9 percent and larger hole area. For trays with smaller hole area, Hsieh and McNulty (63) expect the weeping rate to be smaller than predicted.

Weep traction. fw is the ratio of weep rate from the tray to the total liquid flow rate entering the tray, i.e., fw = W/GPM (6.40)

Above the weep point, W = 0 and fw = 0. At the dump point (Sec. 6.2.14), all liquid weeps from the tray, and fw = 1.0. It was stated (50,70) that when weeping across the tray is uniform, the decrease in tray efficiency is usually considered acceptable with a weep fraction of up to 0.1.

6.2.13 Valve tray weeping

Bolles (71) extended Fair's sieve tray weep point correlation (31; Fig. 6.18) to investigate weeping in valve trays. Some results are depicted in Fig. 6.19. The axes of Fig. 6.19 are identical to those of Fig. 6.18. Each dashed line is the locus of weep points predicted from Bolles' extended Fair correlation. The heavy lines are the "operating lines," i.e..

Flgur* 6.19 Weep point pressure balance for sieve and valve trays, (a) Sieve tray; (6) well-designed valve tray; (c) valve tray with too many valves or with valves that are too light; (d) valve tray with too many valves, but fewer than in c; («) well-designed valve tray with two valve weights. (From. W. L. Boliea, Chem. Eng. Prog., 72 (9), p. 43 (September 1976), reprinted courtesy of the American Institute of Chemical Engineers.) 305

the relationship between the upward force [left-hand side of Eq. (6.31c)] and downward force [right-hand side of Eq. (6.31c)] as column loadings increase. When the heavy line is below the dashed line, the force up is lower than the force up at the weep point, and weeping occurs. When the heavy line is above the dashed line, no weeping occurs. The point of intersection of the heavy and dashed line is the weep point.

In a well-designed valve tray (Fig. 6.196), the weep point is below the vapor load at which the valves open. Throughout the valve-opening process (flat portion of the heavy curve in Fig. 6.196), the operating curve remains above the weep point. The weep point of a sieve tray (Fig. 6.19a) is shown for comparison. Figure 6.19c shows a tray with too many valves, or one with valves that are too light. Here the weep point is high, and the turndown may be even worse than that of a sieve tray. Figure 6.19d shows a tray with fewer valves (or heavier valves) than those in Fig. 6.19c. Here the tray will stop weeping as the vapor load rises, but will resume weeping as valves open up. Turndown again will be poor, but performance will be better than the valve in Fig. 6.19c. Figure 6.19e illustrates a well-designed valve tray with two valve weights. The diagram illustrates how this design overcomes the problem depicted in Fig. 6.19c?.

Several experiences of severe weeping from valve trays have been reported (1,71,75). A well-designed valve tray is unlikely to have too many valves, but trays with light valves are common in an effort to reduce pressure drop. To avoid the turndown problems, manufacturers often specify a valve tray with two valve weights (Fig. 6.19e). When the light valves open, the heavy ones are still shut, which reduces the ratio of slot to active area and avoids weeping. This practice is discussed in detail elsewhere (1,71).

Weep mechanism. Banik (72) made some observations on the mode of weeping from valve trays. At high vapor loads (low weep rates), weeping takes place through the valve openings, is induced by the horizontal liquid velocity, and the weeping liquid descends at an angle. At low vapor loads (high weep rates), liquid weeps through the opening between the valve legs and the hole edge and through the peripheral openings of the valve.

Tests by Banik (72) and Zhang et al. (70) show that weeping from valve trays is nonuniform. In Banik's 4 ft x 2 ft rectangular simulator, most of the weep issued from the inlet half of the tray at low liquid rates (< 3 gpm/in of outlet weir) and from the outlet half of the tray at high liquid rates (>10 gpm/in of outlet weir). The nonuniformity appeared to escalate as weir height increased. This pattern of non-uniformity is similar to that observed by Banik and Lockett (56j on sieve trays. In Zhang et al.'s (70) 5 ft x 1 ft rectangular simulator.

weeping was uniform before valves started to open. Opening of the valves began from the outlet end of the tray, and was accompanied by weeping mostly from the inlet half of the tray.

Unlike sieve trays, valve trays were observed to experience substantial weeping from the inlet row of valves (72). An inlet weir about 1 in tall was shown (72) to roughly half this inlet weep at higher liquid rates (>5 gpm/in) but to be less effective at lower liquid rates. The installation of such an inlet weir ("interrupter" or "breaker" bar) is a common design practice on valve trays.

Factors affecting weep. The mechanical design of the valves has a large impact on the weeping tendency. This is discussed elsewhere (1). In addition, most of the factors that increase weeping tendency in sieve trays also increase weeping tendency in valve trays. These include

■ Increasing valve slot area (1,71)

■ Increasing liquid rate (63,65,71,72) » Increasing weir height (71,72)

■ Contouring valve orifices

Weep prediction. The weep point of valve trays can be calculated from the Bolles extension (71) of Fair's weep point correlation (31). The same correlation (Fig. 6,18) is used, except that the sieve fractional hole area is substituted by the ratio of valve slot area to tray active area. An alternative weep point correlation for valve trays was presented by Klein (73). Hsieh and McNulty (63) extended their sieve tray weep rate correlation (Sec. 6.2.12) to valve trays. The extension is complex, and discussed elsewhere (63).

Vapor channeling. The turndown of valve trays may be restricted by channeling (poor vapor distribution) or by pulsation at low vapor rates rather than by excessive weeping. Zhang's simulator study, and motion pictures discussed by Lieberman (74), suggest that at low throughputs, valves at the low-liquid level regions (near the outlet weirs) open first, while those near the tray inlet remain shut, causing vapor channeling at the tray outlet. On the other hand, tests by Zuiderweg et al. (75) and a case study by Reay (1), both in large-diameter (>13 ft) columns at relatively low liquid rates, suggests that vapor channels through a small aerated zone located in some intermediate position on the tray, with the remaining valves completely shut. On two-pass trays, one panel sometimes tends to be active and the other panel inactive (1). In any of the above cases, vapor channeling induces nonuniform weeping, a reduction in tray efficiency, and possibly other adverse effects (e.g., Ref. 1).

To minimize vapor channeling, valve trays are designed to exceed a minimum unit reference (50). A unit reference is the ratio of the vapor rate to the vapor rate at which all the valves are open (Sec. 6.3.2). A minimum unit reference of -iO, 60, and 80 percent is recommended for one-, two-, and four-pass trays, respectively (50). If the unit reference falls below the minimum, selected valves can be blanked, valve density can be reduced, or the ratio of light to heavy valves can be varied (7,50).

6.2.14 Dumping

As vapor rate is lowered below the weep point, the fraction of liquid falling through the holes increases until a condition is reached where all liquid fed onto a tray weeps through the holes and none reaches the downcomer. This condition is referred to as dump point or seal point.

Below the dump point (100 percent weep), tray efficiency is a small fraction of its normal value, and mass transfer is extremely poor. Since no liquid enters the downcomers, they lose the liquid seal that prevents vapor from rising through them. Operation below the dump point can be accompanied by severe hydraulic instability due to unsealing of downcomers, as was demonstrated by field experience (76). The startup stability diagram (1), which defines the range of vapor and liquid rates needed for satisfactory startup, has the dump point as the lower limit. The tendency of dumping increases when (77-79)

■ Liquid rate decreases (Note: This is the reverse of the effect of liquid rate on the weeping tendency. The effect of liquid rate on weeping and dumping is depicted in Fig. 6.6.)

■ Fractional hole area increases.

■ Weir height increases.

The most extensive studies on dumping were reported by Prince and Chan (77-79). The Chan and Prince dump-point correlation (Fig. 6.20« was recommended by Chase (30). The author has also had favorable experience with the correlation under conditions widely different from those used in its derivation. Alternatively, the dump point can be predicted from a weep rate correlation by setting the weep rate equal to 100 percent of the liquid entering the tray. However, little has been reported by either Lockett and Banik (56) or Hsieh and McNulty <63' about the reliability of dump-point predictions from their correlation.

Was this article helpful?

0 0

Post a comment