f 12. 1.2. EffCCT Of EiJUM6€D DOWNCOftGA VOU/ne case rruoy no l oh 3 j
INSTALLATION An olefins plant heat-pumped front-end deprcpanizer (Figure 12.2.1). Top section diameter was about twice as large as ¡bottom section. Feed was mostly vapor, but some liquid was condensed in the feed chiller.
PROBLEM The column was unstable, and both pressure drop and bottom level fluctuated periodically. The period of fluctuation was about 30 seconds. Amplitude of fluctuation significantly increased as plant rates were raised.
INVESTIGATION Flooding checks were carried out on the column, both by the operator and the designer. These showed that the upper section was at least 20 percent below flood, and the bottom section was 40 percent below flood. The pressure drop, although fluctuating, was not excessively high and did not appear to rise rapidly with an increase in plant rates.
When reflux rate was increased, the column became more stable. During the winter, when colder refrigerant was available, stability also improved. Calculations showed spray regime operation in the upper section, and froth regime operation in the lower section.
Gamma-scans showed that the trays were performing normally. A more sophisticated "Gamma-scan trace" technique was implemented to study the fluctuations. On a number of trays, the source and detector were placed just above the tray liquid level, and the amount of radiation was recorded over a period of time (Figure 12.2.2). It was established that on all plates the liquid level increased by one inch during the pressure kicks, and then quickly dropped back to normal.
THEORIES The flooding theory, although having some evidence against it, was not completely discounted. However, an alternative theory was also formulated.
A check on trays 16 to 20 clearly indicated that these operate under dumping conditions. Column loading in this section was the same as in the bottom section of the column, but the trays were much larger. One possibility was that while most of the liquid "rained" during dumping, som« could have found its way to the downcomer and accumulated until a seal was established. When this took place, the resistance to vapor flow increased, and so did the vapor velocity. The higher velocity would then blow the seal, and the process of sealing and unsealing would then repeat itself (Figure 12.2.3). This theory explained the reduction in fluctuation during higher liquid loads, because under these conditions the downcomer seal tended to stabilize.
CURES As this column operation could not be interrupted during the normal running of the plant, it was economical to cater to both theories. The cures implemented during the next scheduled plant shutdown were:
FIGURE II 2 i
Olefins HMt*Pump«d D«propaniz«r
CfiSE S"TOy nO Z
Was this article helpful?