Case Study No

INSTALLATION A vacuum specialty chemicals column, 10 foot ID, I equipped with valve trays. Melting point of the chemicals is about j 60°F lower than operating temperature. Column has 30 trays, with feed entering 10 trays from the top.

PROBLEM When the column was opened up during a shutdown, the 16 lower valve trays have been distorted downwards, with the main support beam bent and twisted into V-shape about 1 inch deep. Solidification in several downcomers was observed.

THEORY This problem is typical of a vapor-gap damage. A vapor gap was formed under the bottom tray, subjecting the bottom tray to the full hydrostatic head of the liquid above. The tray bent, allowing the liquid to drain, which moved the vapor gap to the tray above and j so on. 1

The vapor gap was probably initiated by a blocked downcomer in the bottom section, with liquid accumulating on the trays above. This could have occurred by one of two mechanisms:•

(i) The bottom downcomer blocked, with liquid accumulating above it. When reboiler heating started, the vapor formed slugged through the trays, and a vapor gap was formed.

(ii) A blockage occurred anywhere in the bottom section. If the rate of drawing vacuum is of the same order as the rate of hydrostatic head build-up on the trays, liquid accumulates on several trays without causing damage. This continues until there is enough pressure difference to draw a vapor slug through the liquid column, or until there is sufficient pressure difference downwards to cause damage to the tray with the blocked downcomer.

CURE Draw vacuum and bring column to temperature prior to feed introduction at startup. Alternatively, heat the feed as much as possible before introducing it to the column.

INSTALLATION A process water stripper, which strips heavy hydrocarbons from process water.

HISTORY A second-hand 28 inch column was used for this purpose. The column was packed with two-inch carbon steel Pall rings. At start-up, the column achieved its design separation, and just achieved its design capacity.

A number of months later, the capacity started falling off, although good separation was still achieved. When the column was opened up, several of the rings had disappeared, while others were reduced to fractions. A check of the system pH indicated that it was normally about 4, and sometimes fell to 2.

It became obvious that ceramic packing was required, and this was installed. Two inch Intalox saddles were placed in the column, and it was returned to operation. When the column was re-started, it became apparent that the column was going to fall short of its initial design capacity, probably because of the lower capacity of ceramic Intalox saddles compared to Pall rings. ' However, the drop in capacity was not great and could be tolerated. A number of months later, a further drop in capacity was observed. When the column was opened up again, there were saddle fractions of all shapes and sizes, produced by packing breakage due to turbulence.

The decision was to replace the packing with stainless steel packing. Two-inch stainless steel Pall rings were specified, but these were in short supply and the column had to be quickly returned to service. One-inch SS Lessing rings were available from .a second-hand column, but these could only fill half the column. It was decided to go ahead and use them. When the column was re-started, design capacity again was not achieved, although separation was good.

By this time, the column itself (which was constructed of carbon steel) was suffering from a multitude of problems, including corrosion, erosion and leakage. At the same time, an additional capacity increase was required. Another second-hand column, 36 inch ID, trayed, and fabricated from stainless steel was available. After modifying its internals, it was used to replace the existing packed column.

When the old column was inspected after being removed from service, another interesting detail was noticed. The column was fabricated from two sections, welded to each other across a four inch ring. Two inches of this ring projected outside, the other two projected inside. It is uncertain whether the purpose of the inside projection was to serve as a primitive wall wiper. The packing, however, was not discontinued near the ring, giving an effective column diameter four inches smaller than the apparent column diameter, with a corresponding cross section area reduction of 26 percent at that location. This ring surely did not help the column in achieving its design capacity.

Installing the trayed- coiumn did not spe-11 the end of the problems. Although capacity was about tripled, it experienced operational difficulties. Pressure drop was high, and the column bottom level changed in a cyclic manner, increasing suddenly, then dropping, over a period of 50 seconds. When the column was opened up, it was discovered that the bottom downcomer was installed backwards, causing a restriction between the bottom of the downcomer and the seal-pan wall (Figure 12.4.1). This caused liquid build-up in the downcomer and onto the tray above. Fortunately, the liquid build-up did not propagate too far up the column. When the build-up on the bottom tray was significant, the tray would dump momentarily, clearing the liquid buildup and causing a high level in the bottom sump. The cycle would then repeat itself.

After modifying the bottom downcomer, the column finally achieved its first trouble-free operation period.

Coiumn

Coiumn

Downcomer

Downcomer

\ Seal pan

\ Seal pan

Pan a. Correct b. Incorrect

Figur« 12.4.1 Incorrectly Installed Downcomer Causes Premature Flooding

INSTALLATION In a large refinery, gas-plant reboilers used a hct pumparcund stream which was used to cool the main fractionator. Beth gas plant debutanizer and the debutanizer absorber were reboiled i using circulating hot oil (Figure 12.5.1).

PROBLEM The debutanizer, which was designed to produce mainly a liquid C3-C4 stream' produced only a small vapor stream. The pumparound circulating pump also appeared defective. Whenever the hot oil flow to the gas plant debutanizer rebciler was raised, the pump discharge pressure would wildly fluctuate. Reboiler heat duty and reflux rate would become erratic. The debutanizer was run at low pressure, 100 psig (design was 200 psig). Only a small amount of vapor, but not liquid, was being produced at the reflux drum.

i j INVESTIGATION A test was carried out in which the gasoline flow to j the absorber, the debutanizer reflux rate and the hot oil flow to the | debutanizer reboiler were raised. Pressure was slowly raised by j throttling the vapor product to improve condensation. When the pressure increased to 130 psig, the hot oil flow began to fluctuate, and at 135 psig the hot oil flow completely ceased. The column appeared to have dumped. When a drain valve on the hot oil side of the reboiler was opened, gasoline came out, indicating a tube leakage.

POST-MORTEM When the debutanizer pressure reached 130 psig, the I reboiler pressure exceeded the hot oil pressure. The low-boiling j gasoline flowed into the hot oil and flashed, generating a large I volume of vapor which backed oil out of the reboiler.

fLOW SCHfMC , CMC STUOf NO. S Í6AJÍD 0* «£» i)

INSTALLATION A water-wash column equipped with two-pass sieve trays.

PROBLEM Following a plant shutdown in which the plant was modified to increase capacity, excessive entrainment was observed from the column.

INVESTIGATION Calculations were carried out using a number of entrainment correlations. The most conservative of those predicted about a third of the observed entrainment rate under the worst possible conditions.

During low-rate operation periods, the entrainment stopped. Increasing the liquid flow rate to the top of the column also reduced entrainment. Gamma-ray scans were carried out and showed nothing abnormal about the column, except that the amount of liquid flowing over one of the top outlet weirs appeared greater than the amount of liquid flowing over the other. The reflux entered the column through a perforated pipe reflux distributor.

CAUSE When the column was opened up in a subsequent shutdown, a welding rag was found inside and half way along the reflux distributor. The rag restricted liquid flow to some sections of the distributor. Either liquid maldistribution on the top tray or high velocity jets issuing from the distributor perforations upstream of the restriction could have caused the entrainment.

The rag was believed to have been left in the reflux line which was modified during the shutdown. Upon restart, the water flow carried the rag into the reflux distributor, where it remained until it was discovered. Removing the rag solved the problem.

INSTALLATION A low-temperature distillation column equipped with sieve trays. The column was piped so that the entire vapor feed stream alwav.- passed through the column. The liquid could be used as reflux to the column, or be bypassed around the trayed section of the column and join the vapor feed (the column then simply acted as a flash drum). The column was designed so that all the liquid could be fed to the top (Figure 12.7.1). Overhead product was superheated in the product heater.

PROBLEM Prior to startup, vapor entered the column while reflux flowed into the vapor feed through the bypass. To establish column action, the bypass was closed and then the reflux control valve was slowly opened. When the reflux control valve was opened, both product and heat recovery stream outlet temperatures would significantly drop. This indicated liquid entrainment, because the heat recovery stream was unable to provide sufficient heat to vaporize a significant quantity of liquid. The presence of liquid in the top product and the low temperature of the product leaving the superheater could not be tolerated because of metallurgical limitations downstream. The column was operated as a flash drum, with subsequent loss of the heavy component to the overhead stream.

ANALYSIS The problem was diagnosed to be a downcomer sealing problem. A start-up stability diagram was constructed (Figure 12.7.2), showing the range of liquid and vapor rates at which the column can be satisfactorily started (1).

The analysis was based ok mathematically modelling the downcomer as a pipe. On this basis, the vapor rate required for satisfactory start-up at a given liquid rate was calculated. It was found that at a pressure of 70 psig, vapor rates which fall below the lower dashed curve on Figure 12.7.2 were too low to stop all the liquid from dumping through the tray perforations. Liquid would not reach the downcomer, and a seal could not be established. Vapor rates above the upper dashed curve on Figure 12.7.2 were too high to permit liquid to descend the downcomer; above this curve the vapor would "blast" the liquid out of the pipe. Satisfactory startup at 70 psig could only be achieved in the area between the two dashed curves.

SOLUTION Previous start-up attempts took place in the shaded area on Figure 12.7.2 at the normal operating pressure of 70 psig. It is clear that the start-up flow rates were well outside the satisfactory start-up range.

Increasing the column pressure to 125 psig brought the range of flow rates required for satisfactory start-up closer to the stability limits (the heavy curves in Figure 12.7.2). With some plant triteming, a reduction in vapor flow rates was achieved, and this brought the operating point to within the stability limits. The column was started up at point "A".

Once the column was started up, the downcomers became sealed and the upper curve ceased to be a limit. The pressure and vapor rate were returned to their normal design values.

Product Heater

Product Heater

Heat Recovery Stream Product

Column Bypass

Vapor

Recycle to Purification Unit

Low-Tfnpratur» Olefins Column

U.OOO-i

Was this article helpful?

0 0

Post a comment