Andrew W. Sloley, Scott W. GoldenGlitsch Inc. Dallas
Simple analysis methods and field data are the S key to pinpointing causes of and solutions to distillation equipment operating problems.
The first step in troubleshooting common field problems is to gather accurate field data and thoroughly analyze the existing operation before attempting to formulate solutions.
The following field troubleshooting case history shows how to use field data and a systematic analysis procedure to identify and remove operating limits in a deisobutanizer (DIB).
PROBLEM SUMMARY
Fig. 1 shows the overall arrangement of trayed DIB column C1. Column C1 is an 8-ft. diameter tower with 61 identical two-pass sieve trays. Tray spacing is 24 in., except at the feed and product draw trays.
Alkylation reactor effluent (main feed, called S1) enters the tower on Tray 50. A butanes stream (S2) feeds on Tray 32 and debutanizer overhead recycle (S3) feeds on Tray 2. Products include a liquid overhead to the depropanizer (S4), a side-stream draw from Tray 55 (main product, S5), and the DIB bottoms (S6).
The refinery's operating objective for (fl was 5 vol % or less isobutane in the bottoms. In actual operation, the bottoms product contained 17.9 vol % isobutane. This increased the recycle stream volume and reduced unit capacity to 23,800 b/d.
SYSTEMATIC APPROACH
Successful troubleshooting requires a systematic approach. An experience-proven method for finding the required solution includes these steps:
- Define the process objective.
- Define the current operation by data gathering and field observation.
- Analyze the test data to determine unit performance.
- Check the equipment capabilities for the current operation.
- Determine the process and equipment modifications required.
- Implement the solution.
DEFINING OBJECTIVE
This first step began with specifying an isobutane content of 5% or less in the bottoms product. Additional objectives included increasing unit throughput to an alkylate feed rate equivalent of 25,000 b/d and restricting process and equipment modifications to those that could be made on C1 with minimal effects on other equipment.
DATA GATHERING
The second step - definition of current operation by gathering data - usually requires two stages:
1. Talk to the operators, engineers, and technicians who routinely work on the unit. This is critical. Without the observations of people who work on the unit every day, troubleshooting attempts will fail before they start.
2. Gather a complete, accurate set of test data that has both heat and material balance closure. Without an accurate "snapshot" of unit operation, any conclusions drawn are lacking a firm foundation. The usual result of inadequate data is an argument of opinion, incomplete solutions, and a general waste of time, resources, and effort.
Data gathering for this analysis consisted of a field test of unit performance and informal conversations with unit operators. The field test included measurement of pressure and temperature profiles and stream composition data.
Fig. 2 shows a schematic of C1 with the measured data shown.
Conversations with the operators were aimed at understanding the history and operating characteristics of the DIB column and drawing out the operators' opinions on the column problem.
DATA ANALYSIS
Analysis of the test data starts with a review for internal consistency and heat and material balance closure.
Based on these data, the material balance error, on a component basis, was 3.6% compared to an ideal set of test data with 0% material balance error. For industrial systems of this type, however, material balance errors of less than 3% are considered good.
The test data error in this case was slightly more than the preferred 3% limit, but was judged close enough to yield an acceptable analysis' The main feed (S1) rate and composition were adjusted to close the component material balance.
The measured pressure drop was 5.9 psi. A normal pressure drop for trays in this service would be approximately 4 psi (calculations show a 4.1-psi pressure drop should be expected for the test loading rates).
The elevated pressure drop indicates incipient flooding. The flooding restricted reflux rates, which led to low isobutane recovery.
For evaluation of hydraulic loads inside C1, a simulation was setup and run with a commercially available simulation package. The simulation also provided a review basis for the fundamental process design and a check on recycle stream disposition. The simulation showed that, at test run conditions, C1 achieved the expected separation.
Fig. 3 shows the separation parameter profile for the isobutane vs. normal butane split. The flat portion in the center of the separation profile shows that a significant section of the tower provides no fractionation between isobutane and normal butane.
EQUIPMENT CHECKS
A number of equipment checks uncovered a collection of small but significant problems.
TRAY DECK
Hydraulic problems inside trayed towers usually arise from design faults in tray decks and auxiliary internals, which include feed distributors, draw sumps, and baffles.
Hydraulic analysis of the tray decks cut C1 into the following regions:
- Above the side draw (Trays 61-55)
- Between the side draw and the main feed (Trays 5450)
- Between the main feed and the secondary feed (Trays 49-33)
- Below the secondary feed (Trays 32-1).
Table 1 summarizes the overall tray hydraulic analysis for the test operation.
The normal design limits for the tray hydraulics in this two-pass tray system are:
- Jet flood, 66-80%
- Downcomer choke flood, 80% or less
- Weir load, minimum 0.25 gpm/in. of weir.
Table 2 summarizes the expected hydraulics for a 25,000 b/d feed case at the target operating conditions. The most stringent hydraulic limit for the test case was the downcomer flooding in the trays below the secondary feed.
The recommended value is 80% for this system, which preserves an operating and uncertainty margin. This margin helps protect against failures caused by data uncertainty, operation upsets, and correlation uncertainty.
In this section, however, even the calculated downcomer flood of 79% was below the 80% limit. And downcomer flooding in the bottom section decreased to 78% for operation at the increased feed rate and target operating conditions. This is because the rate of recycle stream S3 is reduced when running C1 at 5% isobutane bottoms, as compared to 17.9%.
The low liquid loadings on Trays 33-54 led to "blowing." Blowing occurs when the amount of liquid thrown into the downcomer by the agitating vapor is large compared to the amount of liquid entering the downcomer by cross-flow across the tray. There is so little liquid on a tray that the agitation of the vapor blows all of the liquid into droplets.
The nominal minimum liquid loading in this system to avoid blowing is 0.25 gpm/in. of weir length. The actual value of 0.3 gpm/in. only slightly exceeds this minimum. Analysis of the data showed very poor efficiency for Trays 33-54 (less than 20%). But while blowing reduces tray efficiency, normally it does not cause flooding.
Overall, this analysis showed that the trays installed in the tower should meet the target operating rates and specifications.
FEED INLET
While no major tray deck problems were seen in the feed inlet design, a number of minor problems were found, any one of which could cause the C1 column to flood prematurely. The first among these was the S2-stream feed inlet.
An external butanes stream, S2, could be fed to three feed points for optimized isobutane recovery. The feed points are on Trays 32, 24, and 16. While external provision for three feeds was included, only one feed distributor was installed - on Tray 16.
The actual feed point used was on Tray 32.
This led to localized imbalancing on Tray 32 and the potential for flooding the tower.
REBOILER RETURN
The reboiler return was poorly designed. Because the return nozzles were located just 3 in. above the liquid level in the bottom of the tower, some of the aerated liquid was entrained up to the bottom tray. This entrainment resulted in loss of fractionation efficiency and possible premature flooding.
Column C1 has a thermo-syphon reboiler. The driving force for flow in the reboiler comes from the liquid head maintained in C1. Rigorous flow and exchanger analysis showed that reducing the weir height inside C1 would result in unacceptable reboiler duty reduction because of the reduced liquid head available. Therefore, no changes, were made to the reboiler return.
ANTIJUMP BAFFLES
The inboard downcomers lacked antijump baffles. Operation at high rates requires the installation of antijump baffles along the centerlines of inboard downcomers.
Observation of the test systems showed that expansion of the vapor at the outlet weir would pump the liquid over the weir. At a sufficiently high vapor rate, the trajectory carries the liquid completely over the downcomer and onto the opposite side of the tray. The tray then floods prematurely because of increased liquid holdup, caused by cycling the liquid across one side of the tray and back to the other.
Antijump baffles deflect liquid from jumping into the downcomer. For this system, antijump baffles are recommended if the ratio of operating vapor load to active area (OVL/AA) exceeds 80% of the ratio of limiting vapor load to active area (LVL/AA) (see equations).
For C1, LVL/AA is 0.2206 in the lower section of the tower and the OVL/AA is 0. 1879. This shows that C1 is operating at 85.2% of the allowable limit - in excess of the 80% maximum design value. Liquid jumping at the inboard downcomers is therefore a probable cause of tower flooding.
MINOR PROCESS CHANGES
The target for process modifications was to examine possible changes while maintaining the tower internal flood values at currently achievable levels. This was done by performing calculations to constant liquid and vapor limits in the deisobutanizer.
The most stringent conditions were:
- Minimum liquid rate constraint in the section between the side draw and the reactor effluent feed point 0 Maximum vapor rate constraint in the bottom section of C1 (between the S2 and S3 feed points).
EFFLUENT FEED POINT
Changing the feed point of the main feed, S1, would reduce the number of trays in rectification service (top section) and increase the number in stripping service (bottom section). The separation parameter profile (Fig. 3) permits a fundamental analysis of the feed point change.
Both the upper and lower sections of the tower are indicated by a "flat" section in the profile. Changing the feed point only moves trays from one flat section to the other.
The overall isobutane recovery does not change.
Simulations confirmed this conclusion.
S2 DISPOSITION
When reviewing the compositions of the feed and bottoms streams, two process anomalies became apparent. First, the low-concentration S2 stream (approximately 11% isobutane) entered C1 above the medium-concentration (approximately 33% isobutane) recycle stream, S3.
Second, the S2 stream had a lower isobutane content than the tower bottoms (about 18%). This situation implied that, for the test operation, sending the S2 stream to C1 resulted in a downgrading of isobutane in the bottoms. A quick process check confirmed this.
This analysis is not valid, however, when C1 runs at the target operation of 5% isobutane in the bottoms.
FEED PREHEAT
Column C1 receives a subcooled feed. This subcooling condenses rising vapors in the tower and builds up a recycle loop in the bottom of the tower. Feed preheat reduces the buildup of internal loads in C1 by moving the enthalpy of the feed stream closer to the enthalpy of the feed tray.
The exact degree of feed preheat desired depends on feed flow rates, feed composition, and the economic trade off between side-stream purity and isobutane loss. For the target operation basis, the optimum feed preheat temperature is about 117 F. This reduces isobutane losses to the bottoms by 78%.
CHANGES RECOMMENDED
The immediate change recommended was to send stream S2 to the bottoms directly instead of to C1. This temporary change gained an immediate isobutane recovery improvement. The second, permanent change was to feed stream S2 to Tray 16 instead of to Tray 32, after the mechanical problems inside C1 were resolved.
Long-term changes implemented at the following shutdown included installation of:
- Antijump baffles for all inboard downcomer trays between Trays 1 and 32 0 "Picket fence- weirs on Trays 54-50
- A feed preheat exchanger.
After start-up following these changes, operation at less than 5% isobutane loss and less than 5% isobutane in the bottoms, and at 25,000 b/d of feed charge was easily achieved.
The total effect of the modifications was to increase C1 capacity (at specification operation) from 23,800 b/d to 27,400 b/d. While not every detail was corrected (the interior weir in the boot was left unmodified), enough capacity was obtained so that the column's capacity has not limited the overall refinery operation since the modifications were made.
OTHER PROBLEMS
A review of the process and mechanical design of C1 showed several minor problems that had to be solved before the tower could be operated close to its upper capacity limit. These problems arose from two major points:
- Lack of attention to detail on minor points (antijump baffles, feed nozzles, etc.)
- Use of one tray design across an excessive range of liquid rates (the high-to-low-rate ratio was 22:1, as compared to a good design value of 3:1 for sieve trays).
A point meriting special attention is that every change proposed for C1 was checked by evaluating the impact on other parts of the unit. This included taking into account the changes in the S3 recycle stream composition as C1 performance changed, and the impact on the thermosyphon reboiler of changing the weir height in C1's boot.
CUTTING CORNERS
The problems found in this example are classified as "minor" errors. Minor refers only to the savings achieved by cutting comers during initial design. The actual operating effect was a sustained performance downgrade over a 25-year period prior to installing the needed modifications.
Conservative engineering practices and thorough checking would have avoided all the problems found.
BIBLIOGRAPHY
Ballast Tray Design Manual, Bulletin No. 4900, 5th edition, Glitsch Inc.
Keffer, Pamela S., "Sulfuric Acid Alkylation Process Design," Stratco Inc., October 1988.
Kitterman, Layton, and Ross, Mike, "Tray Guides to Avoid Tower Problems," Hydrocarbon Processing, May 1967.
Kitterman, Layton, "Tower Internals and Accessories, " First Congresso Brasileiro de Petro-Quimica, November 19,6.
Copyright 1993 Oil & Gas Journal. All Rights Reserved.
