This Month’s Puzzler
Originally, our column was installed with a large feed preheater. Severe corrosion in that preheater and an increase in throughput required a larger reboiler. Our internal reboiler was replaced by an external vertical thermosiphon using an available bottom nozzle. Now, 30% of the tubes in the preheater are plugged. The trouble is that the nozzle is too small for the thermosiphon.
Our control engineer at the time — he has since retired — set the tower level control as the bottom valve. This worked well enough. A new production supervisor didn’t like that arrangement because it was a challenge starting up the plant. He had a contractor create a cascade control scheme where the level controlled the steam flow to the reboiler while the column bottom temperature served as a subordinate controller for the bottom valve.
With the changes, the reboiler stopped working completely. We can hear steam getting to the reboiler but the thermosiphon is working only erratically if at all. Our plant outage has been extended while we unravel this mystery.
A committee of engineers has come up with a range of possible causes: 1) the preheater isn’t providing sufficient heat because additional tubes were plugged during the outage; 2) the level reading is faulty because there wasn’t time to clean the tank bottoms and fouling has occurred in the bottom pressure leg; and 3) the steam traps aren’t doing their job, causing surges of condensate causing control problems.
The corporate engineer suggested contacting the old engineer for some advice. What do you think caused the problem? What is the best way to address it?
Check Control Stability
Although the more-common approach for bottoms control is tray temperature controlling reboiler heat input (either directly or by cascading temperature as primary and heat input flow as secondary), other approaches also are applied depending on the systems. In a very broad sense, control stability could be affected by the loop and tuning; the process itself (distillation, steam flow, feed composition variations, etc.); and other control loops on the column (e.g., pressure control or reflux control). Look into the following possibilities:
1. Signal noise in the level transmitter could cause variations in steam flow to the reboiler. This not only will affect the reboiler but also vapor/liquid traffic through the column and column stability. If possible, locate level transmitters in a stilling well. If this is not possible, consider increasing the damping setting on the transmitter.
2. Steam flow transmitter problems also could affect bottoms control stability. It’s not clear as to which technology is used for steam flow.
If it is a dp cell without a capillary arrangement, make sure the high-pressure and low-pressure legs are filled equally with a suitable liquid seal such as condensate. A capillary arrangement does not have this problem.
Condensate in the steam supply line will make the steam flow indication erratic, which, in turn, affects control stability. Ensure the steam header to the reboiler is insulated and provided with steam traps to drain condensate. In addition, any closed or failed-in-close-position steam traps on the condensate flow out of the reboiler will stop the reboiler altogether. Temporarily, you can bypass the steam trap — then, if the reboiler starts working, the steam trap is a cause of the problem.
If you use other technologies (e.g., mag meter, turbine meter, etc.) look into their specific problems that could affect stability.
3. For cascade tuning, the secondary control (steam flow) typically should be two-to-three times faster than the primary control.
4. It is not clear whether the thermosiphon is vertical or horizontal. For vertical thermosiphons, if the bottoms level is too low, heat transfer will cause the loss of thermosiphon due to film boiling in the tubes. For vertical thermosiphons, the level should be 70%–75% of the tube length. For horizontal thermosiphons, a bottoms level that is too low may not be adequate to support liquid height to the baffle in the reboiler — this also will make the heat transfer unstable. In either case, keep the set point of the bottoms level at an adequate level to ensure sufficient liquid height in the reboiler.
5. Minimize impact of other loops. For example, ensure you have a reasonably constant feed flow, temperature and composition. Similarly, column pressure control should be relatively stable.
With the original control scheme, you said that the supervisor had startup problems but didn’t specify the type of startup problems. If the column is heat-integrated with other columns, try to decouple the columns during startup. Review the startup procedure to make sure it addresses the potential stability problems during startup. Discuss with the supervisor, contractor and operators the reasoning and experience that prompted dropping the old control scheme.
GC Shah, consultant
Houston
Set Up Three Groups
The column apparently started up successfully and the failure occurred later. Because it is rare that a single distillation column acts alone, your team should consider the usual suspects first: column compositions of all columns. A second group should assess the controls. A third should take the global view and look at the overall supplies to the plant or refinery.
Let’s outline the tasks for the first group.
All good investigations begin with sampling. Then, run simulations to determine if there is any species that could cause fouling at the bottoms. Comparisons help, so look at the old model for the system; make sure it is accurate — never assume it’s complete without a field review. Don’t get stuck relying on one software package: even hand calculations could reveal something.
Clearly, you can’t limit the boundary of the first study group to the first column. Also, besides looking at the performance of each column, review the selection of the feed preheater, condenser and reboiler. The condenser, while not showing up as a problem, still could tell you something; so sample it, too, and review the design package with the others.
It is strange that the reboiler and preheater are “over-sized.” This should ring alarm bells because heat exchangers are designed for a specific velocity range. If solids could collect in a reboiler, low velocity won’t keep the tubes clear. Review this with the manufacturer and designer.
The design-review-team effort can begin while your control team looks at the old trend lines. This should involve comparing the startup, shutdown and steady-state operation of the troubled column and the system in general. Changes done during the outage could affect more than one column; so, other changes may deserve a look.
The third group, when arranging for inventory diversion while the system is down, should talk to the feedstock suppliers. Again, sampling is key. Often, an outage is an opportunity for additional changes to an operation and these changes can have unforeseen effects on production capacity and quality.
Also, it doesn’t hurt to have the maintenance group review the steam system — traps, control valve responses, boiler chemicals, etc.
After reviewing the results of the teams, develop a plan to get the column running again. Finally, consider inspection of the heat exchangers and the tower as soon as possible.
Dirk Willard, consultant,
Wooster, Ohio
Puzzler Is Ending
This issue does not contain a new Puzzler because CP will cease running the department at the end of this year. Responses to the December Puzzler will appear online. Thank you for your contributions over the years.
DIRK WILLARD, Contributing Editor
