THIS MONTH’S PUZZLER
We overhauled our isobutene/n-butane fractionating column during a recent turnaround. We cleaned its sieve trays, replaced the reboiler (a circulating thermosyphon) and slightly lowered the suction nozzle to avoid splashing the bottom tray. The bottom-most thermocouple failed during start-up total reflux; it went full-scale. We switched reboiler control to an available thermocouple higher in the column. We replaced the failed thermocouple after the column was dehydrated. Now, we’re having trouble with the temperature profile below the feed tray. The reboiler differential temperature is sporadic. What’s causing this and what can we do?
LOOK AT THE START-UP
The thermocouple failure was probably incidental to the problems with the new reboiler and the column. If it was reading high before it failed, failing high unfortunately would signify a problem. A high reading would indicate poor circulation or fouling.
The best way to approach this problem is to review the history of the start-up and the design and installation of the new reboiler. Someone may have a case history somewhere that relates in some way to your troubles.
Total reflux is a common start-up technique because the column is isolated from upstream variations; quality control for the bottoms and condensate is easy. Total reflux allows for loop tuning and is a good way to dehydrate the column after hydrotesting. However, this technique can pose problems if the feedstock is contaminated with a material that could gum up the trays or react because of a higher-than-expected bottoms temperature.
The questions to ask are: 1) How different was temperature indicated by the failed thermocouple, assuming it worked briefly, from the one it replaced? 2) How does the temperature at the bottom compare now to what it was with the old reboiler? 3) What does the heat balance around the column and the bottom of the tower — including the reboiler — look like? And 4) is the cooking caused by a failure during start-up or is this problem caused by poor selection or installation of the reboiler?
Cooking can be a real problem during start-up. Imagine heating butter and brown sugar on a stove. Eventually, all the butter boils away and you're left with tar. That's what cooking is. If the trays are fouled, the liquid will build up. There still will be separation, but the condensate reflux ratio will be increased to achieve the isobutene split.
If cooking is the problem, it may be possible to flush more isobutene than normal into the column and hope that some of the gum can be dissolved. Or, you could go the other way and hope that a heavier component would suppress the high volatility of the lighter compound boiling away in the bottom of the column. Another alternative would be to inject a low-viscosity high-molecular-weight solvent into the column to dilute the bottom. Kister recommends all three approaches in "Distillation Operation," McGraw-Hill, 1990.
Dirk Willard, lead process engineer
Fluor Global Services, Inver Grove Heights, Minn.
Our pilot-plant batch reactor uses a new fiberglass packed-bed caustic scrubber for final polishing of vented gas before it goes to a thermal oxidizer. The design pressure is 150 psig for the reactor. The vessel is protected by a 4 in. × 3 in. pressure relief valve set at 35 psig, which we had sized assuming a heater coil valve-open scenario using steady state. The pressure in the reactor frequently spikes above 28 psig — which has blown the relief valve several times. We rely on the pressure drop through a throttling valve and knockout drum to prevent the vent gas from overpressurizing the scrubber, which typically operates with a discharge pressure of 1 to 2 psig. Now, management wants to use the reactor exclusively for heat-of-reaction measurements. We're unsure about how to isolate the fiberglass scrubber in the event of a runaway reaction and how to properly size the reactor's relief device. What's your advice?
Send us your comments, suggestions or solutions for this question by October 10, 2012. We'll include as many of them as possible in the November 2012 issue and all on ChemicalProcessing.com. Send visuals — a sketch is fine. E-mail us at ProcessPuzzler@putman.net or mail to Process Puzzler, Chemical Processing, 555 W. Pierce Road, Suite 301, Itasca, IL 60143. Fax: (630) 467-1120. Please include your name, title, location and company affiliation in the response.
And, of course, if you have a process problem you'd like to pose to our readers, send it along and we'll be pleased to consider it for publication.