This Month’s Puzzler
We installed a new knockout drum with a better mist eliminator (Figure 1) to replace a drum that was giving us problems. The new drum, like its predecessor, is at the inlet of two thermal oxidizer (TOX) systems. “Flow is controlled by a pressure transmitter downstream of the mist eliminator,” notes the production foreman. Operations had project engineering install two redundant transmitters in the old drum. However, the impulse tubing for both transmitters plugged so often that the TOX systems downstream tripped on low pressure every few weeks. Those systems run at 2 psig at the inlet with the trips set at 0.5 psig. The pressure drop across the mist eliminator is about 0.3 psig when clean and about 0.5 psig when dirty, according to the manufacturer; our maintenance engineer says 0.75 psi is more like it. We installed two redundant transmitters on the new drum and their impulse lines also suffer from plugging: six months after the new drum was installed, the TOX systems have started tripping again.
Our grizzled, old maintenance engineer is disgusted. He hoped this problem would be solved by now. He says we should have included a steam purge in a manifold and the mist eliminator; he drew up a sketch in a few minutes when I asked his opinion on how to address the problem (see the figure). The young engineer who ran the project didn’t see a need for it, figuring the improved mist eliminator would do the trick.
What do you reckon that we should do to solve the problem? Why do you think the mist eliminator failed to work effectively to prevent fouling of the impulse tubing? Can you suggest a better way to control the TOX systems or reduce the tripping? Is it possible to improve the performance of the mist eliminator?
Eliminate The Impulse Legs
Consider the following points:
1. Impulse legs to the pressure transmitters are plugging frequently. Eliminate impulse legs altogether by using pressure transmitters with diaphragm seals and capillaries. (Some vendors call them remote diaphragm seals.) [Ed.: for information on such seals, see “Rethink Distillation Column Pressure Measurement.”] Thus, the process stream containing particulates does not come in contact with the capillary. The diaphragm is bolted onto the tank nozzle. The diaphragm senses pressure, which is transmitted to the pressure transmitter. This arrangement should reduce, or even eliminate, the frequent plugging problem.
2. The diaphragm arrangement should give relatively trouble-free operation. However, if you experience problems with deposits sticking on the diaphragm, provide an isolation valve for the diaphragm, so that it can be cleaned and put back in service.
3. Capillaries use a fluid to transmit a pressure signal to the transmitter. In an extremely cold or hot environment, the fluid may not work effectively. Specify to the pressure transmitter vendor the range of ambient temperatures expected. In these situations, you may need to provide a means to maintain temperatures in the working range of the capillary fluid.
4. The problem statement does not specify size distribution of particles in the wet gas. However, if there are particles roughly 10 micron or above, you might consider passing the wet gas through a cyclone separator and then to the knockout drum that contains the mist eliminator. The key is to reduce particulate load to the mesh pads.
5. You also can get cyclonic action by introducing the wet gas stream tangentially in your knockout drum. However, at this stage, such a change may be difficult because it requires installation of a new nozzle followed by pressure-rating certification.
6. Separation efficiency of mist eliminators rises sharply with increasing velocity — until you reach an upper limit where re-entrainment will severely affect efficiency. Thus, there is an operating range of velocities where the mist eliminator is effective. Ensure the velocity of wet gas through the mist eliminator is within the design range.
7. Although chemical species shown in the wet gas are not highly corrosive, if you have older vessels made of, say, carbon steel, you may have corrosion; this could cause the mist eliminator’s plugging. During the next turnaround, take samples of the deposits in the mist eliminator.
8. The problem statement does not indicate any problems with the TOX. However, it seems to me that with considerable droplet carryover in its feed the TOX could experience flame stability problems.
GC Shah, senior advisor
Answer Some Critical Questions
This actually is a two-part problem. The first is fairly simple but the second will require time and personnel resources.
The first problem — the impulse line plugging — can be easily solved by installing capillary pressure transmitters, thus eliminating impulse lines and, therefore, plugging.
Installing two transmitters is puzzling. If one impulse line plugs, what would keep the second from plugging?
The line purging approach, while a good idea for keeping the impulse lines free of plugging, seems to be introducing vapor back into the process downstream of the mist eliminator.
The second part of the puzzler is more difficult, mainly because there are not enough data to attempt to solve the problem. This initiates a few questions:
1. Why was the previous knockout drum replaced? Industrial tanks usually will last longer than the engineer who orders them. It may get to the point that no one will remember who requisitioned it in the first place.
2. Was any tank wall corrosion found in the replaced tank?
3. If the answer to the second question was yes, was the new knockout drum made of corrosion-resistant material or coated to prevent oxidation?
4. What were the parameters of the mist-laden gases entering the mist eliminator (temperature, density, viscosity, etc.) and what were the “others?”
5. Have provisions been made to clean or backwash the mist eliminator? The mist eliminator functioned acceptably for six months and then seemed to require cleaning; notice the increase in pressure drop with age.
6. Has the mist eliminator been installed properly? The sketches indicate it is installed perpendicular to the knockout drum. Would it be better to install it horizontally so that the liquid condensed and the material clogging the impulse lines would drip back into the knockout drum? This would require a change to the internals: fixtures would have to be installed in the drum to hold the mist eliminator and isolate the downstream gas/vapor from the rest of the tank.
7. Finally, because TOX operation is shutting down “every few weeks,” this is not only an operation’s problem but also increases costs and reduces company profits. Therefore, I recommend putting together a team that includes a chemist, a metallurgist, a chemical engineer, and operation’s personnel to thoroughly examine the process and determine the root cause or causes of this problem.
Stephen Curyk, consultant
Lago Vista, Texas
Change The Controls
First of all, pressure doesn’t control flow as the operator described. Pressure control valves being called flow control valves is typical; however, proper flow control would be a good idea. I would suggest an insertion-type device, i.e., a pitot tube, annubar or perhaps even a target. This could give you real flow measurements to aid in a material balance because vent flows feeding TOX systems tend to get forgotten in plant balances.
By redundant, I assume that operations switches back and forth between the transmitters. Or do they run them together? It probably won’t matter because the tubing comes from the same tubing bulkhead nozzle into an 8-in. blind flange. Both tubes will foul at the same time.
The maintenance engineer is on to something. Purging with steam may unplug the impulse lines but it also will cause additional condensation downstream; check flow capacity of the liquid discharge nozzle and pipe from the knockout drum to the TOX fan. Also, if there are agents in the gas stream that will react with steam or impurities in the steam, then corrosion or plugging in the TOX systems could be a problem. Maybe, pulses of steam would be okay but compressed air probably would make more sense. Perhaps a non-fuming, less likely to precipitate acid such as acetic or inhibited hydrochloric acid might do a better job of cleaning the tubing. Another idea would be to build the tubing so it can be disassembled quickly without interfering with plant operation.
Alternatively, you could locate the transmitter diaphragms on the drum flange directly; it’s a low-pressure vessel (ASME: ≤ 15 psig), so modification is okay but follow ASME construction practices anyway. I like the maintenance engineer’s idea of separating the pressure taps, so long as they are not too far apart. Note that different instruments always will produce different readings even if on the same tap.
Steam might work well to clean the mist eliminator. Nitrogen is too expensive at up to $3/100 ft3 of liquid in cylinders at 230 psig; air could be a fire hazard because flammable liquids are present. Steam may provide the added benefit of possibly dissolving some of the solids into the steam condensate. The temperature of the steam coming out of the regulators won’t be saturated, so be mindful of the temperature limit on the mist eliminator materials.
Dirk Willard, consultant
We’re having trouble measuring the solids’ level in our forced-circulation crystallizer that feeds storage bins. The product is a solid with a dielectric constant of about 2.7; however, some impurities have constants of about 12 and others about 5. Screw conveyors transport the water-logged solid to bins feeding re-dissolvers that are the next step in purification. We have difficulty controlling the water content from the crystallizer; excess water flows into the bins. We have used tuning fork contact probes but always have suffered issues related to fouling. Problems included overflows of the bin and grossly inaccurate indicated levels. Out of desperation, we’ve switched to guided radar for continuous measurement.
Is guided radar the right choice? Why do you think the tuning fork probes failed? Can you suggest other options for reducing errors in solid measurement?
Send us your comments, suggestions or solutions for this question by August 12, 2019. We’ll include as many of them as possible in the September 2019 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, 1501 E. Woodfield Rd., Suite 400N, Schaumburg, IL 60173. 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.