Benzene Water Stripper 2

[sshep]

I am designing a steam stripping tower to remove benzene from process water. The feed will be approx 100ppmwt benzene.It is desirable to recover the benzene as a hydrocarbon phase from a decanter style reflux drum, and reflux the aqueous phase back into the column.

My question is: Do I need a rectifying section to get assurance of a phase split at the top of the tower?

We have several towers in this sort of service at various plants, including a design I worked on about 15 years ago (it is in the same situation). Most of these towers do not have a rectifying section and also do not actually produce a hydrocarbon distillate. Either the decanters never worked, or operations gave up on that operating mode. Now all such towers return a benzene rich aqueous condensate (either continuously or intermitently) to the feed tank. Oil is skimmed from the feed tank a few times per year. Simulations show that a hydrocarbon phase should form, but experience indicates otherwise. As I am redesigning a condemed tower I have an opportunity to do something better and would appreciate any advice on this subject.

If more info is needed to give me an answer then please ask.

best wishes, sshep

 

[epoisses]

Sshep, not sure if I understand the problem, re the existing towers, is it the towers that don't work as per the simulation or the decanters?

 

[DickRussell]

Rectification is not the problem; it's the absolute amount of benzene that is. Assuming the stripping part is performing adequately, the high activity (low solubility) of the benzene in low concentration in water will let it be stripped easily out of the water. Solubility is much higher than 100 ppm (around 4-5 times that at 100 F, per API TDB), but the amount of stripping steam needed overhead is only on the order of 5-10% of the feed, depending on operating pressure. You can check the stripping factor, S=KV/L to see what V/L you need at any column temperature; you want S>1.2 to do any significant amount of stripping, and perhaps at least 2.0 for really good stripping with the trays you have, which probably is around 10-20 if the column was built for this purpose originally. Check the Brown-Souders chart for a visual.

You ought to make sure the feed to the top tray is hot,because you don't want to do partial condensation at the top trays with cold feed. You want the second phase formation only in the condenser.

Rectification isn't a good thing for this separation, either, because you can't really go across the hetero azeotrope in the column without going two liquid phase somewhere. Once you have a separate organic liquid phase in the column, it is apt to run down separately and somewhat unpredictably until revaporized by the steam coming up. Breakthrough intermittently is a possibility in that situation.

In the decanter, the oily layer will collect on top, and assuming initially the top level is kept below the top of the baffle the interface will gradually drop. Flow of reflux can be reset by interface level, once it is comfortatbly below the top of the baffle, so as to ensure plenty of space for settling. Due to the very low amount of benzene in the feed, the amount of benzene actually overflowing the baffle will be that low, so perhaps withdrawal from the benzene side of the baffle can't really be controlled, except manually each shift or by on/off, high/low switch.

Assuming the feed came from some upstream operation in which benzene and water/steam were present together in large quantity, with a decanting happening there, you need to consider if it would be best not to run a decanter on the water stripper overhead at all, but to condense (perhaps against the feed to heat it up) and send the condensate back upstream to that decanter. That could be a better place to address the very low amount of benzene in the stripper feed.

HTH

 

[sshep]

DickRussell,

Thanks for your attention and detail. Instead of "refluxing" a rectifying section, you make me think that I should just have a stripping section and then add the aqueous distillate phase back into the feed so that it gets the normal preheat. Mixing like this is not usually efficient, and this is why I asked the question.

I have some experience with azeotropic distillation and have sometimes seen the aqueous side tower be an actual distillation (in the case of an aqueous feed), while the hydrocarbon side is just a stripper. Intuitively, this design decision would appear to be dependent on the amount of hydrocarbon in the feed, but this is where my lack of experience comes in and the reason for my question. In our case the feed is well below saturation (100ppm actual vs 1800ppm saturation).

New question: Your example cites a decanter that includes an interface controller, where as my existing decanters in this service are 3 compartment gravity seperation designs- i.e. center feed with baffled overflow, and underflow-overflow. Do you feel that a interface controller and overflow baffle is a better design for a case of small hydrocarbon flow?

Again my problem is that the hydrocarbon phase (reportedly) does not form. If this was one column then maybe there could be a problem, but this is 5 columns in similar service in three seperate units. Unfortunately only one column is actually still piped up in the original manner so it will be awhile before I can make a visit to see a sample for myself. Obviously the benzene does go somewhere, but in all cases it is via some (or all) aqueous distillate being sent back upstream to the water feed tank. If we could get a hydrocarbon phase we would route it back into the process, but I don't really want to send free water back to the process where it could cause corrosion problems.

thanks again for your help,
sshep

 

[DickRussell]

If indeed the aqueous phase is to be returned to the column, after presumably being cooled quite a bit so as to enhance phase separation in the drum, then yes, it ought to be reheated. Efficiency in distillation isn't the issue. Any benzene in the feed or reflux will be bottled up in the top of the column, blocked by stripping steam, until it gets out overhead. You could be taking a pile of water out with the benzene dissolved in it, depending on how the baffling is functioning. Presuming all you want to do is clean up the water for downstream use as BFW or for disposal, and only very insoluble stuff is in the feed (benzene), then you have only a stripper. The big problem is that very small absolute flow of benzene. To get good phase settling, you need some volume/residence time in decanting, and you'll have an interface. Since practically all the volume entering the drum is water, trying to control the interface will be elusive. Of course, if you do control that level, then increasing pump flow to lower it will be just temporary, because top level will immediately drop below the baffle. I imagine you'd have a constantly cycling interface level, but that may be ok, as long as you have the means of getting the rather irregular benzene volume out of the overflow. I'd really like some others to weigh in on your situation. I imagine others have encountered this large water/tiny organic phase issue before.

If you had miscible chemicals in the feed, and you needed to rectify out the water as well as strip the water, you'd have a different case. Given what you have, my closest experiences are such that we could send all the column overhead (steam plus stripped organics), after condensing, to be sent back upstream, to where a gross oil/water separation was done in the first place.

 

[StoneCold]

sshep
What about absorbing the benzene using a Macro Molecular Polymer instead?

http://www.akzonobelmppsystems.com

I don't have one of these systems myself but I have looked at it as a waste water cleanup step.

Regards
StoneCold