How to Best Implement a Vapor Side Draw 1

[sshep]

Friends,

I have been asked to evaluate and design (if practical) a tower with two side draws. The feed is a binary mix of valuable n-octanol (45wt%) and n-decanol (50wt%) with the remainder being heavy and light tails which must also be removed. The proposed scheme would take a liquid sidedraw above the feed (n-octanol), and a second sidedraw (n-decanol) below the feed. In order to get best removal of heavy components from the lower side draw, a vapor draw is indicated.

The vapor side draw will be condensed. This column would be a vacuum tower running about 200 mmHga as set by the tower bottoms temperature. The side draw condenser can tie-into the main tower vacuum system if needed (i.e. to remove non-condensables). I envision a butterfly valve cound control the side draw, or maybe some sort of variable area condenser, but have little experience with vapor sidedraws.

My question is how to best implement the design and controls for such a vapor sidedraw. I have only limited previous experience to draw on for examples, and worry about the impact on tower pressure or other operating problems which could arise. I appreciate any insight, references, comments or other helpful information which you may be able to give. I can provide any additional process info if needed.

thanks,
sshep

 

[sshep]

Hello,

I have made further analysis and such a tower as I described looks feasible.

I am again requesting recommendations about best practices for implementing and controlling a vapor side draw. I have seen some side recifying towers which draw off vapors using a variable area condenser in the side tower, but my own limited observation of those type controls was that they didn't always look very stable (the classic debate: was it "tuning" or fundamental instability). If I use a control valve to regulate the vapor draw then I will be condensing at even deeper vacuum than my tower which seems unneccesary.

I have a chapter 7 "Control of Side Stream Draw-off Towers" from the Handbook of Seperation Technology, but there is remarkably little detail of how to physically control the side draw and the references are limited. I would trust better to examples of actual operating designs and the experience of my valued peers.

No opinion or comment in this matter is too small.

Thanks,
sshep

 

[lastone]

I wish I could help you more sshep, but I have never operated, seen, or even read about a vapor side draw. Certainly not saying they don't exist, I just don't know anything about them. It kind of seems like if you have a point in the tower where your vapor is meeting the spec you are looking for, why not stop there and make this the top of your column? It's definitely an interesting post!

 

[sshep]

Hello again,

I am reactivating this thread to try and get further advice. Lastone's comment was appreciated, but the only way to make that suggestion happen is via a second column producing a decanol distillate which is what I was trying to avoid (to save cost).

In the course of design I note that a variable area side condenser located below the feed requires the least hardware of any vapor side draw design if my side draw pump takes suction straight from the condenser. I note that the side pump can operate on flow control. The condenser will find a level to meet my flow setpoint (self correcting). The product flow can be reset by some higher composition control cascade (or advanced logic), which may include side condenser level for stability.

While this seems simple I have few examples to draw on. Again, has anyone seen a side vapor draw or can give me advice on what to do (or avoid).

Thanks,
sshep

 

[epoisses]

sshep,
I think you have gone where noone of us has gone before...
But just for my own simple understanding 100 miles behind you, wouldn't you be able to obtain the desired product, at least theoretically, by taking a liquid side stream from a slightly different tray? Could that be the reason why one doesn't see (m)any vapor side stream drawoffs?

 

[StoneCold]

sshep
I would guess that liquid draw offs would be much easier and from a heat balance point of view, probably better.
I am curious how you are going to keep the top and bottom of the column from drying up if you are pulling 95% of the feed off into side draws? Does the simulation look good with that much side draw?
Is this a big column (large risk) or a small experimental column?

Goodluck

 

[sshep]

Friends,

Thanks for these new responses. I have a bit of information relative to your curiosities.

My column would be roughly 120ft T-T, and 5 to 6ft dia; handling 12kpph feed and reflux of 25kpph. We actually have a column in similar service at another site (in Europe) with two side product draws and another intermediate drag (rarely used). That existing column is divided wall in the top section with the upper sidedraw on the product side of the partition and the lower draw below the partition. Both sidedraws are liquid. The accounts I have recieved are that the existing tower works great except for some leakage across the partition (bolted plates), however 20years of experience has shown me that sometimes high level reports of project success cannot really be trusted, and I am looking more into that design. In any event I am sure that I could build on that experience and make it work, however, one obvious theoretical improvement (in my opinion) would be a vapor side draw below the feed for increased heavies removal. Getting how guidance for implementing a sidedraw of vapor was the goal of my post.

I have alot of conventional column design experience. With respect to how to keep from running the column dry with large liquid sidedraws, one sure-fire method is a total trap-out, because you can measure the reflux returning below the feed. The example tower does not use a total trapout. It was instrumented cheaper using some calculational constraints (internal reflux minus sidedraw) and a dry cut switch in the suction line to trip the sidedraw pump if you draw too much. Having worked at sites in other countries, I have come to consider such decisions a matter of style. In the USA we seem to instrument stuff more than other countries (i.e. USA sites more likely to want a total trapout)- would be glad to hear the observations of others.

I have quantitatively considered (but not definitively resolved) the energy implications of a vapor side draw using my simulations. From an overall balance if I needed 5kpph extra reflux in order to take a 5kpph liquid side draw then the energy would be a wash, but the truth is that substantually more reflux would be required to get the same product quality with a lower liquid sidedraw because we are essentually just diluting down the feed heavies which pass by the lower draw-off tray in route to the bottom. A liquid sidedraw below the feed is more of a dilution effect than a VLE seperation (at least with respect to heavy impurities).

Thanks for your interest and best wishes as always,
Sean

p.s. Please feel free to speculate and comment somemore.

 

[DWColumn]

In Butane isomerization units the DIB is usually run with a vapor side cut. This gives a better separation from the C5s which go out the bottom of the tower. This vapor is often 45 to 55 % of the feed, with IC4 being most of the balance.

The type of condenser you mention is effectively what is used for control on some steam heated reboilers. The liquid level blankets the surface and thereby limits the amount of heat input. In the reboiler the process liquid is the cooling medium.

Dividing wall may be the best solution for the problem-- There are a number of papers on this type of design. Chemical Engineering and CEP (Chemical Engineering Progress- AIChE publication) have had several discussions on this topic in the last 3 years. One Japanese company had data (private communication) on a very high volume (90% of feed 99+% purity), high purity side cut. It was a liquid side cut though.