Column Scan to identify operational Problems 1

[phex]

Dear all,

I'm looking for a service company for a special column scan. A little background at first.

We have a process to separate Aromatics from near-boiling Nonaromatics. This is an extractive distillation process using a selective solvent in order to change boiling points and thereby enabling us to separate the aromatics from the feedstream.

In the column, we have a point were two different liquid phases (solvent and HC reflux) are distributed onto a random packing with the still mixed aromatic and nonaromatic vapors rising against these liquids.

In one of our plants, the process works not as expected and we suspect the packing area distribution to be the cause. The problem is that the reason is probably not the hydraulic or liquid distribution, but the concentration distribution (meaning the HC reflux and solvent mixing) over the packing area. We want to have one perfectly mixed liquid phase from the two liquids distributed on top of the column.

In order to verify our suspicion, we are looking for a way to scan this packing for the uniform mixing of those two solvents. My idea was to use some kind of tracer in the HC reflux and scan for the tracer as well as for the hydraulic distribution in this packing.

I am looking for companies operating in Europe, who can do such a scan. Can anybody give me a lead where to look? Google tells me some companies, but most of them operate in the US or do not provide a service as I am looking for.

Is it even possible to do a scan like this?

Many thanks for your help.

Regards,
Chris

 

[AndreChE]

Call Tracerco.

We did column scans (three columns) one month ago in Portugal. They are based in UK, Netherlands, Italy...

http://www.tracerco.com

AndreChE

 

[sshep]

phex,

It sounds like you want to put a radioactive tracer in one liquid and then make scans- I never heard of that and doubt the practicality or safety.

I think what you want is not possible. The normal tower scan only detects density differences. If you run 4 scan lines and the two liquid densities are different enough, you may see a concentration difference, but otherwise you can only scan for uniform (or not) liquid distribution.

best wishes,
sshep

 

[DBenson78]

Can I suggest as AndreChe you call Tracerco, they have a new service which is called Thru Vision

http://www.tracerco.com/products-processdiagnostics-thruvision.htm

This allows you to obtain a 2D density map through the column as opposed to basic data from scan lines.

I hope this has been some help

 

[leeeeroy]

In addition to above comments, this technique is not limited to scanning for density differences. I've heard of this company injecting radioactive tracers into fluids going through packed beds and then measurng the distribution and residence times with externally mounted detectors. This is detailed on their web site:

http://www.tracerco.com/products-processdiagnostics-distributionstudy.htm

 

[sshep]

The tracer link was an interesting read- the example was a more typical pipe response detection which is familiar. A non-volatile tracer should circulate with the solvent so with enough detectors doing the column equivalent of a CAT scan, perhaps you could learn something. This would be a new application of radiographic tower monitoring for me, so if phex finds out something I hope he will enlighten us.

As an alternative it occured to me that since extraction solvent is usually sent back to the column somewhat colder than the temperature profile in the tower, it might be possible to make a temperature survey around the tower just below the solvent feed distributor. Do this either by shooting skin temps or by thermal imaging. This is similar to a normal distribution problem troubleshooting by temperature survey, but with the extra advantage of having cold solvent to give larger temperature gradients.

Incidently our most recent designs for extractive distillation internals have a liquid collector from the bed above (including a welded in wall collector) that fully mixes the collected liquid with the solvent feed and then distributes the mix onto the packing below. This is as opposed to simply letting the liquid fall from the bed above onto the solvent feed distributor.

best wishes,
sshep

 

[phex]

Thanks all for the info provided. I suggested contacting Tracerco and talk with them about the possibility to do a scan like this.

As soon as I have any news, I'll let you know if we've been successful.

Regards,
Chris

 

[AndreChE]

http://www.tracerco.com/pdf/Tracerco_News_V1_E1_3Qtr_06.pdf

 

[STI666]

I formerly worked for Tracerco and they definitely have the technology to tag the feeds, one at a time, and perform a distribution study through your vessel to provide the required information. They use short lived radioisotopes and will perform all of the necessary calculations to ensure the job is safe and no one will be affected by the tracer. I am not sure what the price tag for this service will be though.

I do not believe the Thruvision would give the resolution required to differentiate the feed based on the detectors and method they are using. The density of the packing must also be added in when comparing densities through packed towers. What may be a 5% difference in density between two fluids can become less than 1% when trying to read through the packing.

Scanning Technologies is also a reliable provider of Gamma Scans for trayed and packed distillation towers with operations mainly in North America. They have developed some strategic partnerships and can offer these services to other countries around the world. Their pricing is usually lower than Tracerco/TruTec due to lower overhead.