3 replies to this topic

[Wit Knien]

Good day people,

Currently I'm doing internship (following a course in Chemical Engineering) at a company which uses anhydrous 2-propanol in several process steps, as part of my assignment I am to investigate the possibilities to reduce the amount of 2-propanol that is needed. Sure hits such as reuse of streams which are only slightly contaminated by demiwater, are being performed on production scale, cutting the total demand in half.

The problem lies in the other half of of the 2-propanol waste stream, on a daily basis 30 liters of 2-propanol contaminated with water (~ 4 w.t.%) and polish liquid (~4 w.t.%). Due to the complex composition of the polish liquid, separation has been limited to distillation (the difference between the lowest boiling component of the polish liquid and the 2-propanol is over 100oC).

Trial runs with distillation have been carried out and the aquired distillates have been tested using gaschromatography and mass spectrometry, which are clean. Ergo the distillate contains only water and 2-propanol. To lower the water content to an acceptable limit (up to 1 w.t. %) adsorption using 3A mol sieves is being considered, early trial runs have shown that this is possible to achieve.

Now that we've come to the mathematical model of the adsorber bed, problems are arising. First of all I'm not certain about the difference of using a liquid or vapour feed for the adsorber, personally I think the usage of a liquid feed would increase contact time and the lower temperature has a positive effect on the amount of adsorption (w.t.) 3A sieves can achieve.

I've read the how to guide for adsorber design, posted elsewhere on this forum (http://adsorption.co...dsorberDes2.pdf ) and I've basically become puzzled. The distillation aspect of this project offers hardly any problems in regard to the mathematical model but the adsorber bed is something I've never worked with before. Basically I want to get a rough estimate for the breakthrough time of my fluïdum to be able to make estimates on the dimensions.

The problems I'm currently encountering are:

- Determination of the absorbate diffusivity
- Determination of superficial velocity, I know my end outcome but the reflux is hard to determine with experiments performed on lab scale using a glass setup

Any advice on how to tackle this problem would be greatly appreciated.

[Art Montemayor]

Nick:

I envy you. You have been handed a project that is not only fun and interesting to design and build, but it has all the makings of yielding a success story for you and a great feeling of accomplishment. Adsorption is my favorite Unit Operation because it is so simple and practical in both its design and operation. Regardless of what you are reading or being told, there is no high mathematics (differential equations and integrals) involved in the calculations and certainly no concern for the infamous “break through”.

Your link to the reference you cite does not work – but that is not important for now. What I can advise and counsel you on is more practical and direct and should lead you straight forward to useful work and results leading up to your eventual design.

First and foremost: Remember that you will not be dealing with any author of a textbook or an influential and prestigious academic exercise in explaining how adsorption can be mathematically expressed and taught. You should apply your good, practical Dutch ingenuity (the way other famous Dutch engineers have done in the past) and start to identify where the source of your adsorbent is located and how to obtain basic design information from them. This should be a piece of cake for you since you are located in Europe and molecular sieve technology is well known and available there.

You should seek actual, field data or history on alcohol adsorption on molecular sieves and the expected or recorded sorptivity of propanol on specific molecular sieve types. This information may be available through the mol sieve manufacturer/supplier. Also inquire on what have been successful superficial velocities employed through the mol sieve beds by the alcohol in question. This information, together with the specific heats of the recommended mol sieve (do not attempt to identify the proper mol sieve yourself; you are not the expert. Rely on the manufacturer for that recommendation; they are the experts.) and the maximum regeneration temperature recommended for the adsorbent are essential for the bed design.

Be sure to discuss the various types of potential regeneration schemes you can apply – taking into consideration the inherent trade-offs in each type. You should prepare detailed PFDs of each regeneration cycle so that you can clearly identify the differences between each type of regeneration.

The actual adsorption cycle should be clearly identified by you as to:

NEMA cycle. What is the adsorption drying time? What is the adsorption regeneration time (complete with draining, purging, heating, cooling, and priming)? Take into consideration – as part of your Scope of Work – that this is basically a BATCH process. It is steady state only when one of the adsorbers is on-line; it is batch because the adsorber beds have to be switched and regenerated. Therefore, common sense should let you know that you are going to require peripheral equipment to carry out the total process:

1) An off spec propanol Day tank. This serves as the receptacle for the contaminated propanol that will be subjected to adsorption for cleanup. It should be sized for handling all the expected offspec propanol that will be processed;
2) An off spec propanol feed pump (probably spared);
3) A supply of gaseous nitrogen (if this will be the regen gas used);
4) A polished Day Tank for the product propanol prior to it being pumped back to prime storage and subject to purity analysis;
5) If you regenerate the adsorber beds with a dry, hot gas – such as nitrogen – you will need to drain and purge all liquid from the beds prior to regenerating them. That means you will probably need a dump tank (with a pump) to handle this side stream.
6) Your regen vapors will have to be discarded, flared, or condensed. If you decide to condense them and recycle them, then you will probably need a chiller condenser with chiller fluid;

The above gives you an idea of what is involved in this type of adsorption system. The design is now the next step. I recommend you leave the design to a recognized and leading mol sieve engineering designer who can do all the details in accordance with your scope of work and Design Basis. You can do the preliminary pilot plant and feasibility studies and experiments, but the warranted safe design should be done by experts. You will find that knowing the sorptivity of the mole sieve beds, the recommended superficial velocity, and the regen temperature required that you can easily put together a pilot plant to prove the feasibility of the process and its results.

This is already a very long post, but I thought I should give you all the above “heads up” pointers on how I would go forward with this type of project.

I hope this helps you out, and good luck.

[Wit Knien]

Mister Montemayor,

First of all thank you for your time to reply, furthermore the praise of the Dutch was very very well received. I know that the project I've gotten is very interesting, but because of the fact that the company developed many of the processes internly, it's also challenging.

I certainly got a lot of things to look into now, thank you for that. The link to the absorber design manual did not work from this computer aswell, it was actually a how to guide, posted by you in another thread, I'll add it to this post though.

Not trying to sound cocky or anything, but I'm curious about the pressure swing adsorption, since that would seem to be less of a strain in regard to energy consumption and the complexity of the system.

I'm sorry that I have to keep my post somewhat short but it's the start of the weekend and some other matters are requesting my attention.

Thank you again and I'll look into your suggestions on monday.

Edit: the file upload keeps failing, I tried several times (.pdf file of 536 kb).

http://www.mijnbesta...NIKY6AGAV4Z.pdf

I've uploaded the file there now.

Edited by Wit Knien, 10 August 2012 - 12:02 PM.

[Art Montemayor]

Nick:

I don't think it is worthwhile to think about applying Pressure Swing Adsorption (PSA) on this application. PSA is all about regenerating by expanding a compressed gas within the adsorption beds. First, how are you going to establish a gaseous state? Second, how are you going to establish a COMPRESSED gaseous state for propanol and water? Please let me know how you can propose this.

The simplest and conventional manner of regenerating a mol sieve bed is to drain all liquids from the bed and subsequently heat up the bed with hot gas, followed by cooling it usually with the same gas. Of course, that means you have to fill the bed with your fluid prior to putting the regenerated bed back on stream. All this means that there is a measured, methodical number of steps in the regeneration procedure.

Once you find out from mol sieve suppliers what they recommend as regeneration temperatures for the bed, bear in mind that the highest bed temperature will be lower than the regen gas temperature - usally around 500 to 700 ^oF. Your block valves on the adsorption towers will have to withstand this temperature PLUS the cycling of the temperatures from near ambient to maximum regen temperature. A very special type of block valve is called for here and its selection will take research and testing.