3 replies to this topic

[antnfedak]

Hello!

My name is Anton Fedak. At this time my goal is to optimize a VCM (vinyl-chloride monomer - from PVC industry) condensation unit. This unite is composed of 3 vertical Shell&Tube condensers (TEMA BEM, Class C). The condensation is knockback condensation. At 1st stage of condensation there are 2 condensers working in parallel. At 2nd stage there is 1 condenser. The 2nd stage is connected in series with the 1st stage. The vapor stream flows inside the tubes, and condenses there. The vapor flows upwards. The cooling fluid flows in "parallel flow".

My question is:
since i have the filme of condensate flowing down the tubes and since the condensate that is formed at 2nd stage flows down through the 1st stage, a have a subcooling of a condensate at 1st stage, right? If it so, does this phenomenon cause better condensation at 1st stage or not?

Because i'm using Aspen Plus (i can't figure out how to run Hetran model inside Plus interface, i tried but it gives me a lot of errors, something about work not being defined), detailed model, to simulate the condensation, but the results given by the simulator are far more conservative than the real ones. For exemple, i took a samples of a vapor stream that exits the 1st stage and enters the 2nd, made a chromatography analyses to find out what is the composition of that stream, and it's composition in VCM is aprox. 57(%wt). What i get from Aspen Plus is 86(%wt) in VCM. What could be the problem?

The cooling fluid is water and the temperature difference of water at 1st stage of condensation is 3ºC.
The stream to be condensed has about 1.5% of non-condesables.

Thank you!

Attached Files

•  Flowsheet presentation.bmp   1.22MB   26 downloads

[srfish]

Wouldn't this act as one long tube but give more condensate than a single long tube?

[Art Montemayor]

Anton:

Your question is related to what you are trying to do physically as well as what you are trying to simulate in a computer program. However, we don’t know what Aspen has as its algorithm for predicting the % of VCM condensed, and we don’t know what you fed as input as well as how you set up the simulation. Therefore, I have no comments on what is being simulated.

However, I do have some comments to what you explain and the first one is that your Aspen diagram does not depict what is happening (or what you are describing accurately. The difference is the physical actions of the condensing operation. Refer to my sketches in the attached Excel workbook.

I do not believe that you can do what you depict on your Aspen flow chart. You can’t expect to entrain all the condensed VCM out with the un-condensed vapor out of each vertical condenser. You would be designing a pneumatic conveyor instead of a condenser, if you tried to do that. I do not believe the scheme will work successfully in the field. The principle of a knock back condenser is to literally “knock back” the produced condensate upstream – by gravity forces – and not to knock it up (or downstream).

Refer to the graphical description of the similar apparatus described in U.S. Patent #5.558,746 which describes the operation in detail. In this application, the produced condensate is used to scrub and pre-cool the ascending vapors and consequently remove some of the tars before they access the condenser.

In your setup, you might be obtaining some subcooling – but only if you allow the condensate to fall back down (be “knocked back”). But your sketch doesn’t allow for that – so your question is contradicting what you propose to do.

I suspect you are trying to do something different from the conventional attempt to employ a simple knock back condenser, but you don’t explain what. As SRFish implies, why don’t you just employ a longer (taller) knock back condenser and be done with it?

Attached Files

•  VCM Knock Back Condenser.xlsx   168KB   56 downloads

[antnfedak]

Thank you very much SRFish and, especially, Art Montemayor for your replies.

As for my Aspen flowsheet, i know what a knock back condensation is, the thing is that Aspen Plus does not have a single peace of equipment that allows to add a vapor and condensate streams at the bottom of the condenser. That is why i've put those adiabatic flash separators. All they do is to separete the condensate from the un-condensed vapor. BUT those flash unites doesn't and will not exist
My purpose is to optimize its functioning, but always rely on the knock back condensation!

The vapor stream to be condensed is composed by 97%VCM, 1%methyl-chloride, 0.8%ethyl-chloride and 1.2%N2+O2. The termodynamic model i've chosen is Redlich-Kwong-Soave EO. The vapor stream enters at 48ºC and 6barG (that's why in simulation i use 7bar). The cooling stream enters at 14ºC and 4bar. The temperature difference of cooling fluid is 3ºC.

Is there a way of simulate in Aspen a real knockback condensation?

As for the reply of SRFish, wouldn't a taller condenser imply a grater/larger presser drop (not so much at tube side but at shell side)?

I attach the updated excel file. Take a look. The difference is at the entrance of cooling stream.

Attached Files

•  VCM Knock Back Condenser.xlsx   174.67KB   51 downloads