7 replies to this topic

[niketas]

Dear Community,

 

I would like to model the system shown in the sketch. Basically

• there is a constant flow of gaseous reactor effluent into the vessel,
• coolant at constant temperature being circulated so that temperature inside the vessel is around 0 degrees Celsuis,
• constant venting of the lighter components and continuous GC analysis,
• liquid drop out of the heavier components; however, NO outflow - thus, sampling done ex situ

What modeling/simulation tools are appropriate for this kind of a system?

Is CFD better suited than a thermodynamic equilibrium approach?

 

Your feedback is much appreciated.

Attached Files

•  Separator Sketch.png   25.25KB   3 downloads

Edited by niketas, 06 November 2023 - 08:06 AM.

[latexman]

With no liquid outlet you will need a dynamic model.  Dynamic models are 10-100X harder than steady state models.  I recommend you add a liquid outlet stream.  Let it approximate the sampling frequency.

[Pilesar]

First step would be just a single stage flash calculation of the feed stream at constant temperature and pressure of the vessel. Second step would be saying 'that looks good enough!'

Edited by Pilesar, 06 November 2023 - 09:41 AM.

[Pilesar]

Are you a student? Better to post in the student section so you can get advice you can use. You don't list flow rates, but the process design looks extremely poor to me - just terrible. If this is not for lab scale equipment and very small feed rates, I suggest you start over. The diagram is pretty.

[niketas]

Are you a student? Better to post in the student section so you can get advice you can use. You don't list flow rates, but the process design looks extremely poor to me - just terrible. If this is not for lab scale equipment and very small feed rates, I suggest you start over. The diagram is pretty.

 

Thanks for the feedback! Well, I tend to agree with you on the "poorness" of the design but that's what I need to work with unfortunately - same setup for years if not decades...

 

It is a lab-scale setup with ca. 100 standard cc/min of inlet flow rate & volume of the vessel is 300 mL.

 

OK, claim is - the lightest component (say C₂ or C₃) will not stay dissolved in the liquid (C₄ - C₁₀ range) but instead nearly all of it will leave with the vent.

However, using HYSYS or Aspen Plus or ChemCAD, it is impossible to confirm since at equilibrium there will always be substantial C₂-C₃ in the liquid at the system pressure (say 20 bar).

 

In this case with no liquid outlet, will the system be at equilibrium or not? If not, what is the best way to model the depressuring/venting phenomena?

[niketas]

With no liquid outlet you will need a dynamic model.  Dynamic models are 10-100X harder than steady state models.  I recommend you add a liquid outlet stream.  Let it approximate the sampling frequency.

 

Thank you for the feedback! Tried the dynamic model, too. However, one way or the other, HYSYS resorts to the thermo equilibrium models. Hence, it resulted in the lightest components (C₂ - C₃) staying in the liquid throughout the integration, eventually reaching the steady-state value.

[MrShorty]

I agree with Pilesar, a "simple" flash calculation should provide most of the information needed (even if it doesn't fully answer the question of whether or not the system is fully at equilibrium).

 

Focusing on the main claim (nearly all of the C2/C3/lightest components will pass through the vessel), the one thing I see missing from the problem description is "quality" -- how much of the feed stream is actually condensed in the vessel. For example, if I see that I'm feeding 100 SCCM of gas to the vessel, and the vapor stream to the GC is flowing at 95 SCCM, That suggests that only 5% of the feed is being retained, and, since C2/C3 will strongly prefer the vapor phase, I'm probably justified in saying that nearly all of the C2/C3 in the feed stream is ending up in the vapor stream. The flash calculation should allow you put an estimate to what "nearly all" means.

 

Alternatively, I could have 100 SCCM feed rate, but only 20 SCCM in the vapor stream. That would suggest 80% of the feed is condensing, and it will be difficult to say that "nearly all" of the C2/C3 is coming out in the vapor stream. Again, a flash calculation should allow me to put a number to that claim.

 

Even if you end up deciding to explore the possibility that the vessel contents are not fully equilibrated, an equilibrium flash calculation should allow you to begin evaluating any claims about how much of what components are passing through the vessel and which components are being held up in the vessel.

[Bobby Strain]

You failed to show internals in the separator. And there must be something to assure the reactor effluent is cooled to 0C.

 

Bobby