4 replies to this topic

[rosesdaf]

My distillation column is operating with saturated liquid. I calculated the bubble point of the feed at a given operating pressure, however when I run it on Aspen plus, a vapor fraction of 0.12 appeared. I double checked my calculation but can't seem to spot any error. I am just wondering if it's something that Aspen does or if it has to be my mistake. I lowered the temperature by 0.2C and it resulted in a vapour fraction of 0.

[MrShorty]

What thermo model did you use for your calculation of the bubble point? What thermo model is Aspen using to calculate the bubble point? I would guess that you and Aspen are using different thermo models when you each calculate the bubble point, giving different results (though a difference of only 0.2 C does not seem like a huge error).

 

If that proves to be the case, then you need to decide if 0.2 C error in the bubble point is worth wrestling with, and then decide which thermo model you believe is more accurate. Without more detail, I'm not sure any of us can speak to this process in anything more than general terms.

[rosesdaf]

I calculated the bubble point by hand using y=sum of Ki*xi in my shortcut method and ran it in rigorous simulation using LK-PLOCK.

[MrShorty]

So you're hand calculation is admittedly a simple, shortcut method. When compared to the simulator's more rigorous method, your shortcut is within 0.2 C of the rigorously calculated bubble point. Seems reasonable to me.

 

I guess I'm not understanding exactly what the question is. Do you need to hand calculation and the simulator calculation to agree more closely than 0.2 C? Are you wanting to modify the simplified hand calculation? If you want to modify the hand calculation, what improved model are you wanting to propose? (My mentor who got me started ChemE was always much more willing to help if I offered a proposal of my own for him to critique than if I just came to him with an open question and no proposals.)

[StealthProg]

Even the same methods implemented in different simulators can give different results.

 

Component databases may not have exactly the same criitical properties for components or have a different method for calculating interaction parameters or the interaction parameters may be based on laboratory results.  All of which will lead to slightly different answers.