BP distillation column

[GERLASA]

Hallo!

My question is dealing with the optimization of a unit operation distillation column in a BP plant in Hull (UK).
Formic-, Acetic-, and Propionic Acid are to be distillated. But the variables of separation and energy consumption are conflicting. The distillation of propionic acid does not reach up to 50%. How can the separation efficiency be increased? Or would it be better to take out the propionic acid from the bottom and separate it in a column on its own? Or does the installation of a additional column cost too much? By the way, does anyone know the costs of intalling a distillation column?
I'd be grateful for any help.

 

[StoneCold]

Gerlasa
I would suggest that you simulate your existing conditions in Aspen or Hysis. Then compare the performance of the existing equipment to the simulation. If they don't match up well then you may have some damaged trays or a flaw in the orriginal design. If the simulator matches the real world fairly well then you can use the simulator to help find the best solution to your problem, whether that is to add another column or not I don't know. Once you determine if you need another column then you can have any of the major column manufacturers (Sulzer, Koch-Glitch, etc) quote you a new column.

Good Luck
StoneCold

 

[25362]

Is azeotropic distillation involved ?

 

[GERLASA]

To Stonecold:
Tahnks for your reply. The thing is that i did the simulation in Aspen. But i do not have data about the real performance.

To 25362:
Two of the involved components form an azeotrop. But not propionic acid. Its water and formic acid.

 

[25362]

Sorry. But according to my sources the propionic azeotrope boils at 99.9^oC and contains 82.3 % w/w water.

 

[GERLASA]

To 25362:
But then there wouldn't be a problem with the distillation of propionic acid in this column. Or am I misunderstanding something?
In the column i'm dealing with only valuable acids are supposed to be separated from non-valuable acids. It doesn't matter at this stage if there is an azeotrop or not, as long as it is distilled to the HEAD. The azeotrop will be a problem later.

 

[25362]

As one may see it depends on the amount of water, since the azeotrop is water-rich, insufficent water may leave acid undistilled, being its BP ~ 141^o[/sup[C.

 

[GERLASA]

To 25362:
Hallo again!
Thanks for your hint.
In my simulation i tried to increase the amount of water. But propionic acid was reacting in the opposite way, the distilled fraction is decreasing a lot, what still leaves me wondering. In contrast, if i decrease the amount of water the distilled fraction of propionic is improving. Can this be due to the presence of methyl-ethyl-ketone?

 

[25362]

There is a tertiary azeotrope of water:formic acidropionic acid (18.6:71.9:9.5) with BP=107.6^oC.

MEK forms an azeotrope with water boiling at 73.5^oC containing 11% water. This azeotrope may be "stealing" the water needed to form the other azeotropes that boil at somewhat higher temperatures.
How much water are you adding in your simulation, is it in a ratio to the valuable acids parallel to the azeotropic compositions ?

 

[GERLASA]

I just added an arbitrary amount of water to see what is happening. Actually i do not have so much of a clue of how the azeotropes are formed and how much water they require.
The column has a feed of 31.8 tons/hr and the share of water is 21.9. I increased to 30 and decreased also to 16, without changing the other component shares.
Can i add as much water as i need or is there a danger of overflow?

 

[25362]

Any tower can handle a determined vapor load before it floods. I believe the programs mentioned by StoneCold (which I don't have access to) are capable to deal with azeotropes.

If these show that by reducing the water input you improve the C3 acid recovery, then, if at all practicable, that could be a path to follow.

 

[UmeshMathur]

GERLASA:

This is very late in the day, as your original post was late in June, so I apologize.

Did you include the calculation of vapor phase non-ideality in your simulation (organic acids are known to dimerize partially in the vapor phase)? Ignoring the efects of dimerization can lead to very serious errors in calculated K-values. Note that vapor phase fugacity coefficients for the organic acids can be as low as 0.3 to 0.4. Treating the vapor phase as ideal could therefore result in K-value errors of over 200%.

The best reference for this discussion is "Computer Calculations for Multicomponent Vapor-Liquid and Liquid-Liquid Equilibria" by Prausnitz, Anderson, Eckert, Grens, Hsieh, and O'Connell (Prentice-Hall, 1980). Vapor phase non-ideality for such systems requires considering the “chemical” approach and the Hayden-O’Connell method is useful, as discussed in this book which is a great reference for learning the fundamentals.

For the actual binary interaction parameter estimation, I suggest that you use the activity coefficient regression package that comes with your simulator (so that you're sure the physical properties of the pure components are consistent with what you'll end up using when you do simulations).