Simulating a dividing Wall Column (Petlyuk) On Aspen

[quangkhoa90]

Hi everyone,

I am working on thermal couple column simulation.
While simulating a Petlyuk column, I face up to a problem.
The pre-fractionator has neither reboiler nor condenser, then the first time when the multicomponent feed is introduced to this column, the separation doesn't occur since there is no heat supplied. Then the Petlyuk configuration cannot be simulated.

How can I fix this ? There must be a way to simulate a Petlyuk column on Aspen.

Thank you for helping me out

Quang Khoa LE

 

[sshep]

Hello Quang,

If you try to use a series of Radfrac columns there is the inconvenince of converging two tear streams externally. You can do it, but it is very clunky.

It will work better to use the Multifrac model with an even vapor split to the second tower- this simulates the vapor distribution at the partition. With a little initialization you can usually make a pretty robust (for Aspen-plus) convergence model.

best wishes,
sshep

 

[quangkhoa90]

Hello Sshep

Thank you for your hint.
I tried the Multifrac model, but it seems not easy to make the model converged.
In fact, I have been trying to find the combination of heterogeneous azeotropic distillation and Petlyuk column.
For the Radfrac col, we can specify the Azeotropic convergence option for the column, but I didnot find out how to do it for the Petlyuk column. I think it may be important to mention it in the column spec option.

if you have any suggestion, I would like to know that.

Thank you very much for your anwser

Regards

Quang Khoa LE

 

[sshep]

Hey Quang,

It looks like the problem is your system rather than the towers. Use every trick to simplify this to get better convergence. Fewer components, better column initialization, stream initial estimates, replace 3-phase with free water if possible, check you 3 phase system (TXXY)for stability, etc. The azeotropic convergenge option is not anything special, it is just preconfigured convergence parameters which are suposedly better for azeotropic systems (not neccesarily heterogeneous or for your system).

More data about the system can help.

best wishes,
Sshep

 

[quangkhoa90]

Thank you for your advice Sshep,

I will try to give some good initial estimates.

Regards

Quang Khoa LE

 

[sshep]

Post back on how it goes. Often it seems more important that the estimates be on the proper side of the final answer rather than being close to the answer- i.e. explore a bit. When you do get a converged answer, save it as a separate un-initialized case for future reference.

Good luck,
Sean Shepherd