VLE Data in Aspen Plus (BIPs Regression) 1

[vinramvolero]

Hi,

I have a multi component distillation system modeled in Aspen Plus. We have VLE data at three different pressures. So I have three data points with overhead and bottom compositions. How should I proceed from here to validate my existing Aspen Plus distillation model and regress the existing binary interaction parameters.

Any help is much appreciated.

 

[apetri]

it is not difficult to regress VLE (as well as LLE and SLE data) sets at different temperatures and pressures,
supposing you have access to DDB database or a similar source you obtain X,Y values at different T,P
see herebelow an example for Methanol-Water (free DDB data)

you enter these values in data-regression utility to calculate BIPs , I include an example generated with my copy of Prode Properties and Peng-Robinson-NRTL with Huron-Vidal mixing rule

with these data-regression utilities you can calculate the optimal parameters starting from different properties as equilibrium data, enthalpies etc.

note that you can even process the input data automatically (supposing you have a file to read)
I have not access to Aspen but I am sure there is something equivalent you may look for...

 

[Iomcube]

vinramvolero, REGRESSION UTILITIES are avaiable in DWSIM as well as in Prode (as stated earlier). However unit consistency should be used while migration regression answers across different applications. Even with same units different application emplot different techniques of regression. Am I nor right, apetri ?

 

[apetri]

Iomcube,
if you mean the possibility to transfer the calculated values from Prode (or another simulator) to Aspen, I agree that it could be not easy, many models in Prode and Aspen adopt different parameters and formulations, for example the extended versions of Peng Robinson in Prode Properties are different (they adopt more parameters) than the similar models in Aspen and require different BIPs,
in my aswer I suggest a procedure to follow (presuming Aspen includes a data-regression routine)
I have not experience with Dwsim and I do not know if the models are the same (some sites report it gives the same results of Hysys) but also in that case I suspect it could be not easy and may require some data conversion...

 

[vinramvolero]

Thank you apetri. That was helpful. I just finished this project and here are few lessons.

- The regression in aspen plus is easy to perform.
- I had hysys model with a special property package and all hypos defined in hysys. Migrating this to A+ took real effort.
- Hysys do not have regression analysis. And if you don't have hysys fluid package in A+ then you have to a) find similar property method in A+ b)Regress a lot physical property parameters - This took longer
- Verify the two simulation by doing XY plots and flash calculation
- Proceed with VLE data

 

[vinramvolero]

Hi Iomcube,

Thank you. I never used DWSIM. I am aware that it is open source. I will give it a try.

 

[shvet]

How should I proceed from here to ... regress the existing binary interaction parameters.

See Teaching Module Thermo-001.