Anhydrous Ammonia Boil-off Flash Calculations for Ammonia Rundown to Storage Tank

[ankur61]

I want to perform flash calculations of anhydrous ammonia rundown at a slightly higher temperature (say -32 degrees C)and pressure (say 1.5 bara) to the ammonia storage tank operating at atmospheric pressure. I am using the open source process simulator COCO to perform these flash. I am aware of flash operations in COCO.

However, I am not sure which EOS to set up. As usual, many people set up Peng-Robinson as the EOS but I feel that Peng-Robinson may not be the right choice due to ammonia being highly polar while almost being a non-electrolyte or weak electrolyte. In my opinion PR should not be applied for anhydrous or aqueous ammonia.

What is your suggestion on setting up an EOS for performing the flash calculation while able to use the embedded EOS in the software such as PR, SRK, NRTL, Activity coefficient method of Wilson etc.?

 

[PaoloPemi]

for a binary you may be able to fit (with a VLE data-regression procedure) experimental VLE data in some (maybe limited) range but standard Peng Robinson may show several limits... a more suitable approach could be, for example, a EOS with complex rules (i.e. Peng Robinson + NRTL with Huron Vidal or similar mixing rules),
for low pressures you could also adopt a mixed model (i.e. NRTL for liquid activity and ideal for vapor phase)
again you can fit experimental VLE data to calculate the binary interaction parameters (BIPS).
Also there are models with support for association (see for example CPA), electrolytes etc. etc.
I am not familiar with your software ( I have Prode Properties, see

http://www.prode.com

which offers a free version and it includes the above mentioned models plus a data regression procedure ) and the options available but I do hope you'll fidna suitable solution., best luck.

 

[TiCl4]

Why are you worried about NH3 dissolving in water if it is anhydrous? I have the impression that PR does fine for single-component polar compounds. As a test, I checked PR (1978) flash calculations with pure water and checked to see if the results match steam tables.

PR: 200 C, 15.52 bara
Steam Table: 200 C, 15.55 bara

It's not too far off. It's not conclusive proof, but you may be able to check a few other known single-component polar molecules to see if it predicts well for single-component polars.

 

[georgeverghese]

Alternatively, it may be safer to use the Mollier chart for anhydrous ammonia - see Perry Chem Eng Handbook 7th edn page 2-215, fig 2-6. Trace amounts of water in NH3 can affect thermodynamics significantly.

 

[pierreick]

Hi,
You may want to try the Patel Teja Equation, more appropriate for polar compounds.
reference:
Equations of state Theories and Applications by Chao and Robinson
Chapter 21 Application of cubic equations of state to Polar fluids and fluid mixtures.
Good luck
Pierre

 

[pierreick]

Hi,
IF you don't want to use EOS, several options like Mollier Diagram suggested by George or NIST web where you can get graph or data tabulated:

https://webbook.nist.gov/chemistry/fluid/

It's a pity that the OP did not reply to suggestions.
Good luck
Pierre

 

[ankur61]

Dear All,

Thanks for all your responses. For the time being I will stick to Peng-Robinson since the simulation software does not give a multitude of options for selecting the EOS with some variations in the original PR and SRK EOS.

Regards,

 

[georgeverghese]

These folks used the PR EOS in COCO for the ammonia - water system, which is in agreement with what you intend to do :

https://www.ijstr.org/final-print/m...The-Cape-open-To-Cape-open-coco-Simulator.pdf

 

[shvet]

@ankur61
... I feel that Peng-Robinson may not be the right choice due to ...
Why feeling? Why not to compare the model chosen with the lab tests conducted? NH3 is well researched and thermodynamic data is readily accessible.

Note that right choice depends not on EOS selected but on coefficients a software provides. What Aspen says

Be aware
As per my experience with NH3-H2O stripping any model gives a huge error in viscosity prediction as ignores NH4OH.