2 replies to this topic

[tech_phoenix]

Hello, I am simulating a packed distillation column in Hysys 3.2.
For separation of mixture containing 2500 kg/hr ammonia and 50500 kg/hr water. In this case ammonia is separated as gas from above and purified water is obtained from bottom.

I selected Peng Robinson package for solving this problem. UNIQUAC could also have been used, but binary coefficients were unknown. Is this selection suitable?.

Moreover I used a mixer for making water ammonia mixture, instead of using NH4OH which is aqueous ammonia, as it was not present in the component library. Is this step correct?

[siretb]

I would not trust Peng Robinson, and my choice would go to a system capable of handling electrolytes. because NH3 + H2O --> NH4+ + OH-
You might consider NRTL, that would be, with UNIQUAC, my second choice, if electrolyte package were not available.

Before doing any serious calculation, I'd check the flash / VLE comparing the output of the simulator with the Tables of the Perry of of the handbook, that tabulates, for different temperatures, the partial pressures of NH3 and water, for different NH3 concentrations.
I would also check the heat of mixing. Such data can be found in a library. At infinite dilution it is
-8.33 kcal/mol NH3, at 20°C.

If Peng Robinson passes both tests, that should be OK I doubt that.
If a specific package is available with Hysis, use it.

[Andrei]

tech_phoenix

Any thermodynamic system without interaction parameters will not be suitable for that specific application. No parameters will mean no calculation, or zero parameters will be no interaction. Either way it will not serve the purpose.
It is your choice what you want to simulate: ammonia absorption/desorption in water or ammonia hydroxide decomposition.
The powerful non-ideality of the ammonia-water mixture will not be correctly described by a simple equation of state thermodynamics. Exactly the ammonia dissolution/hydroxide formation is pushing the system into powerful non-ideality.
It is a little bit complicated to go to a electrolyte model, especially because data is even less available in that area. Instead what I suggest is to go to a liquid activity model (try NRTL) or a modified equation of state model (read a little bit about Sour-Peng Robinson).
And by the way, thermodynamics evolved a lot in Hysys since version 3.2, try an upgrade to a later version.