Distillation column with GERG 2008 AGA 2017 2

[mdepeche]

I did some tests with the steady state column in Prode Properties to compare EOS models such as Peng-Robinson, Soave etc. (the Extended PRX-SRKX versions in Prode) with GERG 2008 model which should be more accurate on both vapor-liquid and enthalpy,

for several separations (HC) I observe that results are not too different (GERG 2008 vs. PRX models in Prode),

also, the phase diagrams (GERG 2008 vs. PRX models in Prode) show limited differences (1-2 K or less after tuning EOS) which could explain the limited differences with columns (a similar VLE),

I would ask the Prode experts in this forum, should I prefer GERG 2008 when solving compressors, pipelines etc. or is it preferable to adopt GERG in all cases (VLE etc.) ?

 

[apetri]

in my opinion (not from an expert) multiparameter models such as GERG 2008 do not offer large advantages (compared with extended EOS) for phase equilibria, however for V, H, S (and derived as speed of sound etc.) these models are more accurate,
certainly, if you wish to model a compressor, GERG 2008 can be the best option (considering the model includes 21 chemicals only),
but a extended EOS such as Prode PRX can give values reasonably close to GERG reference,
also Prode can include MBWR which could be another good option, depending from application...

 

[cosh]

I would say it depends from application,
for process design in many cases a 1-3 % error is acceptable,
in other cases (for example fiscal metering etc.) one wish to obtain best accuracy.

 

[mdepeche]

yes, when required one should adopt GERG 2008 or other complex models.

 

[rotw]

Depends on your pressure range. High pressure and composition with various non methane compounds, I would prefer GERG for compressor calcs. Not only because of enthalpy but isentropic exponents and compressibility could effect power and discharge temperature calculations significantly. I think however that it would still be hard to have these results when obtained via GERG and other EOS deviate one another by more than 5 percent.

 

[mdepeche]

rotw,
I agree with you, I compared GERG 2008 and several EOS models with polytropic compression test cases provided by R.A.Huntington in 1985 (and 2017) papers,

for CO2 Pin 75.85 Bara Tin 310 K Pout 413.7 Bara Tout 460 K

Huntington 3 points model (included in Prode Properties) gives these values:

GERG 2008 Head (Kj/Kg) 80.336 Eff. 0.6439

Lee Kesler (Ploecker) Head (Kj/Kg) 80.680 Eff. 0.6326

SRKX (Soave Extended) Head (Kj/Kg) 81.35 Eff. 0.707

other models (Peng Robinson etc.) give similar values (with differences about 1 Kj/Kg on Head),

for these calc's GERG 2008 is probably the best option.