Which Equation of State

[YEngineer]

I have been asked to calculate maximum H2S release rate from an acid gas injection well.
Proposed stream has +/- 40% H2S; 50%CO2 and the rest are HC's. Everything is on paper - no lab data.
This work will also include well modeling to predict friction losses under different wellhead injection pressures and maybe different injection strings.

The client is a midstreamer and has licenses for Pipesim (SLB) and Hysys (Aspen).

The EOS Pipesim uses are:
Soave Redlich
PR76
PR78
Multi reference fluid corresponding states CSMA
Benedigt-Webb-Rubin-Starling
Cubic plus association
GERG 2008

HYSYS might have some more but I prefer to start and finish modeling in Pipesim. From my experience in Lab, I know the predicted results from EOS (WinProp from CMG) were +/-20% off with what we measured in the rig, especially in critical region.

Question: Which of the above EOS is best suited for acid gas modeling?

I did some runs on Pipesim and I could not figure out how to export results to excel or even print them. Help-file is useless. It used to be so easy on pipesim 2003....
If anyone is familiar with pipesim, I will also very much appreciate some instruction on how to export/print results?

thanks

 

[PaoloPemi]

I have a different software (Prode) which includes about the same models (plus some specific additions),
I would say it depends mainly from the properties you wish to calculate, for example, for density of gas mixtures a EOS (with volume translation) can do a decent work altough multiparameter models (GERG, MBWR) are certainly more accurate, however, when comparing (gas phase) density predicted by PRX vs. GERG differences are usually well below the 20% you mention, similar comments for enthalpy and entropy (gas phase), for phase equilibria things can be more difficult if you have some water (wet gas) a possible option could be a EOS plus Activity with complex mixing rules otherwise the same comments as above...

 

[YEngineer]

Thanks Paolo,
phase equilibrium and density are the parameters I am after mostly.
The proposed stream is dry, or with a scenario where water can be present up to 0.01%

 

[PaoloPemi]

small differences in composition may have a large impact on results, for example for (molar fractions) H2S 0.4 CO2 0.5 C1 0.04 C2 0.02 C3 0.02 IC4 0.02 Prode with PRX model calculates the whole phase diagram (up to CricondenBar at about 90 Bar) while with GERG 2008 the procedure calculates both bubble and dew lines but stops at about 80 Bar (for calculating the whole diagram it requires to modify the settings for vapor/liquid identification criteria), anyway if you compare (for this composition) values on bubble , dew and constant fraction lines calculated with Prode PRX or GERG 2008 the differences (phase equilibria test) are about 1 K,
the same for gas density, for example at 340 k 90 Bar.a GERG 2008 predicts 230.56 Kg/M3 while Prode PRX 232.18 Kg/m3 , a small difference, to observe larger differences you must calculate values close to the critical point, Prode PRX predicts for this mixture critical pressure 89.46 Bar.a and temperature 328.51 K , if you calculate gas density close to this point, for example 90 Bar.a 330 K (less than 1K and 1 bar from crtical point) Prode PRX returns for gas density 339.3 Kg/M3 and GERG 2008 316.6 Kg/M3 there is about 7% difference but very close to the critical point... for liquid phase differences are larger due to the well known limits of cubic EOS, anyway for example at 90 Bar.a 310 K Prode PRX predicts 644 Kg/M3 while GERG 2008 returns 633 Kg/M3, similar results with MBWR (multiparameter) models,
I would expect not too different results with Pipesim or Aspen...

 

[georgeverghese]

From memory, many of the EOS you've stated have an upper limit on total acid gas content ( H2S + CO2) of 5% mole basis. Beyond this, would suggest you talk directly with the Thermo - EOS experts in the process simulation company. Not familiar with GERG2008.

 

[PaoloPemi]

agreed, probably the best way is to compare the predicted values with measured data, for example
"Densities of Carbon Dioxide + Hydrogen Sulfide Mixtures from 220 K to 450 K at Pressures up to 25 MPa" Stouffer, Kellerman, Hall, Holste... J. Chem. Eng. Data, 2001, 46 (5), pp 1309–1318

however I would mention that some recent papers report values showing larger errors for GERG model, for example, Alfonso Gonzales Perez in his thesis published on 2016, Etudes expérimentales et modélisation du comportement de phase et des propriétés de transport des mélanges lies à la capture et au stockage du carbone Etudes expérimentales et modélisation du comportement de phase et des propriétés de transport des mélanges lies à la capture et au stockage du carbone, compares a new version of SAFT with GERG and other EOS,
there is a mixture with composition CO2 0.4212 , C1 0.4053 , H2S 0.1735 with values for density, herebelow I report the measured values and those calculated with GERG, PRX and PR models in Prode
T(K) 353.59 P(Bar.A) 140.06 Density EXP, M3/KG 201.88 GERG 196.03 PRX 194.02 PR 197.33
T(K) 353.58 P(Bar.A) 175.11 Density EXP, M3/KG 263.48 GERG 254.96 PRX 250.23 PR 254.45
T(K) 353.71 P(Bar.A) 270.36 Density EXP, M3/KG 395.18 GERG 381.67 PRX 374.52 PR 382.9

from this work it would seem that std. Peng Robinson gives (for this mixture) good results, however other works report different values and I would suggest to compare values...