6 replies to this topic

[Doni Baskoro]

Dear all,

I'am a student and currently designing a surface facility from a gas well which producing 6 MMSCFD of 85% C1, 5% C2, 2% C3, 1% C4 and minor C5-C12 including 0.9% N2, 2.6% CO2 and 100 ppm H2S (Dry gas analysis). We got water too from the well as much as 8 barrel/day, I'm including this stream of water too in simulation.
PT condition, 800 psig 140 F

 Komposisi Gas.xlsx   10.45KB   91 downloads

What I'm going to ask is what kind of fluid package should I used?
I already using :

• All Peng-Robinson basis,
• Peng Robinson for the saturating of gas, then continued by a) NRTL and/or b ) UNIQUAC for the separation (HP, MP and MP)
• All a) NRTL and/or b ) UNIQUAC

Every method gives me a significance difference in condensate spec and I cant decided which one should I choose. Is there any of you guy can help me out? Any hint or tips, much appreciated..

Thx

Edited by Doni Baskoro, 27 November 2011 - 08:55 PM.

[PaoloPemi]

water is allways difficult to manage within a EOS, when possible (see : range of application) with water and hydrocarbons I prefer ISO 18453 (GERG 2004) or GERG 2008 , which are available in Prode Properties.
Base Peng Robinson even with extended alpha correlations (TWU or Mathias) can give unreliable results with water, note that ISO 18453 is a modified form of Peng Robinson (with specific data for water).
Models based on liquid acitivity (NRTL, UNIQUAC etc.) may be the best when coupled with complex mixing rules as Huron Vidal or Wong Sandler but in most cases BIPs are not available (Ok, one can regress but it is a tiresome work and VLE data may not be easy to find) as result one may get unreliable estimates.
If you have VLE data (for example from Dechema or similar collection) complex mixing rules may be a good option (in addition to those mentioned).

[explodingmouse]

The simple answer is: use Peng-Robinson. The PR fluid package is the industry standard for simulating gas streams. For the purpose of gas processing facility design for a relatively normal looking gas (such as yours) the results it gives you are more than adequate. Specifically to answer your question, PR is suitable for a water saturated gas.

The slightly more complicated answer is: which ever package most closely replicates the observed behaviour of the fluid. The problem when you are doing greenfields design is that you cant observe the behaviour of the fluid so you really dont know which fluid package is best. Now your best bet is to use what you have observed to work well in the past - which is why I recommend the PR package.

Keep in mind though, that the PR package describes gas behaviour well - you may have some unit operations downstream which may introduce new components which require a different fluid pacakge, or you may choose a different fluid package if you were specifically interested in the interaction of, for example, condensate and water.

[Doni Baskoro]

Thx for the responses

explodingmouse : The odd when I use peng-robinson thoroughly all the system that it gives me all water (99.9%) even in the first separator (HP) in the liquid stream which I know it shouldn't be there. Then this scenario gives me no condensate. But yes, I will use this FP for the saturation purposes only maybe.

PaoloPemi : Now that I'd like to use PR only for saturating, I'm still not sure what FP (NRTL, UNIQUAC or UNIVAC) used in the separator section, (two separator, HP and MP, and one scrubber). About the lack of BIP, I'm still counting on the estimation which hysys already had.

For now, I dont have any sources to give me an advice about this and perhaps when I finished this project I'll not confidence about the result.

[PaoloPemi]

I am not familiar with your simulator, I do use a different software, Prode Properties, that has the mentioned GERG and EOS with complex mixing rules which I prefer in these cases.
What I can say is that Peng Robinson (and Soave Redlich Kwong) with extended alpha correlation (Stryjek & Vera, Mathias, Twu etc.) do an excellent work in fitting vapor pressure of pure components, however with hydrocarbons and water conventional one-binary-parameter mixing rule (Van der Waals) has several limits, see the note from Stryjek and Vera in PRSV: An improved peng—Robinson equation of state for pure compounds and mixtures,
"The conventional one-binary-parameter mixing rule allows the correlation of the vapor—liquid equilibrium data for a wide variety of binary systems. Only for systems formed by a polar compound (associating or not) and a saturated hydrocarbon, are results poorer than those obtained with conventional excess Gibbs energy functions."
For solving the problem different methods have been proposed, starting from the work of Kabadi and Danner (1985) which proposed a two-parameter mixing rule, Huron & Vidal, Michelsen, Wong & Sandler etc. have proposed complex mixing rules and nowadays suitable (for these mixtures) thermodynamic packages are available .

Edited by PaoloPemi, 28 November 2011 - 02:47 AM.

[explodingmouse]

Doni,
Im not a fluid package expert and im sure paolo is giving you more specific advice with regards to the advanced capabilities of each package, but consider that when you design equipment, you add on at least 10% safety margin, therefore the accuracy to the additional decimal point is not necessary.

In regards to your HP separator, whats the vapour fraction of your dry gas, and your water saturated gas? If your starting gas is at 800psig and 140oF, what T and P are your HP separator operating at? How are you modelling your 8bbl/d free water?

If you mix your dry gas stream at 6MMSCFD with 8 bbl/d at the same condition, and put straight into your HP separator, im really not suprised that you get a liquid outlet thats 99.9% water. You have done nothing to condense any hydrocarbon, any thats present in the water is just what the water phase has scrubbed out.

[Doni Baskoro]

PaoloPemi and explodingmouse, yes you guys are correct, I've checked it again and it appeared that my NRTL package still using ideal gas as the vapor model. Now that I changed it with PengRob (NRTL coupled with PengRob for the vapor) , I did had the same result as all-water stream in the HPSep. The stream specification actually not differ that much with all PengRob basis.

Yes exploding, my bad you're correct, now I'm changing it, pengrob all the way is actually quite okay. Thx for the answers