Which EOS for a low-pressure mixture of HC and non-HC gases?

[KernOily]

Good morning!

I am simulating a small gas treating plant, for the purposes of throughput increase, that is used to cool and compress a low-pressure mixture of gases (we use HYSYS and PIPEPHASE). This gas is casing vapor (casing gas) from oil wells. A typical dry gas analysis (60° F, 0 psig) is (mole fraction):

CO2 0.55
O2 small, 0.05 or so but can be as high as 0.1 or 0.15
N2 small, 0.05 or so
H2 small, 0.05 or so
H2S 0.02 or so
C1 0.35
C2-C6+ remainder

At process conditions of 230° F and 2-3 psig, at the plant inlet, the analysis is:

H20 ±0.85
The rest is normalized from the above analysis.

I have tried PR, SRK, and CS, among others, in my simulator and I can't get a decent phase envelope. Not even close. It is nowhere close to reality in predicting the phase equilibria at process conditions. The process consists of cooling the gas in a shell-and-tube to about 205° F and then secondary cooling in a fin-fan to about 130° F or so before subsequent compression to about 10 psig.

What brought this all up is - we can't get the simulator to match the exchanger duties we measure in the plant. They are off by at least 30%. I can make 'adjustments' to the simulator to make it match but that is not good practice and will not provide a general solution, and I need a general solution.

Do any of you have any experience with low-pressure mixtures of polar and non-polar gase like this? Which EOS have you used? Any other thoughts or recommendations?

Thanks ! ! ! ! ! ! ! ! Thanks!
Pete
P. J. (Pete) Chandler, PE
Mechanical, Piping, Thermal, Hydraulics
Processes Unlimited International, Inc.
Bakersfield, California USA
pjchandl@prou.com

 

[TD2K]

PR or SRK would be typical EOS for this system. Have you been able to do a mass balance on the downstream plant to first verify your wellhead dry gas analyses? If the gas plant is mostly compression then this likely isn't possible but can be a good check if a significant amount of liquids condense.

When you say you can't match the duties you measure in the exchangers, how is this being done? For example, I've had limited success in matching air side calcuated duties from air coolers to the process side even when we did a survey at multiple points across the air cooler outlet measuring air flow and temperature.

 

[MortenA]

With regards to EOS/phase invelope fit:

I dont have any experice that tells me that it should be nessesary to change model at low pressure. Actually gas tends to be more "ideal" at low P/T. Well actually your gas is a little hot but i dont think that actual decomposition needs to be considered.

When you say:"H20 ±0.85" are you 100% certain whether this is absolute mol% or % saturated? And what have you used for input? Quite often i would guess that the water content would be reported in % saturation and this will not be equal to mol% that are normally entered into the software. You would need a water saturation tabel in order to convert to mol% and then enter it from there.

With regards to the HX's

How did you simulate the HX? By the normal "ideal" exchange or by the actual matching component?

If the model predict a different degree of condensation than what you experiense then this could be the cause for a major difference in duty

Best Regards

Morten

 

[KernOily]

Guys thanks for your replies.

>PR or SRK would be typical EOS for this system.

I thought so as well, but...?

>Have you been able to do a mass balance on the downstream plant to first verify your wellhead dry gas analyses? If the gas plant is mostly compression then this likely isn't possible but can be a good check if a significant amount of liquids condense.

The samples for the dry gas analysis are taken at the plant outlet, after cooling and compression. The dry gas is flashed back to the plant inlet conditions to get the equilibrium water saturation and thus the composition at the plant inlet; that is where the 85 mol % H20 comes from. The mass balance works out pretty closely but only AFTER I adjust the exchanger duty. The way I do this is by increasing the UA (exchanger area) artificially until the duty matches the field measurement. So the EOS is not doing the phase equilibrium correctly.

>When you say you can't match the duties you measure in the exchangers, how is this being done?

The gas is first cooled in a S&T HEX using steam generator feedwater. I calculate the duty by using Q=mdotCpdeltaT from the feedwater temperature rise. For example, if I calculate that 10.5MMBtu/h is required to heat the feedwater using tubeside field-measured temperatures of 160° to 205°, and the gas cools shellside from 230° to 205°, the cooling duty reported by HYSYS does not equal 10.5MMBtu/h; it calculates 8MM or so.

>For example, I've had limited success in matching air side calcuated duties from air coolers to the process side even when we did a survey at multiple points across the air cooler outlet measuring air flow and temperature.

I think this speaks to the empirical nature of air-cooler design, fan blade deflection, dirt on the fins, etc. No matter what they say, there is still some 'art' to that business. Good luck with this. I have NEVER been able to get an air cooler to match the manufacturer's data sheet performance, let alome a simulator like HYSYS. I can usually get close though, say within 10-15%.

I remember reading someplace in the PIPEPHASE manual that PR and SRK are not good for mixtures containing water vapor. I'll dig it up and post back here later, gotta go right now. Thanks!
Pete
P. J. (Pete) Chandler, PE
Mechanical, Piping, Thermal, Hydraulics
Processes Unlimited International, Inc.
Bakersfield, California USA
pjchandl@prou.com

 

[MortenA]

I tried to look at a similar model my self and notised a small "warning" issued when you calculate the phase envelope: It done without water! Thats why it look so strange (and the reason why there is no bubble point curve - this light mix of HC's and inert gasses is fully condensable at any P/T

Best regards

Morten