Joule-Thomson Inversion Point

[jesahe]

Hello Everyone,
Using WinSim, I am simulating natural gas stream conditions from an onshore wellhead to downstream of the first choke.

Wellhead Properties
T = 300F
P = 12,000psig
SG = 0.646
ΔP across choke = 5000psi

After the choke, the gas temperature increases a few degrees (T = 302.2F). However, real world observations make me think the temperature should decrease significantly. I suspect that the increase in temperature is due to the high pressure of the well, which puts my Joule-Thomson coefficient past its inversion point. Therefore, as the gas expands it heats up rather than cools down.

Basically, I don't fully trust these results and I would appreciate any comments or direction. Thanks!

 

[dcasto]

Believe it or not, methane will increase in temp during a pressure drop at those high starting pressures.

 

[delittle]

can't you model the valve with a flash operation such as H-P and a proper EOS such as Soave Redlich Kwong ?
Joule-Thomson calculated at initial t, p provides dT/dP at initial conditions, it is NOT constant over that pressure range,
better to integrate enthalpy or another property to estimate the final results,
I have a different software (Prode Properties) but results from different simulators providing the same composition and thermodynamics should be close,
I can't compare the values (you do not provide the composition) but I expect that you should get correct values with the proper method...

 

[MortenA]

Reverse (or negavite) JT is a fact (judging from your text i guess you already knew this). The range that you mention would be the exact point where it occurs.

But i would recommend using PR (or deriviates) for these high pressures since you are outside the recommended range for e.g. SRK (i dont see what EOS you use).I dont see exactly why delittle

Best regards,

Morten

 

[delittle]

Morten,
no problems about different opinions about applications of SRK vs PR equations of state, these EOS are available in almost all simulators and one can test both,
by the way many papers suggest SRK over PR for natural gas mixtures.
At some conditions positive dT/dP may occour but for such large pressure drops (ΔP across choke = 5000psi) I would model the choke with a H-P or equivalent operation (see my previous post) instead of assuming constant dT/dP (i.e. estimating final temperature based on a constant Joule-Thomson).

 

[MortenA]

I used to use winsim a lot, but its been a while. I seem to remember that their valve model was rather flexible. But i would assume that the standard was an isenthalpic process and not using a constant JT coefficient.