water dew point in natural gas mixtures

[cbadia]

I am evaluating the different options for calculating HC+water equilibria with PRODE PROPERTIES,

version 1.2 includes a extended version of Peng Robinson with temperature dependent BIPs,

accuracy on water dew points is good,
normally within 1 K comparing with standards as ISO 18453

this accuracy seems superior to the other packages based on PR which I have evaluated,

for those interested a copy of the software is available from prode com

[mderra]

yes,
the Extended EOS models available in PRODE PROPERTIES offer,
in my opinion,
a reasonable compromise between complexity and accuracy,
for natural gas + water
Extended Peng Robinson with temperature dependent BIPs
offers good accuracy and fast solutions.

[mderra]

yes,
the Extended EOS models available in PRODE PROPERTIES offer,
in my opinion,
a reasonable compromise between complexity and accuracy,
for natural gas + water
Extended Peng Robinson with temperature dependent BIPs
offers good accuracy and fast solutions.

[carlo.stenali]

in my opinion that would depend
from the filed of application,
see the notes on operating manual,

'http://www.prode.com/docs/pppman.pdf'

as alternative you may regress the BIPs for
the required range of temperatures

[asanter]

"as alternative you may regress the BIPs for
the required range of temperatures"

probably that is the best option if you have access to experimental VLE data,
however it's not easy to find C2-H2O, C3-H2O etc VLE data sets
does someone know where to find these data ?

[carlo.stenali]

there is some data available
in GPA and Dechema,
usually I do use temperature dependent (4 parameters) for C1-water
and constant Kij for C2..C6 and water
if there are large amounts of CO2 or H2S temperature dependent Kij
would provide a much better accuracy

[carlo.stenali]

there is some data available
in GPA and Dechema,
usually I do use temperature dependent (4 parameters) for C1-water
and constant Kij for C2..C6 and water
if there are large amounts of CO2 or H2S temperature dependent Kij
would provide a much better accuracy

[matrenel]

thanks Carlo