water dew point in natural gas
water dew point in natural gas
(OP)
I am evaluating the results from different packages (GERG, CPA and Peng Robinson with complex mixing rules) to calculate phase equilibria and thermodynamic properties of natural gas with CO2, glycol and water,
as software I utilize PRODE PROPERTIES,
I notice that setting the BIPs from internal database the results can differ from a set of reference values which I got from literature,
however when I utilize the utility for VLE data regression available with PRODE they match well.
I would ask if there is some criteria to evaluate the range of validity for the BIPs in database,
thank you.
as software I utilize PRODE PROPERTIES,
I notice that setting the BIPs from internal database the results can differ from a set of reference values which I got from literature,
however when I utilize the utility for VLE data regression available with PRODE they match well.
I would ask if there is some criteria to evaluate the range of validity for the BIPs in database,
thank you.





RE: water dew point in natural gas
anyway I understand your point of view, my usual way to proceed (when water with other "problematic" fluids as H2S, CO2 etc. are present) is to find some reliable VLE value and test the models (which is probably what you are doing)
when possible I try to fit experimental data but that may result difficult cause you need to redefine most or all BIPs.
About the range of validity, the operating manual of PRODE PROPERTIES says that BIPS have been calculated with data from DDB / Dechema,
I think they are good when you cannot regress VLE data (which is probably the most common case.)
RE: water dew point in natural gas
the values in the database are good when there is no time or VLE data to regress,
PRODE PROPERTIES includes a large number of BITs (compared to other simulators) so you have a good starting point in most cases.
However one should not forget the limits,
for example if you calculate a phase envelope a single set of BIPs would not be able to cover all the T, P range, the same may be true for columns etc.
In most cases a EOS (with complex mixing rules) may help but still there are limits to consider.