water hydrocarbons dew point in natural gas pipelines

[hcostello]

Our GC gives C1-C7+ , which is the most accurate method for calculating water / hydrocarbons dew points in a natural gas pipeline ?

[hcostello]

no solutions for water/hydrocarbons dew point calc's ?

[matrenel]

dew point should not be difficult to calculate for hydrocarbons (with SRK or PR EOS) but for hydrocarbon + water you will need some specific model,
if the GC interface is based on OPC or MODBUS you may consider Prode Industrial Interface which provides all those properties.

[hcostello]

thanks, OPC is Ok for connecting DCS (actually we have all data in DCS),
I have a few questions concerning the criteria for C7+ characterization, the correlation for water dew point (GPSA or complex model) etc.
Our client does laboratory analysis at regular intervals and we are close to the end of allowable ranges.

[matrenel]

you may download the manual "http://www.prode.com/docs/proind.pdf" for information,
according my experience as process engineer:
1) little amounts of heavy components have large influence on dew point, include only the C7+ components which you know are in the stream with amounts say > 1.0E-6 (you can compare results against a chilled mirror or laboratory instruments)
2) with GPSA correlation you can get a estimate of water dew point, if you know the real amount of water (in stream) use suitable thermodynamics to get more accurate resuts.

[hcostello]

yes, there is a section in ISO 11150 which gives some data about chilled mirrors sensitivity and suggested software settings to match results, however (IMO) ISO standards do not address the problem of estimating with accuracy water dew point in hydrocarbons, we have a spec. for water dew point lower than –6 C , our standard reference is a chilled mirror but it is difficult to discriminate water condensation.
The same for hydrocarbons, according ISO 23874 GC should measure all hydrocarbons in the range C5 to C12 , which is difficult (with standard instruments), I see that Prode permits to characterize C7+ which is Ok.

[matrenel]

for water dew point you may read ISO 18453 developed at GERG, it is based on a modification of Peng Robinson EOS with parameters fitted to vapour pressure of liquid water and ice (different parameters in alpha function above and below 0°C), it has several limitations but it has been verified to be more accurate than generic correlations included in many simulators.
IMO a better approach is based on EOS (Peng Robinson or Soave Redlich Kwong) with complex mixing rules as Wong Sandler, CPA etc. , I would recommend a reading of "Water content of high pressure natural gas: data, prediction and experience from field" developed at Statoil, they compare results from different methods showing the advantages of EOS+complex mixing rules.
I was informed that Prode can include different methods including ISO 18453 and EOS+complex mixing rules and this should give the "most accurate method" according the actual technology.

[hcostello]

that's interesting, I calculate water dew point with Aspen but I get really poor results
I read ISO 18453 , errors are acceptable in the range of application, I have not experience with complex thermodynamics but we are prepared to accept any reasonably good solution (we are mainly concernd with water dew point),
thanks for help.

[dbert]

I agree with matrenel, water dew point calculated with the usual methods available in most simulators can give poor results, I read recently a couple of threads (at cheresources) about HYSYS returning water dew point of -100 C (!!!) : really you should't consider those values.
Recently I have had the opportunity the utilize ISO 18453 on a realtime system (4 weeks continuos operating) and I can say it gives realistic values (within it's limits of application), CPA is another good option, though.

[hcostello]

dbert, did the composition analysis came from a GC ?
How did you measure water content ?
Thanks.

[dbert]

yes, from a GC.
I am not an expert in GC area, I do suppose with a Thermal Conductivity Detector but I'll ask tomorrow.

[dbert]

Hcostello,
I confirm TCD with a packed column and specific settings, range is 300-2000 ppm (water), the absolute value is quite unstable (need frequent calibrations).