liquid water carryover in natural gas

[rotw]

Hi,

Can someone indicate a formula to calculate the quantity of liquid water carry over in a natural gas stream (hydrocarbon gas), knowing the gas composition, actual pressure and temperature ?
It can be also a formula which is function of the relative humidity ?

any indication please ?

Thank you
regards

 

[SNORGY]

Is this free water entrained in the gas stream or molecular water absorbed or adsorbed by the gas molecules in the gas stream?

 

[rotw]

this is just free water entrained in the gas stream.

 

[zdas04]

Then continue to dream.

This month's SPE Oil & Gas Facilities Engineering magazine has an excellent article by Mark Bothamley (Chief Engineer at JM Campbell) that goes into the exotic and esoteric math and the way I read his conclusion is that you gotta guess.

If you want water vapor (which SNORGY was correct in assuming when you said "it can also be a formula that is a function of relative humidity") then there is no formula. There are a family of curves called the McKenna Chart that you can look stuff up, but there isn't an equation to go with the curves. Several have been presented at eng-tips.com but none of them do a very good job across the possible range. You can generate your own relationship based on the ASTM tabular functions (which can be purchased from anyone who sells codes and standards). I did that and spent just over 2,000 hours developing three empirical equations that match every point on the McKenna Chart. I use the equations in programs that I write, but I don't give them away (I figure at my hourly rate I spent nearly $400k developing them). You could do the same. I found that a solve block in MathCad with 7 unknowns gave me the best fit, and it improved when I put temperature in an exponential term, but the content of that term was kind of difficult to determine. It is a different equation less than 3 bara, more than 10 bara, and between 3-10 bara.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[rotw]

Thank you David for your insight.
I was not aware it was that difficult.
I looked into the well known formula used for moist air then I only guessed that it would be a different story for other gas like Hydrocarbons.
But then how this is being calculated then in Hysys when you use a mixer component?
Do they have some complex built in algorithms for that?

 

[zdas04]

Pretty much the standard answer to anything about HySys is "they have complex built in algorithms (and tables) for that".

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[PaoloPemi]

as said by zdas04 there are no simple formulas capable to provide high accuracy,

however you may consider ISO 18453 standard,
it's more accurate than charts,
another good option is GERG 2008 and, of course,
a EOS with complex mixing rules.

All these methods are available in tools which you can utilize
in union with Excel, Marlab or Mathcad,,
or embed in your code,
see for example

http://www.prode.com/en/properties.htm

 

[rotw]

Thanks for the link. I am currently looking at the ISO 18453.

By the way with regard to the pressure of saturation of the water (in order to calculate the relative humidity for instance) and dew point calculation, Is it correct to used the IAWPS steam correlation when the water is mixed to Hydrocarbon ?
Is it a kind of good approximation?

I have heard that in presence of acid gas especially at high pressure, the water dew point prediction using the IAWPS correlation becomes less reliable. is it the case especially for acid gas or are you aware about other gases than acid gas which compromise the use of the correlation above ?

 

[PaoloPemi]

it depends from the model which you decide to adopt,

ISO 18453 has modified Peng Robinson with specific alpha functions and temperature dependent Kij
this because water is a polar fluid and std. VDW mixing rules don't work well.
Note that ISO 18453 has a relatively limited range of application but it is accurate.

With water vapor pressure model (based on IAPWS or another vapor pressure correlation) you treat water as separate component, you cannot not get accurate values with this assumption (although it can work in some cases).

A better approach (especially if you have H2S, CO2 Glycols etc. in the mixture) is to use a EOS with complex mixing rules (Huron Vidal, Wong Sandler etc.) or association term (as Cubic Plus association) or others,

all these models are not easy to code,
as said I prefer to use a process library (PRODE PROPERTIES) which does all the calc's for me.

 

[rotw]

thanks Paolo.

quote
With water vapor pressure model (based on IAPWS or another vapor pressure correlation) you treat water as separate component, you cannot not get accurate values with this assumption (although it can work in some cases).
unquote

What are the cases where it can work ?

 

[PaoloPemi]

it may work well when you can model the components in the mixture as near ideal fluids,
for example at relatively low pressures and with limited interactions with water (not the case of CO2, H2s etc.)

 

[rotw]

understand. But what about low CO2 content (say below 5-10% mol weight) and pressure ranges say below 100-150 bar.
are you confident to treat the water as separate component in these ranges or should be more restrictive ?

 

[PaoloPemi]

you can compare your model with a std. as ISO 18453
or some reliable procedure (see above posts) or experimental data (tables, charts) and see how it works,
this is the best procedure for all cases as I do not know a
simple rule to validate a model apart to compare results against reliable data...

 

[zdas04]

And the 300 kg gorilla in that story is "reliable data". Calibration and repeatability of field instruments is not great and transporting a sample to a reliable instrument will always change it. Also, 5% CO2 is not "low". Take a look at the CO2 saturation curves in chapter 21 of the GPSA Engineering Field Data Book. They are really messy, and include an awful lot of water vapor per volume of gas. If you treat each component in a CH4, CO2, H2S system very carefully, then you can get a result that would lead you to a reasonable decision by combining the water vapor based on a mole percent of the constituents. With more than about 1% of the mixture being any heavier hydrocarbons you have to consider interactions.

Pressures below 100 bara is where the simplifying assumptions start deviating from reality. By the time you get to 4 bara mostly they don't reflect reality at all.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[PaoloPemi]

"Take a look at the CO2 saturation curves in chapter 21 of the GPSA"

I can easily agree that you must decide some criteria to select "reliable data",
this is a common problem in process simulation where one needs good data sources to validate the models.

 

[SNORGY]

Sorry if my word choice is (was) poor...

When I said "...free water entrained in the gas stream...", I was envisioning an otherwise dry - even saturated for that matter - gas stream carrying water droplets at some slip velocity at or higher than the droplet terminal velocity. In such case, an EOS is not of particular importance and GPSA Chapter 7 ought to be able to deal with it. To me, that is what is meant by "...liquid water carryover in a natural gas stream...", i.e., a natural gas stream comes along, picks up liquid water droplets, and carries them over to some other place. That is a transport issue, not an EOS issue.

 

[rotw]

Snorgy,

A more complete title would have been : "...Amount of liquid water carryover in a natural gas stream".
Trying to keep this issue focused to the EOS problem / correlation story that I do confirm was the intended topic

 

[delittle]

ISO 18453 has the advantage to give a "certified" accuracy within the allowable range of application (both usual and extended) which is good.
As noted by zdas04 a chart may result difficult to read and real accuracy in some cases may be questionable.

 

[zdas04]

In every flow I've ever had to deal with, you cannot isolate this problem to just one thing. Water vapor will condense. Liquid water will evaporate. Both things happen in response to the environment they find themselves in at that nano-second. Both the slip/drag and the EOS are important to understanding your ability to transport a volume of fluid from point "A" to point "B".

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[rotw]

The original subject of this post was the amount of water carry over. However this brought another question to mind.

In the case of a hydrocarbon gas comes close to its dew point and finally reach a point (temperature, pressure) where some amount of the gas is condensing and is carried over in the gas stream.
I deem the difficulty to estimate the amount of gas condensate is also a complex task, (or even more complex) and same questions as with water ?
Do we then have to deal also with tricky correlations not necessarily always valid and uncertainties issues, similar of what we have seen with water?