19 replies to this topic

[jamese]

I understand that water content (or water vapor pressure) of any gas above a film of water on a pipewall is given by water dew point tables .

Such tables are commonly available on the internet, for instance;

 

http://www.duncanrog...s/dew-point.asp

 

I have always thought that the water vapor pressure/water content was only dependent on temperature, ie that at any given initial pressure the gas water content is the same regardless of initial system pressure.

This is also suggested by water dew point tables which never make reference to initial system pressure.

 

I appreciate that the water content of a gas can be varied from the initial starting value by moving away from the initial system pressure, dependent on the availability of free water.

 

However, is my understanding of the bit above in ittalics correct?

Or in reality at higher pressures does the water content/vapor pressure start to vary with initial pressure, from the values stated on commonly available dewpoint tables?

 

Thanks

 

James

 

[chemdoc]

you may consider the limits of application of Raoult's law

 

http://en.wikipedia....ki/Raoult's_law

 

at high pressures you may need a EOS or equivalent

[Art Montemayor]

James:

 

You raise a very good and interesting point regarding the water content in a gas and the method(s) of measuring it accurately.  Your initial statement is not, however, correct.  The water content above a film of water on a pipe wall is dependent not only on the temperature of the pipe wall but also on the pressure of the system.  Unfortunately, a lot of degreed engineers continue to put out shabby and incomplete data, reports, and calculations citing so-called “dewpoints” without mentioning the condensed fluid in question (water, in this instance) and the pressure of the system measured.  They continue to believe that the reader can read their minds and understand that the measurement(s) are done at atmospheric (14.696 psia) pressure – when, in most cases, we are dealing with process pressures.  That is the case with respect to your referred source.  Note that the author refers to so-called “free air” – which is left to the whims of your imagination and not defined.  What is not mentioned is that the data given is at atmospheric pressure.

 

If you are a chemical engineer, you should understand and be able to carry out routine dewpoint calculations – the identification of a process temperature where a condensable vapor starts to form the first drop of liquid on a colder surface (such as a heat exchanger tube).  To prove my assertion that the dewpoint is dependent on pressure as well, consider a 100% saturated steam system at 100 psia.  The temperature of this steam is defined thermodynamically as 327.81 ^oF according to the NIST database.  That means that if we subject this steam to a pipewall temperature of 327 ^oF, a drop of water condensate will form on the surface.  Therefore, we have identified the dewpoint of the steam system as such – but at 100 psia.

 

When we have a mixture of vapors – some of which are not condensable (such as air) – we have a similar situation except that we can maintain that the system pressure will remain steady since we are not condensing the non-condensables.  Such is the case of compressed air that is saturated with water moisture.  The higher the system pressure, the less the water vapor content – in accordance with Dalton’s law on partial pressures – and the lower the dewpoint (when referred to atmospheric conditions).  However, the water content in compressed air will drop out by condensation when cooled to a temperature higher than that of the dewpoint referred to at atmospheric conditions.  This is the effect that makes refrigerated compressed air dryers work.

[jamese]

Thanks for the replies.

I think I might be getting this spectacularly wrong;

If I have two cylinders containing water and a gas at the same temperature, the only difference being the pressure.
I thought the vapor density of the water was the same in both cylinders

[Art Montemayor]

^James:

 

^{Sorry, but you are wrong.  The cylinder with the higher system pressure (a pressure made up of the various partial pressures) has LESS water moisture in the vapor form that the cylinder at the lower pressure.  Think of the system as acting as a sponge.  The more pressure, the less water vapor - while at the same temperature.}

[jamese]

Thanks Art.

When I google vapor pressure on the internet, I find many sources that state that at it is only temperature that effects vapor pressure.

How does this fit with the two cylinder example?

[chemdoc]

at low pressures (see previous post) you may use Raoult's law,

where the partial pressure of the water
PP(water) = x(water)*Pvap(water)

as result you get something like

 

Attached Files

•  waterincompressedair.gif   52.24KB   8 downloads

[jamese]

Thanks for the graph and explanation of Raoult's law, I will have a go and see if I can apply it.

 

So this is not true then?;

"One of the golden rules of thermodynamics is that the amount of water vapor that air and gases can hold is a function of its temperature only. In other words, one actual cubic foot of air at atmospheric pressure can hold the same amount of water vapor as one actual cubic foot of air at 100 PSIG pressure, provided that temperatures remain constant."

 

http://www.fluidener...essure-dewpoint

 

Thanks

[chemdoc]

just consider that if you compress a fluid (air + water)

more moles (of air and water vapor) will be contained in the same cubic foot

[jamese]

Thanks.

If the air and water vapor mixture is compressed and the amount of water vapor stays constant then it is only the density of the air that increases.
So I don't understand the graph you posted as it does not show a linear trend for the mass of water vapor/mass of air, as pressure varies.

What am I getting wrong?

[chemdoc]

that graph gives the maximum water content,

for that you may use Raoult's law and water vapor pressure.

[MrShorty]

I feel like adding to the confusion here.

 

So this is not true then?;

"One of the golden rules of thermodynamics is that the amount of water vapor that air and gases can hold is a function of its temperature only. In other words, one actual cubic foot of air at atmospheric pressure can hold the same amount of water vapor as one actual cubic foot of air at 100 PSIG pressure, provided that temperatures remain constant."

 

http://www.fluidener...essure-dewpoint

 

 

I probably would not call it a "golden rule" -- more of a "useful rule of thumb at low pressures to make calculations simpler."

 

Some of this will also depend on exactly which quantity you choose for measuring "water content". A mass(H2O)/mass(gas) or similar measure of concentration will tend to not be constant, as chemdoc's graph shows. If you measure water content in terms of partial pressure or in terms of mass/volume, this will tend to be more constant -- up to a point. Above a certain pressure (depending on other system parameters), other factors come into play, and water content will change.

 

Illustrated in spreadsheet. I've used the PR EOS (NIST's REFPROP program) to calculate dew points for nitrogen water. From this, we can see that up to a certain pressure (near that 100 psig that Fluid Energy mentions), indeed the partial pressure of water is approximately constant. However, as we get closer to 1000 psi, the partial pressure of water increases fairly dramatically.  waterdewpointPREOS.xlsx   19.16KB   50 downloads

[jamese]

OK thanks, your graph is very helpful in understanding.

 

I guess it is the fact that at lower pressures the system is behaving like an ideal gas which gives the straight line (constant vapor pressure) , whilst at higher system pressures real gas behavior is coming into effect?

 

I would like to create a similar graph which plotted the water vapor pressure for a MEG water mixture at 6 DegC.

 

I guess to start with I use Raoult's law which will give a constant water vapor pressure across the lower pressures, but then how do I account for the real gas behavior at higher pressures?

 

Thanks

[chemdoc]

just to clarify things, the original statement

"one actual cubic foot of air at atmospheric pressure can hold the same amount of water vapor as one actual cubic foot of air at 100 PSIG pressure, provided that temperatures remain constant"

is not correct, if you compare "one actual cubic foot of air at 100 PSIG and one actual cubic foot of air at atmospheric pressure",

in fact they can contain different amounts of water vapor (before condensation).

You are correct to say "I guess it is the fact that at lower pressures the system is behaving like an ideal gas",

in fact, for a N2 + H2O mixture, at low pressures and near ambient temperature, you can approximate

fugacity of water with vapor pressure correlation and fugacity of N2 with system pressure.

However at high pressures compressibility of water vapor and N2 are very different from 1 (ideal gas approximation) and you would need a EOS or equivalent system.

It is easy to calculate dew point from vapor pressure correlation when you have only one component in liquid phase (as water) and can assume other is incondensable (as N2 at ambient temperature),

for MEG + Water etc. you may create a simple procedure based on vapor pressures correlations or use a software (for dew points, phase envelopes etc. in Excel I have Prode Properties but for such problems at low pressures different solutions should give equivalent results if you can assume that vapor pressures correlations are suitable for liquid activity, differently you should use specific models).

Edited by chemdoc, 25 March 2014 - 03:32 AM.

[MrShorty]

There are entire texts and courses dedicated to vapor-liquid equilibrium calculations, so I think it is somewhat impractical to expect us to recreate that kind of discussion here. If you need detailed help, I would suggest you pull up your favorite phase equilibrium textbook and review the computations and the theory behind the computations.

 

In short, all VLE calculations start with the basic idea of fi(vapor)=fi(liquid) where fi is the fugacity of component i in each phase. VLE calculations like this are as "simple" as that. The complexity is in deciding how to calculate fugacity in each phase.

 

When we feel like we can assume ideal gas, we can subsitute partial pressure (Pi) for fi(vapor). At higher pressures, we don't want to make that assumption, so, as chemdoc indicated, we use an equation of state to calculate fi(vapor) instead of assuming fi=Pi. Are you familiar enough with EOS's and their solution?

[jamese]

OK it just got a lot more complicated!

The reason for this query is that I will be involved in the dewatering of a gas export pipeline, using a pig train propelled by export gas, operating at export pressure. 
The water ahead of the pig train is separated from the export gas by several pigs, in between the pigs there will be slugs of neat MEG.

During the dewatering operation a trailing film of fluid is left behind the pigs, this means that the slugs of neat MEG dilute with water as they travel down the pipeline.
At the end of the operation there will be a film of MEG/water left on the pipewall, this film of MEG/water is required to not wet the export gas beyond its normal wetness. 

(Simple dilution calculations will give the concentration of MEG/water left on the pipewall, dependent on the trailing film thickness and the sizing of the MEG slugs chosen) (residual MEG/water film better than 98% MEG is achievable)

I know the mass per volume of water normally in the gas under transport conditions, so I was hoping to be able to derive a relationship for the MEG/water concentration whereby I could say at a particular pressure in the pipeline the remaining MEG/water film would not wet the gas.

However, based on your good advice it looks as though this is not going to be easily calculable (by me) and I am going to have to seek assistance from someone specialized in this, that is unless I can make some conservative assumptions?

Thanks

[chemdoc]

a problem with glycol swabbing is the common assumption of a uniform film over the length of the pipeline,
a specialist would consider several factors and produce a more accurate analysis (consider if that may be useful in your case).

Phase equilibria for HC + Water + MEG can be solved with a specific software (see my previous post)

but is only part of the problem.

[RoyenG]

Hi,

 

I see a mixed between water content in the hydrocarbon gas phase where MEG is injected to reduce the dew point and water content in the air as for the instrument air as a source of the complication.

 

The MEG injected to the gas will be depend on the supplier info, but for estimation, using simulator like Hysys is acceptable.

 

I have estimate MEG required for the gas injection using Hysys to reduce the dew point. 

 

Regards,

Royen

Edited by RoyenG, 27 March 2014 - 03:51 AM.

[jamese]

Hi,

 

I have good control of the pigging aspects, it is the VLE part I am struggling with. 

As you suggest I will have to get some assistance from someone who can run a software model.

 

We are generally able to conservatively assume 0,1mm film thickness for an uncoated steel pipeline and 0,05mm for an internally coated pipeline.

Analysis of final concentration of MEG/water mixture in the final slug of MEG received at the end of the pipeline during lots of previous operations show this to be generally correct.

 

We may also have to manage deadleg pipework, where there might be trapped water that can drop out behind the pig train, but this can be managed with a slug of nitrogen gas between pigs, or by stopping the pig train and injecting MEG through dead leg pipework into MEG slug volume between pigs,

 

Thanks

[chemdoc]

a specialist would consider in addition to multiphase equilibria factors as heat transfer, superficial velocities etc.

 

if you can estimate compositions multiphase equilibria for HC + Water + MEG

can be solved with a software such as Prode Properties

(there is a free version available) or a equivalent system,

make sure to select a suitable EOS for Water-MEG (polar),

perhaps the extended versions with temperature dependent parameters should give good results,

as alternative consider CPA or EOS with complex rules,

if you are not familiar with these models better to ask a specialist.