Temperature of Compressed Air in Cylinder

This is Conference Room 3. We deal only in reliable assumptions, and oxymorons.

You want CR4, it's two doors down that way.← That's where they over-think everything and always have three "best" ways to fix the problem.

PWSlack's response sounds good to me.(edit)

Talking to myself:

We made an assumption that the "(in breathing apparatus)" was under water.

We don't know that's the case. Could be on a firefighter's back, too.

OP?

Yes, 40 would indicate 'on a firefighters back'. SCUBA is 60+ depending on depth and exertion.

However the phase of expansion that effects the contents is restricted to the 1st stage drop. I.e. the expansion at mask regulator doesn't effect contents temp, just hose temp.

I think there are too many heat absorption variables to rely on, or factor in, contents temperature in terms of 'refining' or 'compensating' gauge readings.

Within a few minutes of non use, the temp drop these discharge rates have caused will be in the 'unreadable', and/or outside the tolerance and/or resolution of anything but a very high precision gauge.

Assume that, because of the slow rate of discharge of the stored air and because of the high heat transfer capability of the surrounding water in contact with all exterior parts of the equipment including the cylinder, the temperature everywhere is the same. Is the assumption valid?

You are actually dealing with several meduims, Air (N2, O2, He) and H2O. This constituents behave differently with pressure and temperature conditions as they have distinctive properties.

Ideal gas equation Pv=mRT could be used to dry air, however not for water. (see property table of H2O, 3d). Ideal gas equation for water could be applicable to be used further than superheated condition.

at Pressure 300 bars, water is compressed, you could assume to use dry air alone deducting the volume of 6 liters of water in the tank.

You see m=Pv/RT, looking alone in the equation since process is adiabatic, (no heat transfer involve) T will drop consequently as m escapes the tank.

The ideal gas law is PV=nRT with "n" being the number of moles of gas, not the mass of the gas. So I believe a major part of your problem is that you are not applying the correct formula. Now you're absolutely correct that heat transfer is an added complication here. However, heat transfer will usually be a much slower process than the release of a pressurized gas.

Thanks for the feed-back.

It is breathing apparatus for surface use. Firefighting yes, but for any activity requiring respiratory protection - chemical spillage - or looming CBRN incidents - particularly in very low temperature ambients.

The water vapour content is 35mg/m3 when measured at atmospheric pressure which give a dewpoint of +12C at 300 bara - means the cylinder will have 63 mg of vapour in it, thus probably of little consequence in terms of heat content, but very important in terms of enough ice forming (anywhere in the system) as the temperature drops. Hence the reason for estimating the temperature - from which we can decide if a water vapour pressure dewpoint of +12C is low enough.

The air cools as it leaves the cylinder, assumed to follow the Joule Thomson effect - which can be calculated from the pressure drop. But the vapour remaining in the cylinder also has to be considered because it could condense as the internal temperature falls. It is not likely to condense on the cylinder walls because the wall temperature will be relatively warmer, but it would form a fog where the droplets settle in the valve possibly to freeze and maybe block or jam the mechanism.

Some assumptions have to be made about the temperature drop - and I wonder what they are. Logically I guess the ones to make are those likely to cause the maximum temperature drop. Any ideas please.

Have you read this?

But as said, any problems will be at the first stage

If the design is the same as SCUBA, icing is more likely to hold things open than shut

Thanks 34point5

I have read the article now. Very good. Dealing in F and psi was like a trip down Memory Lane. The bit about temperature was more to do with problem of degrees F or C versus K.

The influence of the temperature of ambient air is taken account of before the cylinder is used. Once in equilibrium with ambient, it is the temperature drop due to usage that has to be estimated. Whether the valve sticks open or shut, either way has to be avoided.

The temperature and pressure at the point of use are easy to establish. It is the drop in temperature and pressure during use that matters. Pressure is easy to read from the gauge, but the internal temperature air is largely guesswork. But it has to be known in order to establish the dryness required.

In terms of water vapour the drop in temperature is the dominant factor where condensation is concerned. The accompanying fall in pressure tends to cancel out the condensation because the dewpoint rises. Also very low temperatures that the expanding air falls to will be limited by the temperature difference increasing the heat flow into the air from the surrounding cylinder and pipe walls.

This in itself can generate a problem because the resulting drop in temperature of the metalwork can cause external moisture to condense and freeze in the valve mechanism.

No offense intended but you may be "over training" on this.

Compressed breathing air aught to be largely oil and water free.

The only reason I can think of for doing such a study is the filtration/separation standards have been loosened due to corrosion resistant alloy tanks?

Is this the case?

I.e. please explain what exactly is the aim/idea is here.

E.g. did you get a bad fill? Or did you, or someone, have a reg failure? Or are you trying to set an Industry Standard?

Or perhaps have found a problem with such as Australian Standard 2299.1, 3.13 Breathing Gas Quality or the Britland clone DVIS9?

To 34point5 post 10

I'm acting as an adviser to a company making compressed air breathing purifiers. They in turn input to their trade association, that in turn input to BSI for UK standards, who in turn input to CEN for European standards and ISO for international standards. The latter two are both working on standards that include specifications for maximum limits of toxic contaminants in breathing air including water (although not toxic plays havoc due to corrosion and freezing).

You seem well informed. I have a copy of DVIS9, but not As 2299.1, 3.13. Do you have details you could post here please?

For high pressure, water is currently proposed as 50mg/m3 at 200 bar, and 35mg/m3 for over 200 bar (usually meaning 300 bar). But doubt has crept in that 35mg/m3 is not low enough at 300bar because the pressure dewpoint is +12C. Water will condense if the ambient drops below 12C and then freeze if below 0C. Therefore the limit should be lowered - but to what ?.

Canada calls for 27ppm for 310bar which is 20mg/m3 (at 20C) with a pressure dewpoint of -12C, with an additional requirement for the dewpoint to be 5C below the lowest ambient. Thus -12C is only good for use in an ambient of -7C.

Personally, I feel we are looking for solutioin to a problem that is not there, but an engineering explanation is required if further drying is considered not necessary.

They prefer you purchase the Standards (which you can on-line) and CR4 tends to be a skittish about copyright infringement.

Wal makes a good point - how about you do some tests in a freezer

In the later posts; concern about "zero CO" is going to be A. difficult in in-field recharging equipment limitations. Also it's present in normal expiration, so not that bigger deal in trace levels.

Fire makes a fair bit too, as does smoking and vehicles and cites in general.

To Phph; the 40l number is a fair duration based on infield safe usage. Fire has a range of effects of people, so some will use less air others 'up to that'. The idea is not to run out mid incident.

34point5:

I agree that a specification for 0ppm is not a good idea. 'Zero' is a concept as is 'pure'. We all know what it means but measuring will be difficult. Someone will always want to prove the air is contaminated by coming up with a gadget to find 1 molecule. And why not the've built the Large Hadron Collider to find half a particle.

5ppm is a sensible number because it is high enough to be in reach of many low cost detectors (and purifiers) but low enough to give early warning of fault conditions with the compressor, or of a contaminated ambient. It is a limit that has been in operation for more than 30 years in the UK and Canada. Strictly speaking the UK limit is currently 3ppm. It stems from 10% of the 8hr time weighted average of the Workplace Exposure Limit. It used to be 50ppm (hence the 5ppm) but has since been reduced to 30ppm (hence 3ppm).

I do not have facilities to conduct proper tests. But a simple test can be conducted to prove the cooling effect on the air inside a bottle. Take an clear glass bottle, make sure it is dry inside (no visible signs of any liquid water). Place it inside a fridge. When cold you will see a film of condensation/ice over the whole inside wall of the bottle (and over the outside when room water vapour condenses on it when you take it out of the fridge).

If analogous to a compressed air cylinder, the excess water vapour on the walls is not in the air itself and thus unavailable to block anything as the air is let out. But it will evaporate into the air as the pressure falls. Then it will be available to condense at the lower temperature as the air expands.

I am working on the idea that the amount of water available to condense is due to the change in temperature/pressure. Thus 63 mg (as vapour) at 12C and 300 bara (if completely condensed as a liquid in a single droplet) would have a diameter of 4.94mm and could easily block the orifice in a valve.

In reality water cannot condense from saturated vapour into one single droplet. It will condense as fog typically in the range of 1 to 14um diameter. I suppose they could block the pre-filter in cylinder, but normally these are coarse filters for big particles of muck and dirt. Unless water was in the cylinder in the first place. In which case the original standard has not been met.

The original question was whether 12C is low enough for ambients of less than zero C. On paper it is not. Canada works on 27ppm (assuming 20C reference) which is 20mg/m3 or +6C pdp or -11C depending on which source reference tables are used.

Thanks for all the help.

"But it will evaporate into the air as the pressure falls." ?

"Then it will be available to condense at the lower temperature as the air expands."

The second sentence is very much dependent on the first - to which I have added a question mark.

My inclination is to first find out if this is so.

So far as I know there are no problems with % water (within Specs) in SCUBA in sub-freezing arctic water - but BA is worn inverted so it might be a problem.

I can't actually see a reason for condensate to become available (to a getting colder gas) once adhered to a below freezing body. But perhaps it sublimes? Even so I would think the RH ^{_{(relative humidity)}} would be fractional of potential % for an above freezing tank. But say it heats up at the fire ...

So, if one assumes ice particles suspended in the air - I would also be looking at the relative energy of the gas flow in terms of the pressure gradient across the valve seat. I.e. is it possible for ice particle buildup to physically withstand the pressure difference?

Are you looking at ice granules, crystals or micro-snow, or even formation of clathrate hydrates?

Were any of this to happen, what practice effect would it have? Hence my 'jamb open' comment earlier. I imagine this produces excessive 'run-on' in the second stage - which is highly noticeable and a signal to get out before it runs out.

It does seem fairly easy to find out if there is a problem using a deep freeze cold room and venting a range of % RH fills at a range of flow rates.

If it does 'fail' at Standard, or doesn't; or you can tune everything until you identify the threshold. Then you have establish a practical benchmark that encompasses all the gas and hardware parameters. This then can become the Industry Test Standard for all equipment.

Of course in an ideal world, your gear sets the Standard no one else can meet.

Thanks to all for the time and effort you have spent on my behalf, but replying specifically to 34point5 - post 26.

The very low levels of water vapour required by (some) current standards - 35mg/m3 - for air taken from a 300 bar cylinder when measured at atmospheric pressure - will have a pressure dewpoint of +12C when in the cylinder. Although the +12C is a figure taken from early 'steam' tables - that themselves have been up-dated. Finding the right one is an additional problem. At the moment I am comparing 'differences' to see if they matter, ie, significant.

Then I will estimate the lowest possible temperature using a formula based on assumptions to achieve this, then to assess the effect on the water vapour, and if it really matters, then to assess the assumptions (use a different formula) as to whether they are likely or not, and if it matters, to reflect them back into the equations, perhaps on an endless loop until a sensible answer is found.

Hopefully to produce a standard that we can all meet.

My general experience is that water in breathing air being a natural consequence of air compression, is not filtered/dried properly, to the extent that liquid water sloshes around everywhere in the system. The problem with high pressure air is that liquid water goes undetected until it builds up to dangerous levels, thus already in the valve mechanism ready to freeze when the temperature falls.

Making sure this water cannot occur will be met by applying the current standard. Which in 'theory' is applied but because of the complexity of air quality testing (lack of understanding) it is possible for a test result to show that the air is 'dry' when in fact the cylinder is soaking wet.

As an example, with two cylinders both exactly at the same pressure and temperature, where one has been proved to be perfectly dry whereas the other has 1 litre of water in it. The ambient temperature is such that the 'dry' cylinder is precisely at the dewpoint where it is fully saturated, as will be the air in the 'wet' cylinder (any excess water having condensed. A sample of air taken from each cylinder and measured at atmospheric will give exactly the same result.

So which cylinder has 1 litre of water in it? Forget about shaking the cylinder to listen for water sloshing around (the good old fashioned test), because as often as not, the moisture test is likely to be part of a more sophisticated test for all contaminants where a sample is sent away for analysis.

A 'pass' certificate is issued, and you can carry on using cylinders full of water in the belief you meet the standard. Thus if something goes wrong it is the standard that is brought into question - not the procedure.

I think the better 'test' is venting the cylinder inverted before refilling (SOP) ^{_{(particularly important in the days before alloy tanks)}}

But as BA cylinders are generally worn inverted and I'm unaware of any form of 'draw tube' in these cylinder fittings, I'm not sure how you'd get a build up of liquid without then getting a wet mask/face.

I.e. if a Standard deals with the whole system, it has to have whatever SOP's that are required to maintain the system to standard, written in. Just like changing filters or whatever.

To 34point5

Yes,the test procedures needs to be tightened up but first to establish a standard.

The example I gave for the quality test was to demonstrate circumstances where a 'pass' certificate could be obtained for a cylinder full of water. Assume in practice that a cylinder could be filled via a filtering system that has failed to perform thus allowing excess water vapour to go into the cylinder, but not enough to become obvious by shaking or inverting the cylinder and opening the valve to drain off the water.

A 'pass' cert is evidence to the air filling station that the (faulty) filtering system is working. So the cylinder goes into service in the belief that everything is fine.

A subsequent 'freezing' failure (if it occurred and if investigated), when backed up by a 'pass' cert, would lead to the conclusion that 12C not being dry enough.

Whereas the real problem is not the vapour condensing and freezing from a 12C dewpoint, but excess water (that should not be there in the first place) freezing up due to the drop in temperature caused by expansion.

But in this day and age it is so easy to show that the air temperature could drop well below freezing, so a pressure dewpoint of 12C is not low enough. Just what it should be is another problem, and the one that I hope to answer.

I'd look at it more along the lines of if a certified filling station manages to put a litre of water in a BA cylinder and no one puts a size 10 fire boot up their rear, then what ever number you choose is academic anyway.

12C at 300 bar? are you sure about that?

at 1.031bara saturation temp would be 100C. any temperature lesser than 100C at 1.031 bara water will start to condensate.

Psat is directly proportional to Tsat.

Consider 300bara pressure T sat would be >100C, therefore any exposure of H2O at temperature lower than Tsat and @ Psat H2O will condensate.

Certainly at 300 bara pressure and Temp lower than 100C, you'll find water as compressed inside tank. There will be no vapor in the air. Mixture of both air and water is impossible at that condition.

By this you already messed with your problem, 63mg of vapour in air is false as you describe here.

I presume these units are some sort of escape packs to be used in case of emergencies. The manufacturer will probably have a complete set of instructions for how to charge them - unless you are the manufacturer?

If you are not, then it sounds to me like you are wasting a lot of time and effort. Why not test a couple of units under real life conditions with all the parameters known, i.e. ambient temperature, dewpoint and flowrates and then see if you can provoke ice formation at the reduction valves....but I seriously doubt you would be able to get all 63mg of water vapour to settle out in the valve even under much higher flowrates.

Seeing as the human at rest breathes a nominal 5 l/min, and 15 l/min would count as strenuous exercise, and seeing that it is a demand valve, where does the 40 l/min come from? In my early anaesthetic days, when continuous flows were in use of 3l/min of oxygen and 6 of nitrous oxide, serious cooling of the nitrous regulator was a problem, to the extent that a finned design was used. The oxygen regulator was not a problem, nor did the cylinder temperature appear to differ noticeably from room temperature. I suggest that you carry on calculating the contents from the pressure as before, as the experimental error will be small. From the safety point of view, of course, you never aim to use all the calculated contents anyway.

To phph001post 11

40 l/m is the average flow used to estimate the time a full cylinder of air will last.

It is a rule-of-thumb but proven to be reliable in practice. It comes from the time a BA set will last during extensive user trials, based on a range of hard-work and at-rest activiies, as part of the testing procedure specified in the standards.

I can't vouch for your figures but they seem very low to me.

Well, I did teach physiology for many years, but if Wikipedia puts the figure at 5-8 l/min, I'm not going to argue. I will also assert that the 5-fold increase represented by 40 l/min represents a level of sustained exercise which even trained athletes will find hard to keep up. Do you have a reference to back the 40 figure up?

I think I have sorted out where the 40 l/min comes from. It is not "the average flow used to estimate the time a full cylinder of air will last" but the average or, possibly peak, flow during inspiration. It is a requirement of the demand valve that it delivers this flow without an undue respiratory resistance, and it does not represent the actual gas delivery each minute. Humans spend 1/3 of the time breathing in and 2/3 of the time breathing out, so as a first approximation that 40l/min inspiratory is equivalent to 13 l/min overall. That is an expected value for mild exercise.
Returning to the original question, it is not an appropriate value to include in any calculation, unless the time cycle of respiration is also in the calculation. I remain convinced that the temperature of the cylinder will differ only trivially from the ambient, and that any temperature effect on the regulator will also be relatively small. The best thing to do is to measure it.

To phph001 post 13 & 19

The average flow rate used for estimating duration is 40 l/m - and that is on the low side. Pick any respirator testing web site for additional info. eg:

http://www.ncbi.nlm.nih.gov/pubmed/16857648

The working flow rate I use for assessing pressure and temperature is 40 l/m because my audience will accept it at face value.

I will calculate the max temp drop for both the air remaining in the cylinder and for the air leaving the valve by assuming adiabatic free air expansion, then working back to isothermal to give a band of temperatures that the air might cool to.

But from a physiology position you might be more interested in the carbon monoxide [CO] part of the standard. It was 15ppm, then dropped to 10ppm, but in opposition to those wanting 5ppm. Quite frankly I think CO should be zero. There is none in the atmosphere itself, it is only generated locally in isolated areas at ground level. Compressors do not produce CO except in fault conditions at high temperature. The UK and Canada, and parts of the USA are 5ppm.

We are not talking of how much CO we can tolerate, but what is suitable for a standard for filling cylinders to provide life support to persons working in lethal atmospheres. They should be given the best quality air available, not some cheap air for the convenience of accountants and managers.

With respect, I can see that there is going to be no convincing you unless you measure your own minute volume to get an idea of the normal respiratory parameters, which is what I have been measuring all my professional life. The reference you quote gives data pertaining to subjects who were being tested to exhaustion. That is relevant to considering the inspiratory resistance at high gas flows, but not to the expected emptying rate of a cylinder. I appreciate that your emergency workers wearing rescue apparatus are taking exercise, but you would have to explain how they take such strenuous exercise for continuous periods long enough to empty cylinders at that rate. To be specific, what is the size of your cylinder, and when during a rescue exercise do you actually have to change the cylinder?
I appreciate your concerns with CO, and indeed 0 ppm would be desirable. However, 5 ppm is an acceptable domestic level, see
http://www.epa.gov/iaq/co.html

If you want better than that you are going to have to use a very well maintained (and well cooled) compressor or buy medical air cylinders.

Horace40

1. Nothing wrong with your calculation. The pressure 293.33 bara agrees with my calculation.

2. If you want to be more accurate (which is not necessary) you can use a more complex Equation of State (EOS) such as Peng-Robinson and consider the effect of water vapor.

3. From 1st Law of Thermodynamic, we have dQ = dU = CvdT, for a closed system.

4. Therefore, if dQ = 0, this implies that dT = 0. The assumption that temperature remains constant is correct. This is true if the ambient temperature is the same as the air temperature inside the cylinder.

5. Temperature of air inside the cylinder remains the same, the temperature of air outlet of valve is lower due to adiabatic expansion from high pressure to low pressure. That is, the change in temperature is not for the air inside the cylinder. Therefore, if you measure the temperature outlet of valve, then the temperature is expected to be lower due to adiabatic expansion

Have you considered determining/quantifying the effect by controlled experimentation?

Suck it and see.

Some thoughts on this thread:

The formula PV=nRT is useful in a situation where all but one of the variable is known, and the final variable is desired. Pressure, Volume, number of moles, R(gas constant: dependent on system choice), Temperature. It does not apply where two or more of the variables are unknown. When gas in compressed, as in an air compressor, or expanded, as when gas is taken out of a storage cylinder, work is done, both on the gas being transferred into or out of the cylinder, and on the gas inside the cylinder. This energy must be accounted for in the final energy evaluation. Mechanical energy transfers are rarely 100% efficient, and the energy losses (and where that lost energy winds up) must be identified to determine how much of it is lost to the gas stream. With this information, the perfect gas law can give values for a perfect gas. Since dihydrogen oxide (water) is far from a perfect gas (or liquid), we are fortunate that it has been well studied over the centuries, and experimental data is available. When water is in mixture with the other components of air, and the air is compressed, an interesting result is obtained. Perfect gasses when in a container behave as if they are each alone in the container, even when they are sharing the space. Since water is the most imperfect gas present in air at significant amounts, the water also seems to be alone in the container. This is important because the water has a known amount of gas phase mass that will exist in a container at a given temperature. If more is added, say by pumping air in, or by being filled from an air storage tank, the filled tank when it has reached ambient temperature, will still have the same amount of gas phase mass*, the rest will be condensed into liquid (or solid if ambient temperature is below freezing). This is why air compressors have a drain valve to remove water from the tank. It is also why air from a storage tank (like a SCUBA refill station) will have drier air than air directly from a tankless compressor. Freezing up of a regulator is unlikely from a tank filled from a storage tank, but if it is filled directly from a compressor as may be the case for emergency work, there may be liquid water which if mounted valve down will enter the exit stream and may freeze inside an heavily loaded regulator. These complexities make experimentally determined data valuable.

*correctable for the change in volume caused by the liquid water volume.