Liquid CO2 storage 5

[irishal]

I store liquid CO2 at 20 Barg on my site. This is a new installation designed and constructed to European standards. The vessel is rated to 22 Barg and -40oC to +45oC. We have pressure relief valves set at 22 Barg.

My question is:- Is -40oC an acceptable design temperature?

There is old UK HSE guidance saying -50oC but this has been revoked. The new guidance says it should be designed to a recognised code - a great help!

Can someone send me a chart showing CO2 temperature vs pressure?

 

[Montemayor]

Irishal:

A -40 ^oC design temperature is appropriate for your liquid CO₂ storage tank if the design steel is rated and approved for that working temperature in accordance with the stress required.

You should have all the documentation for your tank and studying it you should be able to find the applicable design calculations and/or specifications. I’m in the USA and have always built my vessels in accordance with ASME Section VIII, so I’m not familiar with European or UK codes. I do know that the steel I use for liquid CO₂ tanks is only good down to -20 ^oF (-28.9 ^oC). I suspect that this is also the case in the UK.

I kind of suspect that the steel grades in both the USA and the UK follow each other very closely in acceptable stress and temperature values and that you will not be able to get to a -40 ^oC design without going into special low temperature steel grades.

The chart you request for saturated liquid CO₂ can be found at the following webpage within the NIST (National Institute of Standards and Technology) website:

http://webbook.nist.gov/cgi/fluid.c...&HUnit=kcal/g&WUnit=m/s&VisUnit=cP&STUnit=N/m

Note that the NIST data tells you the corresponding absolute pressure for your tank at the corresponding temperature. If your tank is designed for an operating pressure of 20 barg, then the corresponding design temperature is approximately -18 ^oC --- which is far different from what you are seeking (-40 ^oC). Liquid CO₂ as stored by you is in the saturated state and the pressure and temperature are related to each other as shown in the NIST table I refer you to. That means that if you intend to operate the tank at 20 barg, then for all practical reasons you have fixed the operating temperature at -18 ^oC ---which not too coincindentally comes out to be almost the same as to what we do here in the USA.

I hope this information helps out. I took the liberty of selecting what I thought would be your preferred SI units for the Thermodynamic values given in the NIST webpage. You can always change these to other units if you so desire by simply going to the NIST homepage. Also note that although this may not be of any interest to you, the NIST Thermo data does not go into the solid (Dry Ice) CO₂ phase. If you later want this data you must go elsewhere to obtain it for now.

 

[irishal]

Thanks Montemayor. The steel used is ASTM A 737 Gr. B and the documentation I have shows it has been successfully impact tested at -40oC which tallies with the design rating.

Unfortunately I don't have the design specification showing how the temperature in the vessel is maintained above -40oC in the event of a release of the CO2 (we use it for fire protection). There is a small evaporator unit controlling the vapour pressure between 19 and 20 Barg. Presumably (here's my limited physics showing) on release of CO2 the vapour space pressure (and temp) will drop, the liquid CO2 (still at -20oC) will begin to flash off, and try to maintain the vapour space pressure. The only thing I can think of is the the discharge connection is sized to limit the release of the CO2 to allow this equilibrium to be maintained without the temperature dropping below -40oC.

Does any of that make sense??

 

[Montemayor]

Irishal:

I’m glad to see that you don’t have “limited physics” background. You’ve done very good in your analysis of the process (I presume that you are maintaining the liquid CO₂ stored as a source for what you are ultimately consuming – or require: gaseous CO₂). You are right on top of one of the important points regarding this application: the Charpy impact test at the design temperature. This is very important to take into consideration in this application and I’ll explain in more detail below.

Allow me to review how Carbon Dioxide is used industrially. Most liquid CO₂ storage is not done for the sake of employing the fluid as a saturated liquid. Most installations are done with liquid CO₂ stored in an insulated pressure vessel but with the CO₂ used downstream as a GASEOUS fluid. Using liquid CO₂ as a liquid downstream of liquid storage is simple, direct, and devoid of a lot of hazards. The liquid is either pumped out of the tank or is transported downstream using the tank’s inherent vapor pressure as the driving force. This application can be a lot of fun or simply boring.

However, when the ultimate use of the CO₂ is as a gas, one has to confront operational control problems and potential hazards. If one simply tries to exploit the gaseous phase existing in the vapor space of the tank for process purposes, what will happen is that the gas phase inventory may be depleted faster than tank liquid can vaporize and replace the gas removed. This is what would happen in most applications because of the need to have the tank insulated and as efficient as can be. The amount of heat required to continuously vaporize liquid CO₂ is such that external, ambient heat cannot be transmitted through the insulated tank walls fast enough to makeup for the gas removed. This results in a decrease of tank vapor space pressure – sometimes down to ridiculously very low pressures that approach 1 barg. At these very low pressures, the corresponding saturated temperature of the tank’s liquid would be approaching -73 ^oC --- a temperature so low that ordinary steels are imperiled and weakened by it to the point of potential failure. These are not rare occurrences, I’m afraid. All too often I’ve witnessed situations where a PSV on a liquid CO₂ tank has stuck open upon relieving (mostly because solid carbon dioxide snow -Dry Ice - formed under the seat) and the pressure in the vessel ultimately decayed down to atmospheric. Under this situation the ultimate temperature reached by the tank is -78 ^oC --- a temperature that often means death to the structural integrity of a carbon steel tank. To make matters worse, after this scenario occurs the tank is allowed to “warm up” and is later put back into service – without anyone being the wiser! At this point one has a seriously and potentially damaged tank that could rupture (or “fracture”) due to stress failure, and no one would ever know that it had gone through a previous severe temperature cycle. What is even worse is that some owners may have the option to sell the tank for future CO₂ service by others and the next user would never know or be able to prove that the tank was defective after going through a possible catastrophic service failure. This is a very bad scenario and that is why I appreciate your concern regarding the control of the vapor pressure inside the tank. I have seen CO₂ tanks equipped with rupture disks (Wrong! Wrong!) that have ruptured and caused the liquid contents to revert to solid Dry Ice inside (@ -78 ^oC). I immediately condemned the tank when it failed the Charpy impact test and personally burned a 24” hole on its shell to clearly put it “out of its misery” --- not allowing it to be used again as a pressure vessel.

The way that the vapor pressure in the tank is kept constant while saturated vapor is withdrawn from the top of the tank is that a liquid vaporizer is employed to generate and replace the withdrawn vapors. The vaporization is usually effected by the use of an immersed electric heater that responds to the action of a simple pressure switch or transmitter that is mounted on the top of the tank. When the pressure decays, the electric heater is activated; when the design pressure is reached, the heater is shut off.

Of course you can also employ ambient air vaporizers but these will be much larger, occupy space, generate a lot of water ice and water drippings. They also are not as responsive to pressure requirements as an electric unit.

Please pardon this lengthy reply. I want to make sure that you understand just how correct you have been in your analysis of the importance of maintaining the correct tank pressure and what can be some consequences. I compliment you on the importance you are giving this application. Others would have not bothered in spending any time on it and simply let it go. I think that you can appreciate that there is a lot of potential harm that you can do to others and yourself by not carefully analyzing the consequences of taking carbon steel below its thermal structural limits. My hat is off to you and I hope that I have helped a little bit in understanding what is happening and what can happen in this application.

 

[abeltio]

my experience with this type of tanks is in fire protection systems.
the tank is designed to -40F (or -40C... geeky trivia: both scales have the same numerical value at that temperature) as a design condition for sudden depressurization.

these type of tanks are usually insulated and have a refrigeration system that keeps the temperature around -25C were the pressure will be less than 22barg

in case the refrigeration system fails... when the temperature rises above -18C the pressure inside the tank will rise above 22barg and the safety valve will release.

in the case of a fire protection system... if the system is actuated... a sudden depressurization will "flash" liquid CO2 into vapor.
Dry ice can precipitate if the pressure falls below the triple point.

http://www.co2clean.com/snowform.htm

the main problem with refrigerated CO2 tanks is the exposure to sun radiation or the loss of AC power to the refrigeration system...
this creates a pressure rise and loss of CO2 thru the protection system.





saludos.
a.

 

[irishal]

Thanks for this guys. This gives me a great understanding of what I need to be on the lookout for.

 

[cryotechnic]

@montemayor
Excelent post! I never looked at it that way. In the business I'm working I don't have much to do with CO2. (at least not yet...)
But it's very interesting. When I read your post it was all clear to me and actually it's so easy.

Instead of a PSV or rupture disc, what would you suggest in this case as a safety device?

Cryotechnic

"Math is the ruler of your potential succes...."

 

[Ashereng]

montemayor,

Would a 1/4 turn full port ball used in a HIPPS be considered an alternative to PSV?

The 1/4 turn FP ball would not be prone to dry ice gunking it up, the seals would be fairly good (316SS?), and the valve can be reclosed after the pressure is relieved.

 

[Montemayor]

Cryotechnic:

I’m glad to find out that I’ve succeeded in getting attention to an important safety point – on what some engineers would consider a “benign” and simple fluid that is classified as a commodity item in the industrial markets. I certainly don’t want to come across as an alarmist, but you have undoubtedly recognized what I have been trying to expound: this typically common situation has the potential to become a “stealth” type of future hazard with catastrophic results.

Fortunately I have not identified any other exact similar “stealth” potential hazards in my 46 years of engineering. There are some close to this – but not as markedly dangerous: the gaseous venting of liquid ammonia tanks down to atmospheric (resulting in -28 ^oF); the similar venting of liquid Propane tanks down to atmospheric (resulting in -40 ^oF). CO₂ is the worse actor in this respect because its triple point is so high. It is the solids formation that creates the havoc. What has been done in the past to combat this potential danger is:

1) Try to avoid a potential 2- or 3-phase flow through the PSV. This means you should position the PSV as high as you possibly can from the liquid level surface. This means more capital investment per tank because you are faced with reducing the filling density of the tank (normally 85% -i.e., 85% of the internal volume is maximum product fill). Lowering the tanks maximum liquid level gives more vapor disengagement space and ensures a purer vapor relief through a PSV.

2) Mechanically design the storage tank to a higher MAWP (I’ve used 350 psig as design) pressure and operate it at a lower pressure (I always try to operate at around 235 psig (which sets the tank at approx. -11.5 ^oF). This allows you to set the PSV at 350 psig and gives you a 115 psi cushion before having to relieve. Of course, like any other option, the tradeoff is more capital money.

Either way you go, it’s going to cost you money – in my experience – to ensure that you are safe. I wish I could offer you a safe solution without any tradeoffs. I’ve never found one.

Ashereng:

Yes, I believe you can use a HIPPS solution approach. However, I believe you will wind up with the same hard decision: you have to ensure that the vapor you are relieving through the ball valve is truly pure vapor. That means you have to locate the valve high enough over the tank liquid that you wind up with the tradeoff I described previously.

The seats in a conventional ball valve are not made of stainless steel; rather, they are, like most ball valves, made of a plastic or a “soft” seat elastomer. That’s what gives the ball design a “leak-proof” feature. These seats are prone to heavy damage if a tough and erosive solid like Dry Ice snow gets introduced into the valve - resulting in leaks. We tried and tried a variety of ball valves at Liquid Carbonic in the early 1960’s when they first came out and while I worked there. The seats were chewed up and eroded by the solid Dry Ice snow everytime we tried to apply them in that service. They worked well in pure Liquid CO₂ and pure CO₂ vapor. It’s the solid CO₂ snow that lends havoc.

 

[Ashereng]

Montemayor,

Yes, I was thinking of SS seats (see posting), or hardened, for the dry ice.

Also, would you be able to heat trace/jacket the valve? Would that not vapourise the dry ice back into CO2 gas? You can start the heating on pressure rise, befor you have to vent, and shut it down after depressurisation?

 

[Montemayor]

Ashereng:

Yes, there should theoretically be no problem in heat tracing a PSV and accomplishing a complete sublimation of the troublesome Dry Ice plugging it - that is, if you can supply the heat tracing fast enough to heat the PSV, the Dry Ice, and the cold CO2 gas exiting through the PSV at the same time. I believe that in the actual, real practice you will find this very difficult to accomplish within the heating and stress constraints presented by the PSV itself and also because of the very cold "heat sink" present in the flowing cold CO2 vapors leaking out through the PSV. The available heating surface offered by a PSV is very limited and is the controlling heat transfer factor.

I still maintain that you should be giving ample vapor disengagement space above the operating liquid CO2 level in order to minimize the amount of liquid CO2 that is entrained up and through the PSV - where it forms the Dry Ice. The PSV is plugged with Dry Ice because liquid CO2 droplets or particles are entrained into the PSV and forced to flash into the corresponding 45% solid + 55% gas mixture through the seat of the PSV. If you ensure that liquid does not reach the PSV, you should have no problems.

 

[cryotechnic]

@Montemayor,
You are suggesting to design the tank to an higher MAWP. I was thinking today and thought, what would be more expensive?
1. Design the tank to a high MAWP.
2. Design the tank to a temperature of -78dgr?

I mean, designing the tank to a temperature of -78dgr, would be the best option I think. In that case the risks are much smaller.
But maybe that's the most expensive option.
Personally I'm not in the business of designing and cost calculations, so I don't know. I was just thinking about it.

I have seen CO2 tanks with a ref unit too. (Like Abletio was mentioning) That was a CO2 tank in a CO2/NH3 cascade refigaration unit.
I allways thought that had something to do with controlling the pressure of the CO2, but never realized the complete issue.
So acctually the ref. unit is to bring down the risk of a high pressure in the tank which will result in opening the PSV of the tank.

Take care,
Cryotechnic.

"Math is the ruler of your potential succes...."

 

[psafety]

I work as a engineering contractor for a very large consumer products manufacturing company. Several years ago a carbon steel vessel failed (min temp -40degF), the root cause a bad weld. But during the relief (thru the failure) caused ice and somewhere close to -70degF. The vessel disintegrated on one end due to relief of gas flipping it from its cradle and the end fractured. The vessel torpedoed thru one corner of the plant. At least one employee was killed. They now use no less than 304 SS with a min temp below that of CO2. Yes, more stringent that "general industry" practice.

 

[irishal]

Since posting the original topic I got the system supplier carry out calculations to prove that the vessel design conditions were appropriate. Given the CO2 discharge rates they were able to demonstrate that the temperature does not drop more than 2oC so giving me an overall final temperature of -22/-23oC. Losing the CO2 from a stuck open safety relief is a diffent matter altogher.

 

[Bala0610]

Hi

A very interesting thread....thanks to all in the chain.

Montemeyor :

Is the low temperature event irreversible? You mention "personally burned a 24” hole on its shell to clearly put it “out of its misery” --- "...is there nothing that can be done (depending on materials employed of course...) to bring the vessel back to original integrity..(Stress relief...)

Anyway how does the metal remember the event (reaching a lower than design temperature) ?

How does one ascertain the suitability of material now?

How's one to check the material's impact test response in service ? Is it visible like hardness change or something like that which can be tested insitu - ina a NDT way ?

Pardon my query - am a chem engr.

I am worried about the flare system integrity - where the name plate design temperature may have been passed by in actual relief situation (after all old estimates of low temperature may not be accurate) - some new calcs show that in certain relief scenarios the header pressure can be higher - which could be dangerous if the vessel had been through a lower than design temp before...

Thanks

 

[TD2K]

Art, your comment here:

" To make matters worse, after this scenario occurs the tank is allowed to “warm up” and is later put back into service – without anyone being the wiser! At this point one has a seriously and potentially damaged tank that could rupture (or “fracture”) due to stress failure, and no one would ever know that it had gone through a previous severe temperature cycle."

I thought that when steel warmed back up it regained its normal properties, Charpy strength, etc, there was no long term 'weakening' of the metal. I do agree if when the metal was cold it could fail in a brittle mechanism if sufficient stress was put on it and being so cold, the allowable stresses it can withstand can be very low.