17 replies to this topic

[Roark]

Hello

During my first two of years of experience in the industry I have been working for a couple of engineering companies mostly on the design of LNG terminals. Traditionally I have considered that a variation in atmospheric pressure impacts the pressure in the vapour space of the tank: for example, BS EN 1473:2007 ("Installation and equiment for liquefied natural gas - Design of onshore installations") states in its section B.7 that:

"if the pressure in the tank is equal to maximum operating pressure, a drop in atmospheric pressure brings about a gaseous discharge frome xpansion of vapour in the crown plus vapour evolved from the overheat of the liquid. Similarly a vacuum condition can arise following an increase in atmospheric pressure."

I have done a few calculations based on this assumption and on a paper by Hashemi and Wesson from 1971 and they have been always accepted by my Lead Engineers in both companies.

However, I am wondering "why?". A full containment LNG tank is a massive 9% Ni steel tank surrounded by tons of insulation and a big concrete wall and a dome, so I fail to see how atmospheric pressure can affect the vapour space within such a rigid, closed system. Perhaps I am missing something very obvious because I never met anybody who challenged this "perceived wisdom", but on the other hand I've asked this same question to a couple of experienced guys in my new (and small) company and none of them were able to give me a satisfactory answer other than "it's always been done like this".

Unfortunaly as I said my new company is rather small and there isn't any guru I could ask, so I wonder if there's any experienced guy (maybe just somebody smarter than me ) who could shed some light on the issue.

Any help would be appreciated.

Regards

[ankur2061]

Roark:

Liquefied Natural Gas and Liquefied Petroleum Gas behave in similar manner inside a storage tank when there is a change in atmospheric pressure ie. the change in the vapor space pressure is directly impacted by the change in atmospheric pressure.

I am providing a link for an article which explains this phenomena very beautifully and precisely. It gives a detailed explanation on the phenomena of vapor expansion (pressure rise) and vapor condensation (vacuum formation) in a closed container due to the change in atmospheric pressure. The examples given are for Liquefied Petroleum Gas (propane) but it is also applicable to LNG (predominantly methane).

http://www.corken.co...ining/cp226.pdf



Regards,
Ankur.

[Roark]

Ankur

Thanks a lot for your link. It really is a really nice document, aiming to explain some physical phenomena in a very intuitive (and yes, beautiful) way. It has not fully helped me, though. Please find below my comments.

I understand that the figure that tries to explain better the phenomenom is the attached file internally_driven_boiling.jpg (page 5 in your document). However, there is a fundamental difference between this scenario and a full containment LNG tank: in the example in your booklet, the volume of the fluid is variable, so when the weight is removed (i.e. the pressure goes down), the vapour space, which is in contact with the atmosphere, increases.

This does not happen on an LNG tank (see attached file full_containment_tank.jpg), where the vapour space is separated from the atmosphere by a rigid, concrete dome.

No matter how hard I try I do not get it: if the volume of the vapour space within the tank is constant, how can there be any interaction whatsoever between the atmosphere and the contents of the tank?

Regards

Attached Files

•  internally_driven_boiling.JPG   33.66KB   101 downloads
•  full_containment_tank.JPG   48.53KB   80 downloads

[Roark]

Nobody's up for the challenge? : )

[Art Montemayor]

Roark:

I started out in the LNG industry in May of 1977 when I worked as a Project Manager for El Paso LNG, the World’s leading LNG company at that time and who had the Algeria LNG production to import into Cove Point Terminal in the USA. We had a fleet of nine (9) LNG Carriers, each with a net capacity of 125,000 m³ of LNG. We had to close the company because the Carter Administration in the USA would not accept the Algerian contract continuation. Now we are facing a lack of sufficient energetics. Go figure.

I think you are making a concern related to an LNG terminal phenomena termed “LNG Rollover”. I personally have never read or heard that “a vacuum condition can arise following an increase in atmospheric pressure.” The so-called “standard” you cite is unknown to me – mainly because it was written many years after we had already successfully transported, stored, vaporized, and pipelined a LOT of natural gas. I think you are mostly referring to the on-going failure to describe accurately what is the basic scope of the LNG industry and the related phenomena of rollover.

For example, go to http://www.igu.org/h...d11684.pdf and you will read about rollover and its effect.

I hope that you are a Chemical Engineer, although with a soubriquet like Roark (Roark's Formulas for Stress and Strain) you may be a Civil or Mechanical engineering colleague. In any event, the subject of vapor pressure becomes the prime actor in this explanation that I am going to attempt. Yours is not a challenge, I think, as it is a lack of understanding that what is being handled is A SATURATED LIQUID and, in some cases, A SATURATED LIQUID MIXTURE. If you read the referenced paper – and a lot of literature (including this thread) – you will find that most authors fail to even mention that the LNG industry is based 100% on storing, transporting, and handling a saturated liquid composed of what are gases at normal, atmospheric conditions. Everything on God’s Green Earth has a vapor pressure – even steel and concrete. These “tough” materials may not show it (because their vapor pressure is miniscule and un-measurable) – but they still have it. Molecules are forever trying to escape the bonds that hold them together into a liquid or solid phase. LNG is no different – especially since we only condense it to the point where it forms a saturated liquid at the prevailing atmospheric pressure where it resides. And we keep it at that condition (pressure and temperature) while it is stored. Note that THERE IS NO SUB-COOLING (SUPER-COOLED LIQUID) involved – only saturated liquid in equilibrium with its vapor that hangs above it in the vapor expansion space of a storage tank. That all means is that if there is any heat added to the LNG – such as heat or friction leaks – while the atmospheric pressure is held constant, the additional heat will vaporize some LNG because this represents latent heat of vaporization being added. This happens because the LNG is just sitting there as saturated liquid, awaiting the first amount of heat added to convert back into a saturated vapor.

The same effect can occur within the LNG storage (or transport) tank if the atmospheric pressure is reduced: the top layer of LNG will try to reach that absolute pressure level and in doing so, will start to vaporize some of its content – at first, on its top surface. This is the same effect that we have when we subject a vessel with liquid water to a vacuum condition: the water will evaporate and reach a lower temperature that corresponds to the system’s lower absolute pressure.

Rollover can occur because of several reasons. But the basic, underlying reason is that a STRATIFIED LAYER OF DIFFERENT DENSITY SATURATED LIQUID IS SUDDENLY CREATED and a “decantation” effect is initiated – the heavier liquid layer trys to precipitate to the bottom of the tank, displacing the lower, lighter layer. A sudden evolution of vaporization occurs when the warmer, lighter layer hits the top section and the amount of released SATURATED vapor can overload the tank relief devices and overpressure the vessel – a potential for disaster.

Note that the above is often described for stationary LNG RECEPTION TERMINALS – and not for transport tanks on carriers or production plant storage tanks. This is probably because it is the reception terminals that suffer the consequences of accumulating different mixtures of LNG (remember: there is NO ONE STANDARD COMPOSITION FOR LNG – all LNG is different in compositon) and this aggravates the need to keep and store a homogeneous product that responds to one set of physical and engineering criteria and conditions. Different LNGs have different saturation pressures and temperatures – albeit slight.

In all of the above, the ABSOLUTE pressure has a role to play because it is the basis of the driving force that affects vaporization or condensation and – as I keep emphasizing – it all has to do with SATURATION conditions – not super-cooled. I dare not continue on with this response because I could go on ad infinitum. There are literally hundreds – if not thousands – of articles written on this subject. I wish all these authors would base themselves on the basic scope of LNG: it is a saturated liquid stored at saturated conditions that are fixed (or modified) by the prevailing local conditions of absolute pressure and temperature. I think that is the reason why fellows like yourself and I tend to get confused with all this printed matter.

[Roark]

Art

Thanks a lot for your comprehensive answer. As usual, it's a delight to read your posts: I have enjoyed a lot reading your contributions on a few internet forums over the last couple of years and before addressing the issue on topic let me thank you for the sustantial amount of knowledge you have shared with us all.

Slightly off-topic too, I am a Chemical Engineer indeed. I am aware of that book, but in my case the origin of my nick goes back to Howard Roark, a fictional character in Ayn Rand's novel "The Fountainhead". An architect who refused to compromise. To be honest, the book is quite bad but I adored the approach of the character, despite I realise it is highly ideal and very impractical for real life, especially if you are in the engineering business. Oh well : )

Back on topic, yes, I am familiar with the concept of rollover: it is usually the sizing case for the PRVs, estimated as 100 times the normal boil-off rate of the tank due to heat leak (usually 0.05% volume%/day based on pure methane). However, I think that the crux of this issue is not related to the very nasty (and scary!) phenomenom of rollover.

I am aware that stored LNG is a cryogenic liquid at saturared conditions (i.e. at its bubble point at a certain operating pressure). As a consequence, a decrease of pressure in the vapour space of the tank will result in the vapourisation of a portion of LNG and the consequent decrease in temperature. Keywords: "a decrease of pressure in the vapour space of the tank". My question is: how can a decrease of atmospheric pressure translate into a decrease of pressure within the vapour space of the tank if it is a closed, rigid system made of concrete lined with stainless steel with no vent or any connection with the atmosphere? The operating pressure of the tank is kept slightly above atmospheric (between 100 and 250 mbarg). I wonder how an increase/decrease of pressure outside the tank (i.e. the atmosphere) causes a proportional change of pressure inside the tank.

I have thought about some examples to let me understand the issue, but I haven't quite got there. For example, imagine a car, ideally insulated from the atmosphere. The pressure inside the car is the same pressure as outside the car, i.e. atmospheric, 0 mbarg. The driver, bless him, losses control of the car and ends up 10 m under the sea. The pressure outside the car is then close to 1000 mbarg, but if we admit that the car is perfectly insulated, and considering that it can mechanically withstand 1000 mbarg of pressure difference between the outside and the inside without collapsing, what would be the pressure of the air inside the car? There is no deformation of the car (i.e. volume inside is constant), no inflow or outflow of water/air, no change in temperature... How can the pressure vary inside the car?

Again, thanks everybody for reading. Any feedback on this will be appreciated : )

Edited by Roark, 01 February 2011 - 04:19 AM.

[Art Montemayor]

Roark:

Yes, I remember seeing Gary Cooper in the film, The Fountainhead, in 1949. I later read the book in College since I double-majored in English and Chem Engineering. Very wordy. I had difficulties understanding Ayn Rand - especially pronouncing her first name.

After reading your further details, I believe I am on to what is confusing both of us. I suspect the following:

The person who wrote BS EN 1473:2007 ("Installation and equipment for liquefied natural gas - Design of onshore installations") is a mechanical engineer by training and what is being referred to is the ancient and continuing practice of MEs to resort to looking at pressure vessel design as internal gauge pressure and external gauge pressure – another thing that I have difficulties understanding. But that is how they express the pressure stresses acting on both side of a pressure vessel wall. My first mentor, a British naval engineer and an excellent mechanical engineer, once explained this practice as a general one that allows for all situations involving a pressure vessel -– for example the case where you have a vessel inside of another vessel, or the case of a submarine hull. That helped me to try to understand this method of calculating around pressures.

If we read the statement:

"if the pressure in the tank is equal to maximum operating pressure, a drop in atmospheric pressure brings about a gaseous discharge from expansion of vapour in the crown plus vapour evolved from the overheat of the liquid. Similarly a vacuum condition can arise following an increase in atmospheric pressure."

In the sense that if the atmospheric pressure increases (however slight) and the vapor pressure inside the LNG storage tank remains the same (and it will, since the temperature remains constant), then there will exist a larger pressure difference across the wall of the LNG tank and the pallet or plug of the relief valve and RELATIVE TO THE EXTERNAL PRESSURE, we have a "relative vacuum".

I realize that this is difficult to understand (& explain) – because what is not being discussed is the absolute pressures that exist in the atmosphere and inside the tank. We ChE define a partial vacuum as any pressure below atmospheric pressure (whatever the standard). Mechanical engineers like to look at a vessel as one that has pressures on both sides of its walls (which is basically the truth). When the external pressure increases while the internal pressure remains the same, they are concerned about the integrity of the walls and vessel. We ChEs like to simplify and assume that the atmospheric pressure will remain the same and consequently all we have to worry about is the decrease of the internal vessel pressure since the external is assumed to remain constant (which is basically NOT the truth).

Have I succeeded in explaining --- or have I muddled up the entire attempt? Perhaps someone like my friend and fellow Forum member, Katmar, can jump into this thread and help unwind us from this complex understanding.

[Zauberberg]

API 2000 recommends barometric pressure to be taken into account when calculating venting requirements for refrigerated storage tanks, and provides the formula for calculating volume of expanded vapor in the enclosed vapor space due to changes in barometric pressure.

We need to consider the fact that pressure in process equipment always exists as gauge pressure and therefore always dependent on ambient conditions. If barometric pressure decreases, pressure control valve on the storage tank needs to evacuate more vapors in order to maintain the system gauge pressure. And that is the answer on your query, I believe: simply consider that any decrease in barometric pressure will call for displacing larger amount of vapors from the tank, in order to maintain the set point on the PCV.

[katmar]

Art, my apologies for not replying sooner. I had not opened this thread and seen your request.

I really have nothing to add to your explanation, which is very clear and complete. I hope that Roark will come back to say that he now understands the process, or to explain where he is still struggling.

Although I have nothing to add, let me just emphasize what Art and Zaubergerg have said - the temperature/pressure relationship of the LNG involves the absolute pressure, while the mechanically based pressure controllers (or relief valves) see gauge pressure. The pressure controllers therefore control the internal pressure relative to the atmospheric pressure. When the absolute and gauge pressures diverge because the atmospheric pressure has changed then you will get either an over-pressure or an under-pressure inside the fixed volume vessel.

Edited by katmar, 04 February 2011 - 07:25 AM.

[Bong Kook Seo]

Hello

It is help for you to understand the phenomena if you think about diving into the sea. Diver usually get up to the surface slowly to prevent oxygen boil off in the blood vessel. Eventhough the body is isolated from the surround pressure change, body volume itself is changing as surround pressure change. As surround pressure is going down, blood vessel volume is increasing. If the volumn changement is very quickly, the oxygen in the blood is boil out to compensate the volumn change.

By the way, I need your help to calculate BOG rate when the barometric pressure is changed. I found the equation in BS EN 1473 but it is difficult to understand the explain in the code. Can you send me the calculation example?
my email address is bkseo1218@gmail.com

[daveyboy]

Hi guys if we just recap back to the original query

"if the pressure in the tank is equal to maximum operating pressure, a drop in atmospheric pressure brings about a gaseous discharge frome xpansion of vapour in the crown plus vapour evolved from the overheat of the liquid. Similarly a vacuum condition can arise following an increase in atmospheric pressure."

The key hear is max operating pressure i.e the differential pressure across the tank has increased to maximum then you must drop the pressure in the tank to reduce the pressure differential and If the cargo is at saturated conditions i.e -159.6 oC and 1180mbar (abs) by dropping the pressure you will reduce the saturation temperature (boiling temperature) of the liquid to say -160oC. Conversely if you have a tank of LNG at say -160oC 1080mbar saturated and the atmospheric pressure increases this has the effect of increasing the saturation temperature (increased the boiling temperature) this will change the state of the vapour and cause it to collapse back into a liquid, this change in volume (600:1) will cause a partial vacuum but only until equilibrium conditions, as the pressure falls so does the boiling temp then it will boil again until it finds its equilibrium. The best example is take a cup of water at 80oC in an elevator ascend in the elevator and watch as it starts to boil at say 20,000ft (not sure if mount Everest has elevators) and then decend and watch it stop boiling... LNG is exactly the same. I trust this helps. Dont over complicate with phenomena such as layering, stratification etc ... if the tanks were infinitely strong you could maintain the inside pressure to what you wanted.... but they aint and even a few mbars over a huge area causes serious stresses on the tank material.

[S.AHMAD]

1. Please allow me to add further confusion on .this interesting subject.
2. First of all allow me to made following statement:
a. Pressure inside an open tank changes with atmospheric pressure
b. Pressure inside an enclosed tank does not change with atmospheric pressure.
Statement a is ovious and no further explanation needed.
3. Let me explain for statement b by using example as below
4. Let say the LNG tank gauge pressure shows 0 psig and the local atmospheric pressure is 14.7 psia. This means that the absolute pressure of LNG tank is 14.7 + 0 = 14.7 psia
5. Assuming that the local atmospheric changes to 15 psia. The gauge pressure will be 14.7 - 15 = -0.3 psig. This negative pressure is by definition is a vacuum. However, the absolute pressure inside the tank remains 14.7 psia.
6. Supposing now the atmospheric pressure changes to 13.7 psia. The gauge pressure will r ead 14.7 - 13.7 = +1 psig. This positive pressure indicate an increase of tank pressure. BUT in reality, the pressure inside the tank still remains at 14.7 psia.
7. Hope the above examples give better insight of the confusion.

Edited by S.AHMAD, 18 January 2012 - 10:24 PM.

[Neelakantan]

hi all
probably this is a dead topic; but i am reading all LNG related topics in Cheresources especially where ART and Zauberbeg have some thing to say. the reason is i am now involved as a consulting engineer for a mini/satellite LNG terminal in India. Though our design is with pressurised bullets (banks are made of two or three such bullets), inplace of flatbottomed tanks as we are using a FSU for the main storage.
However, i am explaining to my younger juniors and one such point of discussion was the topic of this thread. Reading this thread brings out the various confusions that an engineer can have; the basic point here is that the pressure regulation is like a weighted valve and the tank internal pressure is a differential pressure over the atmospheric pressure. Thus any change in atmospheric pressure will change the pressure ever so slightly, but the fluid is a saturated mixture which tends to condense or evaporate as the pressure changes. the second thing to remember is the volume difference (1:620) between liquid and gas and this becomes critical when the vapour space is not large.

however, i am happy that engineers are able to question accepted things and get answers for the correctness of idea. This, i am saying, because i am trying to break the general view of all LNG engineers to have double walled vessels! I am trying to convince that the landbased storage that we are proposing is a max of 10000 m3 which will be a buffer to move the liquid to vaporisers. This storage is being built on a modular progressive manner (while the operation will start with small capacity and the storage will be augmented with incresed cash flow as against the present concept of a huge capex at the beginning itself) and so i am suggesting a PUF/PIR insulation and pressurised bullets.

one lingering worry is that after three years of operation we will be having about 20~30 pressurised bullets of 300~500 m3 capacity and will definitely require higher overall capital outlay as the surface area will be more and fire-fighting equipment will be more, BOG generation will be more etc.

regards
neelakantan

[vova]

For Ankur.With all thinkable changes in ambient pressure the methane temperature variation will  be within 1K. Therefore,heat input in tank will be practically unchanged.When atmospheric pressure rises ,pressure in tank ullage and interphase temperature rise also.The rest of liquid also begins to heat up to be around the saturation curve at its depth.This transition to new state requires, if to recall sizes  of LNG tanks,long time.During this period the BOG value decreases initially ,but then  approaches to value observed before the pressure growth.The same but in vice- versa manner will be observed with the atmospheric pressure decrease:the ullage pressure decreases,temperature decreases,BOG value increases but finally returns to its initial value,corresponding to external heat input. With best wshes from Moscow, Vladimir Voyteshonok,vvoyt@rambler.ru 

[TRM]

Roark,

I have same problem with the statement that atmospheric pressure will have direct effect on BOG in an LNG tank.

If the tank is completely self contained, there should be no effect to the BOG.

But if there are PSVs set at very low pressure, so that changes in atmospheric pressure constitutes a large percentage of its set pressure, then the PSV may open prematurely (when atmospheric pressure drops) or would not open at the set point when atmospheric pressure increases.

A low pressure PSV is simply an aluminum plate that sits on top of an opening, and set pressure is adjusted by adjusting the weight of the plate. Atmospheric pressure has a significant contribution to its set pressure.

 

A vacum safety valve is just the opposite. An increase in atmospheric pressure might force air into the tank. Or, so it seems, the tank creates a  vacuum that draws air in.

Wording of the code must be revised. They are wrong.

[PingPong]

During normal operation the pressure control system of the tank maintains a slight overpressure (say X millibar) with respect to atmospheric pressure by regulating the capacity of the BOG compressor.

 

If atmospheric pressure drops, the absolute pressure in the tank will also drop by the same amount, assuming the BOG compressor can handle the new situation, because the control system will try to maintain the overpressure of X millibar being its setpoint.

Note that the BS EN 1437 section B.7 requires that one designs for an atmospheric pressure drop of 2000 Pa/h (20 millibar/h) and a total pressure drop of 10 kPa (100 millibar). This may sound extreme but that is what safety is about: design for the worst possible scenario, even if that happens only seldom.

 

That drop in absolute pressure in the tank means not only that the volume of the vapor in the top of the tank expands, but also that the top layer of liquid is all of a sudden too warm with respect to the absolute pressure above it, and as a consequence it partly vaporizes whereby the required heat of vaporization comes from the liquid itself, being slightly too warm for the new lower absolute pressure. Whether the BOG compressor in your case can handle that extra vapor? Risk of rollover?

 

Note also that setpressure of relief devices are based on a fixed overpressure, not on a fixed absolute pressure.

 

With the above in mind, please reread all the other earlier postings once more. It has all been explained by others already in much greater detail, but maybe now it becomes more clear to you.

Edited by PingPong, 13 August 2013 - 08:36 AM.

[vova]

In my comment  dated by 02.05.2013 it was believed by default that LNG tank is under slight overpressure to ambient.Any variation in atmospheric pressure will result in variation of the tank absolute pressure with the same sign. This will give  some BOG temporal variation. Having in mind cost,sizes and complexity of LNG tanks an idea of PSV in form of aluminum flap valve( mr.Roark) looks  doubtful and more sophisticated system is needed. The possible danger of air entering is also overestimated- growing of ambient pressure generates the following chain of events: decreasing of BOG,rising of absolute pressure in tank ullage,some condensing at interphase with additional  

 growth of liquid temperature followed by heating of LNG rest until  reaching a new state.So,any cryogenic tank being heated and drained will give opposing reaction to pressure rise with some delay rather than give up.In general, the LNG fuelled tank  behaves like boiling kettle-heat is supplied, vapor is removed.In our case , the overpressure is determined by external heating  and hydraulic losses in draining line and a tracking system which holds the specified overpressure value(mr. Ping Pong) is quite desirable.As for a danger of rollover when ambient pressure drops,of course, this works pro rather than contra , anyway,estimations would be good. By the way,it is possible to observe , more exactly to hear, the rollover in abovementioned kettle on a loud grumbling sounds from it when water calcium sediments cover  its bottom. Sincerely, Vova,vvoyt@rambler.PS.Initial verson of comment  from 19.08.2013. was sent by mistake to mr. Chris Haslego. Take, please, my pardons.    

[NAP]

 

PingPong:

 

Only if the PT that is giving the set point to the overhead pressure control is configured in a way that its reference atm. pressure is not a fixed value of 14.7 psia but a real time changing value by sensing the actual atmospheric pressure.

 

On the coast of lousiana, I havent observed any noticable change in LNG tanks' (4 of them) operating pressure with changing atmospheric pressure due to depressions during hurricane season...

 

The "theory" of LNG tank pressure changing with change in atmospheric pressure always puzzles me as I havent experienced any such thing with a cloased system LNG tank with PTs having fixed reference pressure...