Example for deflagration bladder tank, preferable bolted enamel steel?

[MartinLe]

Does anyone have a reprot or news about a deflagration in a low pressure (10-50 mbar) gas storage tank (or tank with other purpose but constructed a similar way, e.g. biogas digester), preferable a bolted construction from glas fused steel plates.

In a current project, someone raised the question of explosion in a sewage gas tank and how far the individual plates would be flying. In my (and the company erecting the tanks) opinion, first the membrane would rip, if this would not suffiently release the pressure the tank would rip open along one seam similar to a sausage.

But we don't know if and whan this ever happened.

The gas system is at a positive pressure. The tanks interior is treated as an ex-hazard zone (only atex approved equipment, lightning protection, warning signs). I'm not sure the someone who raised the question is very well informed and it is likely that we never hear this specific line of enquiry again, but the question piqued my curiosity what damage a deflagration in such a tank would actually do.

 

[LittleInch]

Well a lot depends on your tank construction. This sounds like it could be a fixed roof tank?. The roof will fail before anything else does or the roof / shell seam (frangible roof).

The info below was quite easy to find - the second one with some videos is very interesting, especially the one where the whole tank lifts off like a rocket(!!). You don't actually see it land. the second tank the roof comes off and acts like a frisbee.

Now whether a gas storage tank would do that and how you would get to an a explosive position is more for you to deal with. Methane Gas has a relatively small explosive mix window, but never say never.

The commonality in many of the damaged tanks below is that the roof separates from the shell. Now that can result in the roof coming off like a single plate or a jet of gas / flame coming out from the open portion.

https://aiche.onlinelibrary.wiley.com/doi/full/10.1002/prs.10459

http://www.assentech.co.uk/storage-tank-incidents/

https://www.reliableplant.com/Read/23721/lessons-prevent-death-tanks

https://www.reliableplant.com/Read/23721/lessons-prevent-death-tanks

https://www.researchgate.net/public...hane_gas_explosion_in_fixed_roof_storage_tank

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[MartinLe]

Thanks, I could only look at the page with the videos and the reliableplant article. Spectacular images and a very sobering read.

I must admit that I don't know all the applible codes in my area that well. When I talked to the contractor, he told me that the connection roof-wall must fail before the connection wall-ground, to prevent a tsunami of whatever liquid is in the tank - or the kind of rocket start we see in one of the videos.

 

[LittleInch]

Well it's basic physics. If your tank shell is made of something decent then it is about twice as strong as the roof and even stronger than the roof to wall joint if this is designed as a low pressure tank.

Plus your contractor is correct. Having the floor plates fail means you have a rocket. Usually these are pressure vessels or API 620 tanks where pressure can be a lot higher so the roof and roof to wall seam is a lot stronger.

This is the famous one where an air tank failed during an air test. It started off at ground level...

TBH I think a low pressure tank won't disintegrate, but the roof could lift off and go some distance. Very unlikely but a low probability, high consequence event.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[JStephen]

Usually, that is not a consideration in tank design.
Tanks of any size storing flammable liquids and having a fixed roof will either have a frangible roof (intended to fail prior to the shell-to-bottom weld) or will have emergency venting. However, in either case, the design consideration is the vapor-boil-off issue, NOT combustion inside the tank. More specifically, the emergency vents are not assumed to be adequate to vent an explosion.
NFPA 30 specifies property line set-back distances for tanks, but I'd be pretty sure these are for a tank fire, not a tank explosion.
I have heard of calculations being performed to determine a "safe" distance for pressure vessels that are pneumatically tested rather than hydrotested. Of course, the pressures involved are a lot higher, too. But I don't have any specific leads.
I have seen photos and one example in person where a atmospheric tank shell failed due to corrosion, weld flaws, etc. In those cases, the shell just unwrapped, so there was some hazard in the immediate area, but nothing went flying, either.
I remember hearing of one tank that blew the roof off due to internal combustion, and the roof basically flipped like a flapjack and came back down on the tank, this being a 50' or 60' tank. There again, obviously a hazard in the immediate vicinity, but not blowing pieces over a square mile like a pressure vessel.
I believe API and/or the Welding Research Council have published some info on frangibility of tanks that might be on interest.
I believe there is some information available on venting silos for dust explosions that also might be of interest.

 

[MartinLe]

In my specific case, the tank will be filled with sewage gas (methane/CO_2 mix). So far, all procedures are designed to prevent an explosion or deflagration, not to mitigate any effects.

@littleinch: My guess is that in the instances where the tank lifted, it was welded and the roof was about the same strength as the walls.

Consider a tank built like this

https://www.glstanks.com/en/production

or this

https://www.harvestore.de/referenzen/gasspeicher/niederdruckgasspeicher-bremen.html

... If the wall (and not the membrane inside the cylinder, followed by the roof wall joint), would the bolted connections around one plate fail, or rather one vertical seam? The connection is far weaker than the plate itself. But all of this is very speculative.

 

[human909]

How low pressure are we talking? Would explosion relief panels make sense?

Avoiding means of explosion ignition is usually a good line of first defence. But if it is at all practical providing relief in the event of an explosion is a significant safety improvement.

 

[LittleInch]

I still think the roof or roof / wall connection would go first, even for a bolted tank.

I can't see a single plate being blown far and wide for this sort of tank. The plate connection isn't that much weaker than the plate or it wouldn't seal or hold up under filling.

If it ever did go bang your roof needs to be the weak point by design and will most likely flip up, vent the explosion then collapse down back onto top of the walls.

There were a large number of types of roof in your links - which one have you got? Dome? cone up? flat?



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[MartinLe]

Cone up roof. galvanized steel structural beams inside, stainless steel plates outside (again, bolted together).

I still think the membrane between gas sspace and air space will fail first:

Our relief panel? gas space is yellow, air space blue

The working pressure of the tank is low, 40mbar. The pressure of an explosion would depend on how much air actually enters the system to form a mix. The scenario I can think of where an explosion in the tank might happen is when the tank is opened for maintenance and not all gas has been flushed out. The tank would be practically empty then.

 

[LittleInch]

I've never seen anything quite like this before...

Is that vertical pipe / guide for real? Presumably give you a good idea of how much gas in in there and centres the internal roof.

Any pictures?

Looks like a sealed internal "floating" roof design.

Usually the key risk is a hazardous atmosphere in the space between the internal roof / membrane and the outer roof due to a leak in the membrane or seal.

Yes the membrane would go pretty fast, but this would then pretty soon after ( a few milliseconds) overpressure the void above the internal roof and blow the outer roof off. IMHO.

The other risk from an explosion in the lower section of a bolted tank is that the wall, floor joint breaks and the whole tank lifts off. What is the joint between the wall and the floor?



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Alistair_Heaton]

I have but the bottom half was under the ground due to the worst case situation of it being half full when the void went bang.

The fluid transfers the shock wave down and as you say if the floor doesn't go then the sides split or any valves in the side come out at exceptional high speeds.

You can use an engine cylinder analysis to get real numbers. Its none linear as hell.

 

[MartinLe]

AFAIK this type of gas tank is mostly or only used for sewage gas and biogas.

The central pipe guides the ballast plate and ensures it stays level. It's actually built as several telescoping pipes. Measurement of gas volume is done by radar distance meters from the roof, additionally there's a rope connected to a dial for a rough measurement.

Yes, in all of these installations the air space (and the space around the vents) is considered an ex zone 1 (as defined in ATEX regulation, probably corresponds to Division 1 in the US system).

The wall-floor joint is an angle bracket outisde, bolted to the concrete platform the tank rests on. There's another bracket inside, the tank has an inner steel floor that is bolted to this inner bracket.

The roof is a galvanized steel frame, the roof plates are bolted on top. According to the contractor, the seams between the plates are the most likely to break if an explosion occurs in or expands into the air space.

The tank is constructed for a pressure <100 mbar, all pipe fittings, valves etc. are PN 10.

Can't post pictures from the site and so far I haven't found good, indicative pictures of the exact design online.

 

[LittleInch]

Well that looks a fairly strong tank.

If someone is still worried you might want to then think about "explosion vent panels".Just Google that term and lots of info turns up.

They use simular on offshore rigs to vent any explosions out the side rather than contained volumes.

I don't think this type of tank conforms to any specific standard but if there is see what it says about overpressure.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Alistair_Heaton]

Sounds like the old coal gas tanks they used to use in the UK.

Per say they never had a roof so no issues with the void trapping gas mixtures.

I suppose you have two methods one is vent panels and the other is making sure that no explosive mixture can occur in the void. Either by forced ventilation or passive wind sails. An inert gas system wouldn't be cost effective.

 

[LittleInch]

Most of them were simple gas holders with water seals of a few inches.

These are a giant balloons which are compressed a bit by the weight of this metal movable lid.

If the OP can't find any decent pictures or drawing then it shows they are not too common.

Most seem to be these double bladder tanks, but this looks close

http://claroglobal.com/biogas-storage-digester-covers/

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Alistair_Heaton]

Reckon it would be pretty easy to find out what does go on.

Big can of hot chocolate with press lid. Balloon full of water half filling it. Circle of thin plate on top of balloon.

Two small holes one in the top lid, one in the side a quarter way down.

Inject methane into the side and light the top hole. Remove the methane supply so air can get in the side hole. And walk swiftly away before the top flame dies down. You could also put a known mix in and use a spark plug to ignite and not bother with the holes.

I would only loosely have the lid on for the first attempt. Second one ram it down.

Next level up is stick some strain gauges round the bottom. And make a FEA model up and use the strains to confirm the model and then do a model of a full size version.

That would have been a fun days work when I was a research assistant. The chief lab technician would have been after my blood after the first firing though.

 

[LittleInch]

Getting back to the OP, I think the void above the bladder is your issue.

Good that it's rated zone 1, but a bang in that area could release large fragments or lift the roof if you don't have specific blow off panels or a frangible roof.

A gas sensor in this area looks like a must have to me and perhaps allied to large emergency vents which can be opened to avoid an explosive atmosphere before you depressurise the tank.

This should come up in your HAZID or QRA.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Alistair_Heaton]

I agree.

I would expect not only monitoring but also forced ventilation.

I have a suspicion/gut feel that the bottom of the tank would go as well due to the piston effect of the plate on the top.


Because the membrane has support all the way round it then it will just expand up to the limits of travel and transfer the load to the surrounding structure. It might not even burst until the structure fails. So you have the top going up and the bottom going sideways.


How much kinetic energy the side panels will have i really don;t have a clue but it could be a fair bit round the opening when the membrane eventually goes.