High Pressure Flame-Gas Tank Safety

Look at pictures of bleves. The tank is usually enveloped in flames long before any failures occur. Nothing on the outside of the tank will survive. Electrical resistance would be increasing with tank temperature. Forget batteries and wires, they're vaporized long ago.

Nicely put. I've only seen pictures. That's enough for me.

BLEVEs can also occur without flames, such as with rupturing boilers or refrigerant vessels. For instance, steam expands by about 1600:1 compared to liquid water. Around 1980(??) there was an incident in Everett, WA, USA, in which a boiler ruptured, tore itself away from its piping, and rocketed 300 feet along the ground. It crashed into a building, crushing a victim.

I GA'd oldsalts answer as it covers most things but just a couple of things

Vapour doesn't have a latent heat thats why there is a reduction in heat conduction and the metal temperature rises.

Initially the fire impinges on the liquid surface, and heats it eventually starting to boil the liquid at the relief pressure. Vapourising liquid takes a lot of energy and boiling liquids have a very high heat transfer coefficient, the combination of which means that the heat of the fire is quickly transmitted through the steel to the liquid and absorbed by the change of phase.

Once some of the liquid has boiled away in part and the fire impinges on the gas part of the vessel, heat transfer rates fall and the gas can only absorb energy by rising in temperature. This causes the steel temperature to increase. So long as there is still boiling liquid in the vessel it will be at the relieving pressure. The vessel fails because the vessel has a design pressure / temperature pair and if you exceed the design temperature the strength of the steel fails and can rupture at the notional design pressure. At the fire temperature the bursting pressure of the vessel can be below its design pressure for the design temperature.

Also I am curious - why would you use monitors and not put the fire out? I see the whole cooling idea. Drenching the area in water absorbs the energy of the fire by boiling the water and either keeps the steel vessel in its design temperature range or cools it. I guess the fire indicates the presence of hydrocarbons and if you put the fire out without addressing the leak you will end up with a pool of flammable liquid, possibly a cloud - propane vapourises at cold ambient temperatures which would give you the risk of a fuel air explosion which is not to be sniffed at. (Or am I down the wrong path here)

In my experience this problem is addressed with a separate blowdown system. This is an automated system which has the risks addressed above however the intention is:

You have a simple block valve in a vent line from the vessel. (Typically of the size of the relief valve outlet). You have a fire and gas safety system - including IR / UV sensors, big red buttons for the operators etc. If these are tripped (or say 2 out of 3 are activated) then the valve opens and depressures the contents of the vessel to the flare. Remember BLEVEs occur in pressurised normal temperature storage vessels. This valve is not spring loaded and so once open provides a free path to the flare dropping the pressure to atmospheric or thereabouts.

However it is automated so there are risks. However normally these valves have a spring acting on the closure mechanism trying to open the valve. On the other side is the plant instrument air holding the valve closed. If there is a failure of the electronics the Inst Air cylinder is depressured and the spring acts to open the valve. These valves always fail to the open position which hopefully gets round the being consumed by fire problem.

In this way you drop the pressure of the vessel whilst it is hot - obviously there is a down side in losing all the product but having seen pictures of BLEVEs I don't want to be within 5 miles preferably 50 of one going off

I am glad I managed to work out why you don't want to put the fire out.

One thing we should have mentioned is that the ground under an LPG or similar tank should be sloped to a controlled area. This way any leaks are directed away from the tank so if they ignite it reduces the fuel available at the tank. However as Propane/Butane vaporise it will not control vapour releases and so the sloped floor may just move the hazard.

As to the Emergency Blowdown System I was trying to describe. It essentially comes from API 521 and so I think I would challenge your comment about its lack of reliability. However my experience is just as a Process Engineer. We have recently installed these systems on the refrigerant accumulators for a Propane and Mixed Refrigerant Chilling system for some very large refrigerated atmospheric tanks in the mid east. These vessels were 4 - 5 m diam, 15 - 20m long running at up to 35 bar so are similar to pressurised LPG storage vessels. The blowdown system described was the key BLEVE protection.

I see the merit of installing a bursting disk although I haven't had that experience myself. However this system has the flaw that if the spring loaded relief valve operates as designed then the vessel is kept at its design pressure and so will fail. The key to BLEVE protection as I understand it is to get the pressure of the system down as quickly as possible if the vessel is under fire. The bursting disk achieves this but if the relief valve works the bursting disk will never open. I accept there are risks with the blowdown valve but they are designed with as many fail safes as possible. A loss of power, air, control signal will all cause the valve to open. Sometimes the air supply is routed through a short plastic tube that will melt in a fire causing the air actuator to depressure and allowing the spring to drive the valve fully open. Obviously it is key that these events occur before the fire fully develops as the heat may warp valves and paths and so the valve may stick in the closed position.

After installing all of this it is still essential that the emergency crews are fully trained on the mechanism of BLEVE so that the appropriate actions are taken when the worst happens. I have seen the videos read the reports and still too many people are killed.

One further thing for our original questioner - if you google blowdown vlaves most of the hits will be for boiler blowdown vlaves which are to remove a build up of slats and other crap from boiler drums so be careful on your searching. But API 521 is a good place to start

The burst pressure of the rupture disc is spec'd for less than the test pressure of the storage vessel, usually less than the max working pressure. That way the vessel is more likely to be usable after the overpressurization incident (if not damaged by other factors). The intent is that the fracture disc will be replaced due to age having never being necessary and not due to use.

The release point of the spring pressure relief valve is at an even lower pressure than the fracture disc. This way if the spring press vlv should not open the fracture disc will open before the vessel becomes compromised.

An example could be-

Spring loaded press relief vlv- 200psi; Fracture disc 225psi; max operating press of vessel 240psi.

Perhaps I am slanted towards pessimism having done many Haz Ops analysis's for the development of processes and Accident Investigations after things failed but many fail safe things fail and cause deaths. The Titanic was fail safe.

Thanks for the responses guys. I especially enjoyed Old Salt's 1/4 mile away experience. Glad to hear you're okay.

I see what your saying about anything with circuitry or wiring possibly failing under high heat or costing to much to shield.

I'm interested in this rupture disk. So following your example: "Spring loaded press relief valve- 200psi; Fracture disc 225psi; max operating press of vessel 240psi." Say you have a bleve scenario where heat has weakened the tank, and the real time tank integrity is down to say 180 psi (below the pressure relief valve. Could you assume the rupture disk integrity would also deteriorated down <180 psi given that the flame is on it? If not I'm wondering if material engineers could work on developing this.

Also I'm assuming most tanks have one of these rupture disk on the top or on one side. Would it be worth the cost to install multiple rupture disk around the sides of the tank where a bleve risk heat source is likely to hit the tank?

Thanks

Con...

I believe that I understannd your line of reasoning....but there are some publications that you should read.

The virtues of a "multiple relief device" design for fire protection is contained in NFPA publications and has been in use for many years......you should particularly review NFPA-30 and its references on tank venting (I think that its NFPA-2000 that you would want)

Many US tanks and vessels in the process chemical industry have multiple protective devices

The site specific requirements for leak collection and "fire zones" are typically reviewed during a detailed HAZOP analysis which is required by US federal statutes.

"Old Salt" hit the nail on the head with all of his comments....bravo !

-MJC