Heat exchanger exposed to fire

[chemeng20612]

Hi,

I'm currently trying to determine the temperature in the metal surface of an insulated heat exchanger when it is exposed to fire. I mean, if there is a fire and the temperature in the surface of the insulation is 1112ºF, what temperature can i expect to have in the metal surface of the heat exchanger if the thermal conductivity of the insulation is 0.62 btu*in/h*ft2*ºF and the insulation thickness is 0.4 in.

If anyone can help me out with this, i would really appreciate it, i can't seem to find the right approach.

Thanks

 

[dvd]

The metal surface probably can't get much hotter than the adiabatic flame temperature. What are you burning?

 

[marty007]

How is the heat being removed from the inside surface of the metal? If there is nowhere for the heat to escape, then it'll be a time-dependent solution with the end result of the metal being the same temperature as the outside of insulation (1112°F).

Otherwise, it would be a pretty simple radial heat transfer equation with a few layers:
1) Conduction through insulation
2) Conduction through vessel wall
3) Convection from inside surface of vessel wall to... (see initial question).

You could also add a contact resistance between the insulation and the vessel wall if you want to be more precise, but that would lessen any conservatism.

If the inside of the vessel was filled with atmospheric water for example, you'd have a very high convective coefficient on the inside surface of the vessel wall once you hit the boiling case.

If the inside of the vessel is under vacuum, you wouldn't have much of a convective coefficient and the metal temperature would likely rise very high.

A lot more information is required.

 

[MortenA]

Have you loked at API 521, they may cover it. I tried quickly to seee if there where som numbers but couldnt.

However, you can calculate the heat transfer rate with insulation, and then back calculate the inside temperature of the insulation and assume that the wall has the same temperate. This will be conservative due to the high conductivity of the metal if the vessel is liquid filled - and worst case if its filled only with a gas.

 

[chemeng20612]

Marty007 inside the surface the vessel is filled with crude oil which is what i find complicated because the temperature of the liquid will start to raise with time so i can't assume steady state and i haven't figure out what model of transient heat transfer to use. Any suggestions? Am i looking at it in the wrong way?

MortenA i was exactly looking at API 521 right before i had this doubt. In API 521 is said that in order to consider insulation in the calculations, the insulation must be fireproof, in this case it isn´t so i'm trying to figure out how much my current insulation works in case of a fire. Any ideas?

 

[MortenA]

If its not fireproof - and you are designing in accordance with this RP then you must disregard the insulation. But normally you would not worry that much about the vessel wall temperature - excactly because the vessel is full of liquid! The liquid will keep the walls cold. The temperature of the fluid will only rise until your PSV lifts at the boiling point at the SP. I believe that in general its assumed that for a liquid filled vessel the vessel wall will not get _much_ warmer than the liquid inside (due to the high conductivity of the metal and metal/liquid heat transfer)

 

[moltenmetal]

Equipment filled with gases and high boiling liquids can't be protected against overpressure due to fire with pressure relief valves. You need a different device: either a non-resealing pressure relief device such as a rupture disk or bulkling pin device, or a thermal relief device such as a fusible plug. In your case, you have a "crude oil" which probably will distill volatiles for a long time until you get to metal surface temperatures above 400C, at which point you'll start making a bunch of coke. Temperature could rise quickly after that, raising the metal temperature and lowering its mechanical properties, but the relief valve will still re-seat at the same pressure. Eventually, that could lead to a failure. But if you have a fire which persists uncontrolled for that long, something in that system is going to fail catastrophically, whether it be the exchanger or pipe connected to it. Actuators will melt off of valves etc. too during such a prolonged fire.

As to the heat input rate used for relief valve sizing for fire cases, see API 520/521- don't make stuff up on your own. There are three values: one assuming adequate drainage and firefighting, and another assuming that plus a fire-durable insulation (i.e. rigid insulation with stainless steel cladding), and I guess a third I've never seen or used for instances where you can't count on adequate firefighting and drainage.

 

[racookpe1978]

Do you seriously propose heating a metal tank filled with oil by burning an open flame held up to the outside wall of the tank?

 

[MortenA]

@moltenmetal - i dont quite follow you. Are you saying that the thousands of vessels out there protected by a PSV for the fire case - are not protected?

 

[moltenmetal]

Are they full of gas only, or do they contain only a high-boiling liquid? Is fire the only case they need protection from? If prolonged fire credible? If so, a PSV is the wrong device to select. Whether or not the selection is unsafe depends on how credible a prolonged fire is, and just how prolonged it'll need to be to take the vessel beyond yield. If the vessel could reach the temperature at which the relief set pressure develops stresses exceeding the UTS of the material in the vessel, the vessel is likely to fail, which by definition means it wasn't adequately protected. Then again, lots of stuff will be failing during a prolonged fire, so it may be flanges or piping that go before the shell does.