2 replies to this topic

[jprocess]

Dear All:
As you know for fire case relief calculation th designer should specify fire case heat load and liquid contents latent heat.
The normal approach is as follows:
The related equation for calculation of fire heat load is a function of wetted exposed area and environmental factor.So if you have equipment dimensions you have the wetted area and consequently the fire heat load.Also using a simulator you can obtain latent heat of vaporization.Dividing these 2 values will obtain fire relief load.
My question is that how reliable is the fire heat load equation?Who knows that based on which concepts,
API have developed it?
My question is because of a case study that is related to relief load calculation for an oil desalter.
The engineer who is responcible for basic design have been followed the above mentioned approach that seems to be blindly usage of equations.
It is easy to test the equation in a simulator like HYSYS.You know the total mass also the fire heat load from API equation.Defining a material stream entering a heater showed us that the liquid contents will not vaporize after 3-4 hours so the heat load seems to be wrong!
Your valuable comments are appreciated.
Cheers.

[gvdlans]

From what I remember the API heat load equations are empirical equations based on experiments where different sizes of vessels were exposed to an open gasoline pool fire. The actual heat loads will most certainly be different because of the fuel may be different or because you will have a different type of fire (e.g. jet fire or a local fire, see article "Size depressurization and relief devices for pressurized systems exposed to fires", Per Salater e.a. , Chemical Engineering Progress, September 2002, pages 38-45). The authors are/were employees of Norsk Hydro and indeed in Norway they use far more extensive calculations for relief device and depressurizaton systems than what is recommended in API RP 521. See also NORSOK S-001 annex G ( http://www.standard....704_0/S-001.pdf )

This may sound alarming but please bear in mind that your relief valve should be only one of many protection layers that form part of the strategy against external fires. Other protection layers are for example:
- plant layout
- automatic or manual depressurization
- passive fire protection (fireproofing)
- active fire protection (including fixed water spray protection, manual intervention by fire brigade using mobile or fixed fire monitors and/or hose streams and application of firefighting foam to the pool fires)
- escape routes, evacuation plans
- internal and external emergency plans

A relief valve may buy you (valuable) time in order to properly activate depressurization and active fire protection systems, but a relief valve alone will not prevent catastrophic failure of the vessel. Reason is that the relief valve will at best keep the pressure at 1.2 times the design pressure, but at some point the temperature of the steel will rise leading to the vessel failure. This lesson was learnt the hard way at Feyzin refinery in France in 1966, see http://www.hse.gov.u...asefeyzin66.htm. It is believed that the fire brigade did not attempt to cool the burning LPG sphere because they thought that the relief valve on the sphere would prevent it from rupturing. In total 18 people were killed in this accident.

[pleckner]

"gvdlans" gave one source. I want to add two more and these will directly answer your question:

(1) Crozier R.A. Jr., "Sizing Relief Valves for Fire Emergencies", Chemical Engineering Magazine, October 28, 1985

(2) Stickles R. P., Melhem G. A. and D. R. Eckhardt, "Improve the Design of Fire Emergency Relief Systems", Chemical Enginering Progress, November 1995

These are must have articles.

Basically the equations we all use are based on full-scale fire tests conducted in the 1940s using various petroleum fuels. The familiar coefficients for heat flux were determined from a series of plots. Do they have any real relevance? Only for the fuels they were derived from and only for non-reactive systems. Are we conservitive or not if we use them? The answer is both yes and no (clear as mud).

There is nothing that says we have to follow these simplified equations. Indeed, the authors of reference (2) suggest a more performance-based approach, which includes (for non-reactive systems):

a) Establish the maximum credible quantity of available fuel source
Determine fuel burning proper ties (burning rate, flame emisive power and temperature, and heat of combustion)
c) Compare fule burning properties and select the fuel with the highest fire flux potentail
d) Determine the absorbed heat rate and fire duration based on the actual fuel burning characteristics, distance to fire and insulation properties
e) Determine type of fire (pool or jet)

I hope I've wetted your appetites.