Credit for thermal insulation for fire PSV

[tempelm]

I am in disagreement with some of my colleagues as to whether to take credit for insulation on a propane dehydrator vessel for the fire sizing case. The vessel is insulated and I have selected an environment factor F as per table 5 of API Std 521 in accordance with the insulation type and thickness. This has the effect of reducing the PSV size from a J orifice to an E.

Section 4.4.13.2.7.1 of API Std 521 states that credit is "typically not" taken for thermal insulation because it does not "usually meet" fire protection requirements. This seems like pretty imprecise language for a standard.

I have taken credit for insulation for fire sizing and fire is the only sizing case. My argument is that the purpose of the relief valves is to avoid BLEVE of these vessels long enough that personnel can safely evacuate. Realistically the insulation will stay functional during the time it will take to evacuate.

Part of the API criteria for considering whether insulation and jacketing will stay functional during a fire is that it is not dislodged from water streams or compromised from high temperature. By this time, any personnel in the area will be long gone. If personnel choose to fight the fire, that is an assessed risk outside the normal design. No size of PSV will prevent the vessel wall from failing in a sizeable fire and resulting in an explosion.

It might seem easy to just put on valves sized with no credit for insulation, but there is a high cost to going with larger relief valves, mostly in the increased size of flare pipe and associated piperack steel.

Any thoughts or interpretations of the PSV sizing standard?

 

[pennpiper]

Tempelm,
It sounds like you are splitting hairs, and you have not given important basic information.
1. Vessel size?
2. Line sizes, up stream/down stream with J Orifice?
2.1 PSV size with J Orifice?
3. Line sizes, up stream/down stream with E Orifice?
3.1 PSV size with E Orifice?
4. Line size of Flare Header?
5. Type and thickness of insulation?

Sometimes its possible to do all the right things and still get bad results

 

[georgeverghese]

Generally, bog standard thermal insulation is not good for integrity in a fire scenario - you have to explicitly state that you wish to have insulation of a quality to withstand a fire event. Insulation good for fire is an intumescent coating (looks a bit like concrete) from companies like Chartek (remember a grade called Chartek III) that is coated on the vessel.

Firecase insulation is expected to maintain its integrity and the mechanical integrity of the protected vessel for a reasonable length of time (30minutes I recall)

Having said that, perhaps all is not lost - you must have sized this firecase RV using classical rules with an average latent heat for propane at the RV set pressure. It may be possible to get a significant reduction in relief load if you were to simulate this using a firecase depressure routine in your process simulator. In this case, you would choose an RO size that enables keeping the vessel pressure at a max of set pressure throughout the entire fire event (which would typically be 15-30minutes-check). Using this procedure, the simulator would usually give you a lower relief load than you would get with the standard procedure, because it also accounts for the sensible heat to raise liquid temp in each time step. Check the API for what duration needs to be used when the vessel pressure is initially well below setpoint.

 

[georgeverghese]

Error in my previous advice

" In this case, you would choose an RO size that enables keeping the vessel pressure at a max of set pressure throughout the entire fire event (which would typically be 15-30minutes-check)"

Set pressure should be corrected to relieving pressure

Which, in this case is 1.21 x MAWP of vessel

Firecase only RV may be set at 1.1 x MAWP

Also, discharge line hydraulics would permit a developed backpressure of up to 21% of MAWP even for conventional RV, since set press = 1.1 x MAWP

In some cases, at relieving pressure, vessel temp may rise up to be beyond critical temp, in which case, relief would be for gas expansion only.

 

[mbt22]

Personally I interpret that statement in API 521 as meaning only take credit for thermal insulation if it does meet fire protection requirements.

The API criteria, as you say, are fairly clear - if your normal insulation meets them, take credit for it; if it doesn't, then don't.

I would suggest that someone choosing to fight a fire is not outside normal design. Most sites have a fire team, monitors and the like - suggesting that insulation to mitigate the consequences of a fire case should be designed on the assumption that the fire isn't attacked seems difficult to justify.

Matt

 

[MortenA]

I agree with the other commentors (and your colleagues): You are reading the bible like the devil. IF your insulation meets fire protection requirements, then by all means take it into account - if not dont. The BLEVE scenario is for when the vessel is "boiled dry" but if your assumption fails, then over pressure will occur before as the PSV wont be able to relieve the boil of.

Best regards, Morten