12 replies to this topic

[Kryz]

Dear Forum,

I have very quick question about difference between pool and jet fire.

We sized our PSV for pool fire. We have comments from third party: "The use of API 521 for PSV sizing in fire events should be done so with caution. API 521 utilises thermal doses based on pool fires which generally result in lower intensity fires than those expected for jet fire scenarios which are the more likely scenarios for this facility."

We are dealing with LNG plant. I wonder is it something to worry about. I done some estimation and find the Norwegian Technology Standards Institution, which was shown in one the problem in this Forum. It says that heat flux for jet fire would be 250 kW/m2, and the pool depends on condition from 100 to 200 kW/m2. I was trying to use factor and multiply by ratio jet/pool the required relief rate, and then recheck already purchased PSV rated capacity.

1. Is it correct approach?
2. How more would be required in orifice area by percentage if any rule of thumb is known?
3. Do you know any standards, paper, reference etc. about jet fire sizing?

It is also worth to mention that we size our fire case based that there is no insulation (very conservative), but in fact we have insulation (perlite). The calcs shown that for insulated vessel required relief rate would be much smaller - the vapour rate are 10 times less.

Therefore have we spare capacity?

Is really jet fire a case in LNG plant as someone suggests?

Any suggestion.

Thanks, Kryz.

[gvdlans]

I have a few remarks here:

- As I wrote in previous posts, the pressure safety valve is only one of multiple layers of defense against catastrophic rupture of a pressure vessel as result of a fire. Other protection layers can include fire detection, active (e.g. spray nozzle systems, manual intervention, water monitors that can be used to direct water on the area of jet fire impingement) and passive fire protection (e.g. firewalls, fireproof insulation), emergency depressurization, plant layout, drainage, location and orientation of flanges etc. The HSE Engineer should design a fire protection strategy that is a combination of these protection layers and this strategy should ensure that the risk is reduced to a tolerable and As Low As Reasonably Practicable level (in other words: it should result in a safe plant).

- Although heat flux (in kW/m2) for a jet fire is higher than for a pool fire, the affected area will in general be smaller. Especially for large vessels it is unlikely that the complete vessel will be engulfed in a jet fire. Therefore the total heat input (in kW) may well be higher for a pool fire than for a jet fire.

- It is my experience that PSVs are only sized for pool fires, not for jet fires.

- You seem to be referring to the article "Size Depressurization and Relief Devices for Pressurized Segments Exposed to Fires", by Per Salater e.a., Chemical Engineering Progress, September 2002, p. 38-45. Note that this article is basically a summary of the "Guidelines for Protection of Pressurised Systems Exposed to Fire", by Scandpower Risk Management AS (reference 2 in the article in CEP). A few years ago, this document could be downloaded from the Scandpower website.

- In case you want to take credit for insulation you should ensure that:
* insulation materials functions effectively while withstanding flame temperatures during a fire
* the insulation should withstand fire exposure for min. 2 hours
* the insulation should resist dislodgment by high-pressure water streams used for fire fighting
* in case you want to design for jet fires, the insulation should withstand the forces from the jet fire (insulation should be jet fire tested and certified)

[JoeWong]

Kryz,
Pool fire and Jet fire…

PSV is a final safeguarding device to prevent equipment or system from catastrophic failure. However, this device may not protecting equipment and/or system from FIRE attack. We may rely on other active and passive protecting system to safeguard the system. Gvdlan has indicated number of protecting system in earlier response.

Let me take an example, a wet gas KO drum operate at LOW pressure & LOW temperature. External FIRE (regardless of POOL or JET fire) heating up it gas via metal wall. As gas is low in heat transfer (~ 7 - 20 kW/m2), the heat absorbed by vessel wall do not dissipate into gas fast enough and tends to stay in the metal wall. Vessel wall (flanges, gasket, etc) probably failed before the internal pressure reached it PSV set pressure. However, I have seen many designers considered steady heat flow and assuming heat balance between external heat input and internal energy built up but without considering the low heat transfer effect. Installing PSV does not protecting the system from Fire attacks vessel regardless of pool nor jet fire.

Some engineers working with me have done the same…I clarified with them however I don’t stop them using this basis…as I can not think of any basis NOT to consider relieve load due to external fire which code has made it compulsory. In practical, I would focus on emergency depressurization system, providing higher SIL instrumented protective system, etc.

Jet fire…huh…it initiated from flanges leakage, cracked pipe, damaged of some fitting attached to large pipe, etc…it subject to internal pressure at release point…it subject to flame direction, relative distance between release point and impinged vessel…One will always remember that jet flame contain very high momentum which impacting to the impinged vessel (but reduce according as pressure depleted), etc. There are so many parameters, transient parameters to be exact, evolve with time…One shall understand that the fire impinged area may expose to very high localised heat flux (>300 kW/m2), as the flame travel around the vessel, the heat flux reduced accordingly…the quoted figure of 250 kW/m2 (or 300 kW/m2 in some document) is just an average heat flux…

As for the fire impinged area, i reserved myself to comments...

Just to support my long winded story, API 521, section 5.16 has stated that "...a relief device might not prevent vessel failure from jet fire impingement." Hence, i would consider PSV shall not consider jet fire heat flux.

Personally, I would suggest :
i) PSV will only be sized for pool fire using the API 521 equation
ii) Focus on emergency depressurization. Use jet fire heat flux rather than pool fire heat flux. CAUTION : if you use HYSYS, the model only use API 521 equation. Some configuration is required to estimate depressurization rate based on jet fire heat flux.
iii) While conduct depressurizations study, ensure maximum allowable working pressure due to reduce wall stress (reduce according to time) is always below internal pressure exert on the vessel wall.
iv) Increase wall thickness or apply external fire proofing if necessary.
iv) Conduct Scenario and Quantitative Risk Analysis, Jet fire flame pattern analysis etc if above measures are too excessive




As for credit in insulation, I doubt you should take any credit. Who can guarantee insulation will not damage throughout the life of the plant ? You can imagine one day morning you found that insulation on the protected vessel (where you have taken credit) damaged, are you going to shutdown your plant or let your operator taking the risk ?

Sometime we even have problem with reliability of fire proofing with coating. Credit taken in insulation probably will be counter paid off by high insurance premium…


Hope this helps.

JoeWong

[Kryz]

Great thanks to gvdlans and JoeWong,

It makes me confident that we are not in worry.

I was expecting by the way that jest fire has more local feature, therefore total heat input given to vessel under fire even may be smaller; despite that fire heat flux intensity is bigger.

Certainly the various layers provide sufficient protection as a stated in response from you.

Moreover if temperature locally will be very high then vessel will fail anyway unrespectable if PSV is installed or not. More like thermal failure not overpressure.

Finally I wish advice you and obviously agree with you that insulation credit was not taken into consideration. We sized PSV assuming totally damage insulation and use the heat absorbed by vessel in such conditions to get relief rate. PSV was sized accordingly.

So thanks again,

Kryz.

[pleckner]

We should also start to consider the use of a nonreclosable relief device, e.g. rupture disk, for fire scenarios instead of PSVs. Think about it, a PSV basically maintains pressure in the vessel by opening upon reaching set point and closing upon reaching blowdown. In a fire, the last thing you really want is to maintain the pressure! What you really want to do is depressure. The best way is via a nonrelocable relief device. A PSV is good if we want to save the vessel contents. What are we going to save after a fire?!?

[gvdlans]

Good point Phil!

Advantage of a rupture disk is that it is a simple device with a high reliability. Disadvantage (compared to an emergency depressurization valve) is that it doesn't "know" whether the overpressure is caused by a fire or some other upset...

The emergency depressurization valve can either be activated by an operator or automatically initiated by a fire detection system. Therefore it does "know" whether there is a fire or not...

The fact that a PSV does not depressurize and hence cannot prevent catastrophic rupture is why I do not consider a PSV as the main protection layer against fires. Strong point is that PSV reacts rapidly and therefore it can provide valuable time that is needed to activate other protection layers (emergency depressurization and active fire protection).

[pleckner]

Code allows the provision for a relief device to be installed for fire protection in addition to a relief device provided for protection against "other" scenarios. You can set this second relief deivce (rupture disk) at up to 110% of MAWP while the PSV is set at MAWP. Therefore, you can still provide protection against other scenarios and design the rupture disk to protect only against fire. See ASME Section VIII, Div. 1, paragraph U-134 and some of the spaghetti it references. You have to read this section several times to finally get it; at least I did.

In essence, this is telling the rupture disk to "know" the pressurization is due to a fire.

The trick is to design the rupture disk properly against premature bursting. If one would like to pursue this avenue, one should talk to the rupture disk vendor but I believe it is do-able.

I don't like depressurization valves that are activated by a human. One activated by a fire detection system is the way to go but this too must be designed as a "fail safe" system.

[Kryz]

I wish to also add my thank to Phil for his valuable input apart from those given yesterday.

Unfortunately this fire PSV I am talking about is not only protecting against fire scenario.

Other contingency is relief of compressed vapour space when pumping LNG liquid from tank to tank - in case vapour path would be closed for any reason. Or called otherwise vapour displacement during vessel transfer.
To explain you more we are dealing with LNG storage tanks at temperature -150 deg C and operating pressure 290 kPag. And we do not want to loose content when any overpressure event occurs. We obviously have additional PSV to protect against pump full flow (max speed and max possible motor power) and we fit them somewhere else on the piping. They are main protection against vessels overfill.

By again it is good point consider of using rupture disks in fire relief.

Kind regards,

Kryz.

[JoeWong]

All,
Please allow me to drop some simple opinions...

Configuration
I like and agree with configuration proposed by Phil to some extent...Rupture disk designed for fire case set at 110% MAWP while PSV design "other case" set at 100% MAWP...This configuration may work for most conditions except that the FIRE relief load is lower than "OTHER case" relief load. The PSV (generally pop open type for gas service) is capable of handling FIRE generated load. Internal pressure may not reach the RD burst pressure.

As i describe earlier, gas is having low heat transfer, heat is potentially stay in metal wall and weaken the wall. Low heat transfer potentially causing lower fire relief load. Thus, this further reduce the chances of rupture disc / buckling pin burst.

However, this configuration will definitely works if fire load is much higher than other case relief load.

Tolerance & Set Pressure
Rupture disc may have high manufacturing tolerances (i have experienced with 5%). Due to this tolerance, we may need to reduce the set pressure to ~105% MAWP to ensure the rupture disc will definite burst when internal pressure reach 105%+5% = 110% of MAWP. However, on the low side, the rupture disc may burst at 105%-5%=100%...

API (based on ASME) allowed vessel overpressure to 121% for FIRE case...so i think we may set at 115% MAWP.

Some company may not allowed 121% MAWP...

I welcome your comments to correct my misconception (if any).



JoeWong

[pleckner]

I invite all interested to read my series on rupture disks found on this website. I describe how to size the thing, how to set burst pressure taking into account manufacturing range, burst tolerance, operating pressure to burst pressure, etc.

However, let me respond to some of the comments.

@Kryz, the point is that the PSV is used to protect against all relieving scenarios BUT fire. And of course the fire scenario would have to be the controlling scenario for this scheme to work. The PSV should only be for the lesser relief requirements. Also, the PSV would have to be set at a pressure less than MAWP for this scheme to work. This is not necessarily a problem.

You cannot set the supplemental relief device any higher than 110% of MAWP, period; so forget about 115% MAWP.

Quoting from my previous post, "The trick is to design the rupture disk properly against premature bursting. If one would like to pursue this avenue, one should talk to the rupture disk vendor but I believe it is do-able." If the vendor says because of tolerance it can't be done, OK then this becomes a non-issue. It can't hurt to check it out before eliminating the concept altogether. Remember, I said, quoting, "We should also start to consider the use of a nonreclosable relief device, e.g. rupture disk, for fire scenarios instead of PSVs.". I didn't say do it, I said consider it.

[JoeWong]

You cannot set the supplemental relief device any higher than 110% of MAWP, period; so forget about 115% MAWP.

Phil,
You are right. API code mentioned set pressure should not exceed 110% of MAWP but...actually i am still unclear...

Since ASME allow relieving pressure upto 121% for FIRE case, with the configuration of PSV + RD, the PSV (1st) set pressure is at 100% MAWP. If the RD tolerance is 5%, can we set at 121%-5% = 116% maximum ?

My understanding on this issue...If 2nd device is PSV, due to slow response time, initiate reaction force from the spring, pressure accumulation is possible. So, the maximum setpressure shall not more than 110% of MAWP. But for RD, fast acting, I guess there is no pressure accumulation. Isn't we can set at higher pressure but <116% ?

It just a question...

JoeWong

[pleckner]

RD tolerance applies to the stamped burst pressure only and this (the stamped burst pressure)cannot be any higher than 110% of MAWP. The tolerance is recognized by the authorities as a variation that happens due to the nature of imperfection.

In your writings, you are implying at the least that the tolerance will be +-5% and this is not always true. The tolerance is only allowed to vary by +-5%. It may be 0%! This is why one needs to consult with the manufacturer first.

I also think you are reading into the reasons for these various set pressures way too much. I admit not knowing the reason the committee came up with these variations but I don't think it has anything to do with slow repsonse time, etc. I think it has more to do with practicality in protection with some economic considerations thrown in. The same thing for inlet pressure loss. Why 3%? Why not 2% or 5%? They needed a number to guide people and what was chosen gives them peace of mind considering safety factors and testing pressures.

[vira]

Hi,

You should size the PSV considering a Pool Fire only considering a global average heat load of 100kW/m2.

Jet fire heat loads are point loads specifically to test the material strength agianst rupture from Fire , as per latest Scandpower guidelines and Jet Fire heat flux load to be used is 350 kW/m2.


In any case full relief from a relief valve should not be excepted as a BDV or a Blow Down valve will open first in the event of the fire , hence what is required is to also size the BDV orifice considering a pool fire @100kW/m2 of global average heat load.