33 replies to this topic

[S.AHMAD]

Ankur
1. Thank you for sharing. The slide is very informative.
2. I seek your permission to use the slide contents for training young engineers.

Edited by S.AHMAD, 14 April 2012 - 01:01 AM.

[ankur2061]

Ahmad,

I am not the originator of the powerpoint presentation. I have lot of well wishers and friends who keep sharing things.

You are free to use it to educate young engineers.

Isn't that all of us veterans are trying to do through "Cheresoureces"?

Regards,
Ankur.

[S.AHMAD]

Ankur
Thanks a lot. I intend to extract the contents into my own slides. Thank you again.

[kkala]

Comments on post No 25 by ankur2061 are as follows.
I am willing to bet my job if somebody can prove me that a PSV designed for fire case can protect a sphere from destruction if a BLEVE occurs in the vessel.
Things would be much worse without PSVs covering fire case, any BLEVE would occur sooner.
Note:-A sphere is totally destroyed by BLEVE; meaning of Ankur's statement may be that PSVs sized for fire will not save the sphere from BLEVE during fire. I think PRVs can save the sphere, depending on conditions, or at least delay BLEVE.
-In Chiba accident (post No 21), 16 spheres were involved. 11-13 of them were "burnt or damaged" without BLEVE, which is important.
- Depressurizing during fire (not through PSVs) has advantages, but depressurizing valves cannot replace relevant PSVs, as far as I know. Concerning the two depressurizing systems of Aspropirgos Refinery (Greece), (http://www.cheresour...ves-on-spheres/ '> http://www.cheresour...ves-on-spheres/ ) , PSVs sized for fire have also remained. I think depressurization is (still) rare in spheres.

Even an external vapor cloud explosion at some distance would be enough to destroy the vessel due to the shock waves it will create and no fire PSV is going to help protect the vessel.
An explosion will most probaly cause LPG leakages and subsequent fire, so the vessel (s) have to be protected against fire. Refer to post No 20, useful views on the topic, para 3.
Most severe consequences from a BLEVE come from fireball radiation, so rest vessels have to be protected.

Predicting intensity of fires and explosions is at best an exercise in trying to console ourselves that we provided a protection device to protect the vessel.
To know more about the devastation and unpredictable nature of fires and explosions refer the attachment provided.
Effectiveness of fire protection practices is another cup of tea. As a result of developing human experience they may not be perfect, but they have effect. I believe we cannot ignore them, but we can be reasonable conservative in application. Refer also to post No 20, useful views on the topic, para 8.

Readers can read posts No 20 and 23 for more data supporting kkala's views.
Kindly note that responses to post No 24 by kkala are expected (references of LPG spheres with PSVs not sized for fire).

[Technical Bard]

I would argue that some of the worst consequences of BLEVE are high speed projectile shrapnel from the vessel. There are numerous examples in the accident literature of vessel parts weighing dozens or hundreds of kilograms impacting other equipment/buildings up to 1000 metres from the blast itself. Yes, the rising fireball radiation is damaging, but is not the only thing of concern. A notable example of this is Flixborough, UK (1974) which had a significant impact on the layout of storage bullets and other safety systems related to pressurized storage.

[kkala]

Certainly blast of a vessel under external fire can cause a lot of damage to people and equipment, due to overpressure wave and propelled metallic fragments. In submitted safety studies (e.g. refinery desulphurization unit) we used to write that thermal radiation due to BLEVE of a vessel containing LPG has more severe consequences though. This according to references (TNO institute of Netherlands?) of previous work place, that I can no longer check. Some notes on the matter:
1 (Β). Domino effect area after BLEVE is considered within a fireball radius in safety studies. Actually it is the projection of fireball onto ground (no affection of pressure wave / fragments).
2 (P). http://airfresh-soc...ster-1984.html : BLEVEs propelled metal pieces to 1200 m, apart from fires. Domino effect after BLEVE in Aspropirgos Refinery seems on bigger area than fireball radius (but I think this is not right, probably it does not concern BLEVE).
3. Risk Analysis by M Assael and K. Kakosimos (Tziola, 2008, Greek book) reports:
- (Β)"However the most important consequences occur when the fluid is flammable, since then a fireball is formed".
- (P) Metallic fragments are propelled up to about 800 m.
- Fireball has usually diameter 100 - 150 m.
- Deaths due to thermal radiation have been observed up to 75 m distance around.
-Typical fireball surface radiation 150-300 kW/m2.
- Flixborough accident (1 Jun 74) was caused by escape of cycloexane leading to vapor cloud explosion.
Note: BLEVEs may have resulted as domino effects. Similar cause as in Feyzin or San Juanico accident.
(Β) fireball more serious (P)=overpressure wave and propelled metal more serious.

Having not "practical" experience of BLEVE events, I can only say there is an wide opinion considering thermal radiation effect as more serious, on a statistical basis. I also think that effect of fires, including those resulting for leakages due to (P), is more serious. Data differ from place to place and supporting data would be welcomed to clarify this interesting issue.

It is also noted that: Protection of vessel against overpressure waves and propelled fragments is not a fire protection issue. But fire can be the result of such accident, needing fire protection measures on the vessels facing this risk.

Edited by kkala, 16 April 2012 - 10:32 AM.

[ankur2061]

No PSV alone can protect a sphere once the BLEVE occurs because the vessel is ruptured immediately on the occurence of a BLEVE. The probability of preventing a BLEVE due to a fire would involve a combination of measures such as water / foam spray or water sprinklers (active fire protection) , fire proof or fire retardant coatings (passive fire protection) and pressure relief.

The definition of BLEVE itself as per the presentation attached in Post # 25 is as follows:


The result of a vessel failure in a fire and release of a pressurized liquid rapidly into the fire. A pressure wave, a fire ball, vessel fragments and burning liquid droplets are usually the result.

Also when referred to slide 67, an unconfined vapor cloud explosion (UVCE) has a greater coverage in terms of distance compared to BLEVE. For damage to structures and human injuries due to a UVCE refer slides 55 to 57 of the attached presentation.

Regards,
Ankur.

[kkala]

1.1 In case of fire risk concerning LPG sphere(s), PRVs sized for fire are necessary. This does not mean adequate (no post has said so).
1.2 API 2510 A (1996), 6.4.3: If there is no flame contact on the shells of the adjacent vessels and the relief valves have not lifted, if there is sufficient distance between the fire and the vessel shells of concern and cooling can be delayed or not applied at all. When the relief valves lift, the need for cooling water must be assessed.
Note: This looks like a firefighting (not design) instruction. I would try to specify permanent water cooling (post No 23, 5.8) at any case. See same API 2510 A, 6.6.3.1.5.
1.3 Once BLEVE has occurred, vessel is destroyed. Target of PSV sized for fire and water cooling is to delay or prevent BLEVE (post 29).
1.4 References of LPG spheres without PSVs sized for fire (ie without having risk of fire) have been requested (post 24), it must be a rare case.
1.5 Of course passive fire protection is necessary to extend time of structure survival during fire.
1.6 For use of foam and other substances against LPG fires, look at http://www.cheresour...n-storage-tank/ '> http://www.cheresour...n-storage-tank/ , post 10 and on.
1.7. Emergency depressurization could save from BLEVE, being important even if the sphere is destroyed. But probably it is not (yet) considered feasible concerning spheres in most of refineries, http://www.cheresour...pcv-on-spheres/ '> http://www.cheresour...pcv-on-spheres/ .
2. Yes, vapor cloud explosion can "strike" farther than a BLEVE.

[kkala]

As a conclusion from posts No 8 - 23, I believe that LPG spheres should have PSVs sized for fire on them. An exceptional case (if any) has to be decided after patient investigation in my opinion. References / specific examples of such cases were requested in posts No 20, 23 (para 4), 24, 29, without response (so far).
Of course additional measures should be also considered (emergency depressurization, water deluge or spray systems, etc). A PSV sized for fire may not be enough to avoid BLEVE, yet it is necessary as a last resort to delay it.
"After an accident, if you can convince the inspector that the relevant design had been according to accepted codes (or safety level was higher than their requirements) you get free of charge". I read something similar in the past, judged it right. So much care is necessary, even though improper operation or maintenance can be a more frequent cause of accidents than design pitfalls.

Note: I do not know whether PSVs sized for fire are mandatory on LPG spheres by code at any case.

 

Editing note, Feb 16 2013: <http://www.cheresour...age/#entry69308> could be useful to a reader interested in LPG vessel layout and fire protection.

Edited by kkala, 16 February 2013 - 12:36 PM.