8 replies to this topic

[JoeWong]

Hi,

In determining fire relief load due to external pool fire for vapor – liquid system, wetted surface area needs to be calculated follow by heat load determination per API Std 521 / ISO 23251, eq. (6) in section 5.15.2.2.1. Normally, liquid level at LAH will be considered expose to external fire for conservativeness.

As external heat enters the system, liquid which normally at saturation condition will boil and form bubble & it's rise to vapor phase. Together with liquid expansion, liquid swelling occurs, liquid level rise and subsequently wetted surface area increase.

Question #1 :
Should we consider initial liquid level (LAH) or liquid level at t by taking into account swelling effect ?

Question #2 :
How to determine liquid level at t ?

Hope someone can advise. Thanks in advance.

JoeWong

[Art Montemayor]

Joe:

I'm hoping your query can generate a reply from Phil Leckner, who has probably done thousands of these type of calculations in his career. I've only done hundreds compared to him.

However, I'll give you my frank opinion as to your query. You seem to killing an application with details that detract from the real issue: Protect the personnel and the equipment - first and foremost. To carry out this quality of scope of work, what a process engineer does is that he/she applies what are conventionally (& expectedly) very conservative steps and logic. For example, the height of the exposed wetted surface is sometimes taken as 25 feet above the pool fire; I usually apply 35 feet. The surface area is always calculated in a nominal fashion and very conservative. The resultant relief device should be designed for the worse case and the actual applied orifice size is also conservatively selected. I may be wrong, (& I'll let Phil beat me on the head if I am) but you are worrying about an insignificant factor when it comes down to the amount of volume "swell" that the liquid undergoes. With the contingencies and the conservativeness of the design I would expect that the liquid expansion effect would be practically un-noticed. If I expect the liquid volume to increase due to thermal expansion, I simply apply more wetted surface area to the calcs - say 5 - 10%.

I believe it is more important to reflect and concentrate on having the worse case covered, having a conservative design applied, and paying particular attention to the seriousness of depleting all your liquid heat sink within your vessel in the case of a fire case. In other words, there are far more important and critical factors to worry about than the volume growth of the liquid contents during a fire. If I'm right, thank God! A fire case is a serious and potentially fatal scenario and there are a lot of other things to worry about.

[JoeWong]

Dear Mr. Montemayor,
Thanks for your response. I wish my query may generate response from Mr. Leckner, EGT01, Latexman, etc who are responsive and actively involve in this area.

I am hoping my query is not "killing" an application which has been commonly adopted in many application. Instead, to generate more comments, ideas, responses, etc and ultimately share among us (at least for present generation).

My personal opinion is as an engineer in general and particularly process engineer, one shall take a conventional, conservative & logical approach as minimum basis/ground, he/she shall also put additional thoughts on practial, reasonable, not overly conservative approach or ignoring some effects that has been considered "not important".

I am in EPC business for years (still reasonably young in this industry) and has received a lot of queries (not sure they are reasonable) from international clients. Let take your example, API mentioned that flame can be as high as 40 feet, however, non-equal distribution of heat flux and other factors, recommended 25 feet as reasonable approach. One may take additional 5 feet (if not mistaken NFPA), other may take more margin for conservatism. Regardless of what margin you put, the client is always request reasoning for the basis and "forcing" us to take the approach favour to them e.g. cost saving.

The level swelling in vessel is one "common" question. Some company practice advise LAH to be used but some advise NLL to be used. Personally i have conducted a details study on a V-L vessel using rigorous (Fluid Temp stepwise rise (~ 2 degC) with RV discharge) method. I found that the liquid expansion is is quite substantial and yet taking into account rising bubbles.

Further studies on

- DIERS design manual (vapor-liquid disengagement)
- and some papers from Dr. Fauske " Pressure Relief and Venting: Some Practical Considerations Related to Hazard Control"

has caused me think further...what i usually do following those company general practices, is it safe enough ?

Appreciate everyone can contributes their opinion, experiences, etc.



best regards,

JoeWong

[pleckner]

Art gives me too much credit but I thank him for it!

Is it safe enough? A good question and one that I truly believe can never be known. How do you know if any PSV has been sized 100% correctly? The first gut answer would be that the place didn't blow up. But then again, are the scenarios you witnessed ever what the PSV was actually sized for? Were the physical properties of the material exactly what you considered when the PSV was sized in the first place? Did that vessel pressure during relief really never exceed MAWP? Remember, the vessel may have been tested at 1.3 or 1.5 times design pressure so even if the PSV was just too small, would you have ever seen it? During a fire, did material leak out from a flange that failed and thus LOWERED the liquid level in the tank?

So , we may unfortunately find out if the PSV is too small but will we ever know if it is truly big enough?!?

And this is why we rely on standards and company practices, yours or your client's. This is why we keep abreast of new research on what is going on. Look, until the 7th Edition of API RP520 (January 2000), API recommended two-phase PSV sizing by taking the liquid relief area and taking the gas/vapor relief area and adding the two together. However we knew this was wrong for years but many people still calculated two-phase relief this way siting API RP520 (6th edition at the time). Most of us who knew better used DIERS way before API revised RP520!

In this case, I don't think there is much justification for adding wetted surface area for liquid swelling into the fire calcualtion except in deternining if you have a two-phase mixture or not coming out the PSV. I like what Art does, just increase the height above the sustainable fire beyond the 25 feet recommended by API. I worked for an E&C years ago that had 50 ft. as their standard! If the vessel is insulated, just don't take credit for the insulation even if you can!

Until "other" acceptable and justifiable criteria emerge within the industry, there are a number of ways of increasing the conservatism without spending significant calculation hours, as I show above.

One last thing, for the liquid height in a vessel, by default I use the highest expected operating level unless told otherwise by my client.

[JoeWong]

Mr. Leckner,
Thanks for your advice.

I am totally agree with you...there are so many things (unknown) can happen without us realising them...sometime one effect is cancelled / compensated by another single / multiple effect, etc.
Nevertheless, personally i think as an engineer, we should always take the oppurtunity in any development phase included design to minimise the risk at practical level (ALARP principles).

In your earlier advice "I don't think there is much justification for adding wetted surface area for liquid swelling into the fire calcualtion except in deternining if you have a two-phase mixture or not coming out the PSV..."

I understood the statement as follow : Liquid swelling is not consider for wetted surface area (heat input calculation) but consider if liquid swelling which potentially results liquid entrainment into PSV and size the PSV for two phase relief. Is my interpretation correct ? If so, how do we "gel" these issues together ?

I like your last statement : "Until "other" acceptable and justifiable criteria emerge within the industry, there are a number of ways of increasing the conservatism without spending significant calculation hours, as I show above."
We can always provide margin from different places in order to increase conservatism and minimising the risks.

best regards,

JoeWong

[pleckner]

Joe,

Your interpretation of my statement:

In this case, I don't think there is much justification for adding wetted surface area for liquid swelling into the fire calcualtion except in deternining if you have a two-phase mixture or not coming out the PSV.

Is correct.

My feeling as previously stated is that we do not need to "gel" the two issues into one. We calculate the wetted surface area to get the relieving rate, and we determine if the swelling effect causes a two-phase relief. We don't have to make it any more complicated than that. Now, if you would feel more comfortable to do dynamic simulations or use computational fluid dynamics to account for liquid swelling in the heat input determination, then by all means go for it. I only hope you have the manhours budgeted for this!

[JoeWong]

Joe,

Your interpretation of my statement:

In this case, I don't think there is much justification for adding wetted surface area for liquid swelling into the fire calcualtion except in deternining if you have a two-phase mixture or not coming out the PSV.

Is correct.

My feeling as previously stated is that we do not need to "gel" the two issues into one. We calculate the wetted surface area to get the relieving rate, and we determine if the swelling effect causes a two-phase relief. We don't have to make it any more complicated than that. Now, if you would feel more comfortable to do dynamic simulations or use computational fluid dynamics to account for liquid swelling in the heat input determination, then by all means go for it. I only hope you have the manhours budgeted for this!

Dear Mr. Leckner,

Appreaciate your quick response. Thanks for your time and advices.

Recently i read a paper prensented in API 521 Subcommitee, Sept. 2002,
<< A simple relief valve rating / sizing method for fire relief from Liquid filled vessels >>.
by Bob Sparrow & Helen Mott.

The approach used is made sense to me. Heat gained from fire is calculated using API formula. Wetted surface area has been calculated for entire vessel. It couple with DIERS approach, the Bubbly-viscous and/or Churn-turbulent model. I am hoping that this approach can be extended to investigate liquid swelling effect.

If someone has more advance information, wish you can share with us.

regards,

JoeWong

[pleckner]

Joe,

Is there a copy of that paper that you can share with us?

Thanks.

[JoeWong]

Mr. Leckner,

Sure no problem.

As this document is no personal product and to avoid uneasy feeling from certain party by parking it in the post, appreciate if you can inform you email address or drop an email to to at webwormcpt.wwcpt@gmail.com. I will definitely forward you a copy.

Mr. Montemayor,
I would appreciate to get your opinion on this doc. Wish you can do the same.

Thanks in advance.


JoeWong

P.S. : I have tried to park the document via the Cheresource email/ message facilities but seems no way i can do this.