PSV Fire Case vessel full of liquid 6

[JASR76]

Hi all!!

I have a question about Fire Case in a vessel full of liquid (HX, etc..) engulfed by fire using API 521 recommendations.

a) At the beggining the liquid will have a thermal expansion and the PSV will relieve liquid.

b) After that, an increase in temp will relieve vapors generated with some liquid.

c) Finally (some hours?), level in equipment will reduce and only vapours will leave the PSV.

My question is:

1/ Which scenario do you normally use to size the valve?

2/ Do you calculate all cases and compare? or do you normally consider a typical case, for example, take the worst case in terms of duty/latent heat and consider ALL as vapour in worst case.

I dont know if there is a normal procedure as good engineering practice and I would like to know your always interesting opinions.

Regards

JASR76

 

[JASR76]

In this particular case:

a) we are not at bubble point temp (liquid is subcooled at operating conditions)

b) exchanger is always full of liquid and PSV located in this vessel only "sees" liquid

c) Fire is the critical case to study

So my question is not how to avoid this biphasic relieving locating PSV or any other method, my question is (once you really have this contingency):

1.- How do you study this kind of fire case with a vessel full of liquid?

2.- Do you somehow consider a biphasic discharge? API equations are different for vapour and for biphasic reliefs. Orifice could be quite different, considering that my fluid (c4) even could flash/vaporize across the PSV.

I recommend reading API 521 3.15.3.2.

Your comments will be highly appreciated

JASR

 

[Latexman]

For fire exposure and the PSV nozzle is below the liquid level, size for two-phase flow.

Good luck,
Latexman

 

[JASR76]

Thank you, Latexman

How would you calculate massrate and which vapour fraction would you consider?


Regards

JASR

 

[Latexman]

JASR76,

It is currently considered good engineering practice to use the methodology of The Design Institute for Emergency Relief Systems (DIERS) Project Manual. You should obtain that book or use a methodology in a reputable journal that references DIERS.

Good luck,
Latexman

 

[CJKruger]

JASR76,

Regarding your original question.

Before DIERS, most people just sized the liquid full fire case PSV for the calculated vapor load (vapor flow based on wetted surface area). Initially of course it will relief some liquid, but it was assumed that the PSV will be large enough for it.

With DIERS and the experiment Latexman mentioned, things become more complicated. I have not read through the DIERS report, but feel that opening a 2" valve and depressurizing a vessels is not the same as relieving a vessel.

 

[JASR76]

THATS RIGHT CJKruger !!

That was the kind of answer I was waiting for. It was assumed the relief valve was sized as vapour and big enough to assume some of liquid at the initial moments.

I dont have DIERS papers in my possesion to know if DIERS suggest some strategy to deal with these kind of problems

Anyone knows?

THANK YOU ALL

 

[Latexman]

CJKruger and JASR76,

And 150 years ago, no one put relief devices on their boilers! At least read and understand DIERS before you pooh-pooh them.

Good luck,
Latexman

 

[CJKruger]

Latexman

You are reading things in my reply that is not there. Also, please be specific if you disagree with a particular sentence in my reply. A generalized statement from you to imply that I am reluctant to change to newer and better things is both untrue and unnecessary.

 

[Latexman]

I have not read through the DIERS report, but feel that opening a 2" valve and depressurizing a vessels is not the same as relieving a vessel.

To me, "opening a 2" valve and depressurizing a vessel" is exactly how any non-closing relief device, like a rupture disk (RD), works.

This is how the DIERS team simulated a relief opening, and, more importantly, maintained control of the time the "relief" opened. With a RD or PSV, you know the approximate pressure it's going to open, but it's hard to know exactly when it will open with normal variability in the reliefs and the process. So they could double check that their instrumentation was ready to go and know for a fact exactly when the "relief" opened they used fast actuating ball valves. I know if I was present at a test such as this, I'd much prefer knowing when things were going to get rowdy than be caught by surprise!

DIERS was not studying the specific fluid flow of a specific manufacturer, model, and size of a PSV or RD. They were studying two-phase flow onset, two-phase flow and various flow models for two phase flow, and vapor disengagement dynamics, so as long as the device that opened could be characterized fluid dynamically, the simpler and cheaper, the better.


Good luck,
Latexman

 

[CJKruger]

Latexman

Thanks for the comprehensive reply.

Yes, I agree opening a 2" valve is a good simulation for a rupture disk.

But, I still do not think it is a good simulation for a spring loaded relief valve. Thus, I would need more justification to change away from a procedure that has been used by me (and others) countless times.

 

[Latexman]

CJKruger,

Would you also agree that "opening a 2" valve is a good simulation" of a fully open spring loaded relief valve?

Good luck,
Latexman

 

[CJKruger]

Latexman

Yes, while it remains fully open. In general I suspect a spring loaded relief valve is closer to a back pressure control valve than a full open ball valve.