Relief rate from dry vessel

[MortenA]

APi 521 specifies that the relieving temperature is calculated from T1=P1/Pn*Tn
But for cases where P1 and Pn is far apart then the change in Z may matter.

If you can calculate the condition where the density of the gas is the same at relieving pressure (P1) as it was at the normal P and T (Pn and Tn) using e.g. HYSYS should you then use this value for the temperature instead of the more simple approach suggested by API?

Best regards

Morten

 

[someguy79]

Is this a case of a control volume filled with gas, no flow, and heat input?

 

[MortenA]

yes someguy excactly:

Closed volume (vessel blocked in), filled with gas (density*volume remains constant) and fire heats the gas so that pressure increase.

Since the volume dosnt change, then the density dosnt change and if you know your relief pressure you can then calculate the temperature at witch the density at relieving pressure is equal to the initial (before heating) density.

This is in essense what api does - btu for an ideal gas. If i with no major touble could do it for a real gas - should i then do this or is there some safety margin build into assuming ideal gas (there is - but is it intentional)?

Best regards

Morten

 

[someguy79]

API RP521 - 2007 Section 5.1 Principal sources of overpressure
"The basis for determining individual relieving rates that result from various causes of overpressure is presented in Clause 5 in the form of general considerations and guidelines. Good judgment, rather than blind adherence to these guidelines should be followed in each case."

Also in Section 5.3 of that code:
"...This maximum usually occurs near the critical temperature. In such cases, the assumption of an ideal gas can be too conservative, and Equation (8) (see 5.15.2.2.2) oversizes the pressure-relief valve."

These statements do not explicitly permit the use of such tools as Hysys for determination of fluid properties. However, it seems to me that there are enough other areas in the code allowing you to use good judgment, that if you are reasonably certain your results should be more accurate, they should be permissible.

If anyone has a different interpretation, I would be quite curious to hear it too.

 

[MortenA]

II kno about the "good judgement" i was ua little concerned that a margin was build into the equation and me using more accracy would spoil this?

But i will look more closely at that last quote of your (and maybe go out and get me a net API 521 - im pretty certain its a little older

Best regards

Morten

 

[JoeWong88]

MortenA,
Please allow me to drop my 2-cents opinion...

In the earlier thread related to vessel wall temp, i have highlighted some points related to fire

There are many factors affecting fire type, heat flux, etc and wall temperature would be varies.

Heat Flux Varies
Pressure containing equipment exposing to fire, the wall temperature is very much subject to type of fire, relative distance, momentum of fire, how a fire flame impinging equipment, what type of material on fire, etc and the results is heat flux would be localized.

Low Heat Transfer
Low heat transfer between wall and vapor would leads to wall temperature far higher than the gas temperature.

Environment
Environment factor such as humidity, wind condition, etc affect the heat transmission.

As above factors are difficult to quantify, thus wall temperature of equipment would be difficult to be estimated. Personally i would always recommend to put extra efforts in fire detection, prevention and evacuation of risk instead concentrating in heat flux, wall temperature for gas expansion case. Read more HERE.

Personally i wouldn't pay much focus & attention on this PSV sizing but other prevention measures would be my concern...

Please ignore the following if you are not confortable...

From gas temperature perspective, i would think the equation has considered isothermal expansion with ideal gas condition for conservatism (i guess). The constant density path method that you have considered to use is closer to real gas but still make minimium diference at low pressure and low MW system. If your system pressure and MW are rather high (sorry... can't quote any figure - still investigating) then more margin should be considered (agian sorry can't advise any figure) in order to make it closer to API equation...

Another method which i have adopted for gas expansion is
i) establish the stepwise relieving rate base on constant density path
ii) establish the PSV sizing by adopting the idea highlighted in "Using ideal k for PSV sizing" by Aubry Shackelford.

Infact the second step has built in some conservatism.

The idea in the article "Using ideal k for PSV sizing" by Aubry Shackelford has been used in number of application. My experience is the PSV would be one size bigger for high pressure & high MW system (even LPG at storage condition which is not so high).

Hope this help.