Relieving Load Cacls - Rigorous Method to Determine Heat of Vapourization of a LPG 5

[Sirius P.Eng.]

In trying to determine an accurate latent heat of vapoourisation for an LPG tank exposed to an open pool fire, I recently came across this procedure (see link below) to determine the latent heat of vapourisation for multi-component mixtures using Aspen HYSYS.

Title: Determine Latent Heat for Multi-component & Relieving Area using Rigorous Method in HYSYS

http://webwormcpt.blogspot.com/2008/06/determine-latent-heat-and-maximum.html

I have not been able to reproduce/follow the procedure because process steps 1 and 2 are unclear.

Anyone with experience on such rigorous methods may correct the procedure if required or explain clearly what was meant.

Alternatively, if there are other rigorous/highly accurate methods (apart from a dynamic simulation of course) please share.

The LPG Mixture I am looking at has the composition in mole percent:

Propane: 60%
i-Butane: 10%
n-Butane: 30%

I have ruled out using the latent heat of hexane (115 kJ/kg) as suggested in API 521 because it would result in a high relief load.

 

[Latexman]

I use Aspen Plus, not HYSYS. For fire case with components with varying properties, I like to use the dynamic tools within the relief sizing environment. No need to just use one heat of vaporization. It uses the value correct for the time.

Good Luck,
Latexman
Pats' Pub's Proprietor

 

[PaoloPemi]

if you do not wish to adopt a predefined procedure, you can integrate directly with the so called direct integration method discussed in many posts at eng-tips, cheresources etc.
as first step , given a PSV discharge pressure (or a initial value for P), you need to estimate temperature,
for that you can solve a flash operation with specified volume and pressure,
for example, with PRODE PROPERTIES library and Excel, Python, LibreOffice or equivalent tools
t = VPF(stream,p,v)
at that point you can calculate latent heat as difference (Vapor enthalpy - Liquid enthalpy)
dH = (StrSGH(stream)-StrSLH(stream))
to adopt in first step of direct integration,
at each step, you need to estimate heat by fire and the amount of vapor discharged (mass & heat balance),
with PRODE you can solve, for example, a H-P flash operation,
if you include corrections for parameters as heat transfer between vapor and liquid inside vessel etc. the final result should not be too different from predefined procedures (dynamic simulation),

anyway, a point to consider with a PSV is material stress, where possible I prefer rupture disks or depressurization which you can simulate with a equivalent procedure (variable P istead of fixed P).

 

[don1980]

I noted that the referenced article is from 2008 which predates Aspen's Safety Analysis functionality for sizing relief devices. The current functionality in Aspen allows for a good and very easy way to account for the Hvap of multi-component mixtures. Aspen allows the user to specify a fraction of the fluid to be vaporized. Then the program calculates an "Hvap" by dividing the mass of that material (the fraction you specified) by the heat required to vaporize it. So this is an "Hvap" for the composition that makes up that specified fraction. The program defaults to a 10% vaporization fraction, but you can enter any value you choose.

When you have multiple components, there's no single right or wrong answer to the question of what Hvap should be used. Be aware that the higher the fraction specified in Aspen, the less conservative the design. The most conservative approach is to base the orifice size on the single component that has the lowest Hvap. But that often excessively conservative. Thus, the method in Aspen is a good one. It allows you to base the design of a specified fraction of the mixture - a fraction made up of more than just the lightest component.

 

[georgeverghese]

A similar approach is possible with the firecase relief utility, within the Batch Operations Utility in Pro II. It gives values for max relief load much smaller than with the manual API approach using a fixed latent heat.