1 reply to this topic

[DalMc78]

I am sizing relief devices for several atmospheric flammable solvent storage tanks and used solvent receiver vessels. These steel vessels already exist with a design pressure of 0.7 barg, and have open vents providing PV relief which are being replaced. The project is to bring the vessels up to current company standards by adding PV breather valves and emergency relief devices. I am trying to determine at what level two-phase flow will not occur from the vessels in the event of fire to see if it can be accomodated by the process. Using the DIERS method I get two phase flow at very low operating levels but from my reading on the subject I understand that for non-foaming, non-reactive systems in a fire case the level swell will be significantly lower than that found using the DIERS method and am looking for a way to calculate this.

DIERS (Appendix 1-B6) says:

"For essentially liquid filled atmospheric storage vessels where little or no over pressure can be tolerated, an emergency relief device can be sized for vapour flow for a non reactive scenario (i.e. fire exposure) if the vapour velocity into the vent at the swelled liquid height is less than that critical entrainment velocity and the vessel contains a non-viscous and non foamy fluid."

The problem to me is fully evaluating the level swell which is caused by liquid thermal expansion and volume increase due to bubble voidage in the zone near to the vessel wall. Fauske et al have written a number of papers (see www.fauske.com/chemPapers.html) on the subject of two-phase onset. The one referenced in the DIERS section above, "The onset of Two-phase venting via entrainment in liquid filled storage vessels exposed to fire", says the effect of level swell due to bubbles can be safely be ignored.

Another of the Fauske papers written but not available on the website, ”Vapour-Liquid Disengagement in Atmospheric Liquid Storage Vessels Subjected to External Heat Source”, allows the level swell due to bubbles at the walls to be evaluated but uses data for water only. Chart data would need to be extrapolated to use fluids other than water. I am not entirely confident of the applicability of doing this but to me is is better and more conservative to calulate the swell due to bubbles rather than ignore the effect.

The DIERS method seems to be overly conservative regarding tank inventories to avoid two phase flow in a fire relief case as it is explicitly intended for use where bubbles are present throughout the liquid. It would be good to find a reference that summarises the necessary approach for low pressure vessel fire cases but I have not been able to find any.

Do others think the Fauske "Bubble swell" method described can be generalised to cover other liquids via extapolation? Is it safe to ignore bubble level swell as stated in the DIERS Fauske reference?

Thanks.

[rxnarang]

Yes I do. Good research

API 521 says the same thing.
Rajiv