DIERS, Fires and 2 phase relief

[babelfish]

OK, folks open season on thoughts with regards Fires, DIERS and two-phase relief.

There is much deabte going on in the UK about the use of the "HSE Workbook" for the fire case scenario. For those of you outside the UK this is just the DIERS work restated in an easy to understand way.

What is anyone and everyone's take on using the DIERS "techniques" analysing the fire case?

My stance is that two phase flow doesn't normally apply. The most up to date stuff I can find in on Fauske's web site published in CEP in 2000 (Properly Size Vents for Nonreactive and Reactive Chemicals). There are others.

My main concern is that everyone is getting sucked in to spending more than necessary designing and building relief systems without applying their brains.

Any and all points of view welcome.

Are we all speaking the same language?

 

[JLSeagull]

Welcome.
I am in control systems engineering. Our group uses PSV sizing that does not consider two-phase flow. We process provides data for a relief valve that reflects two-phase flow we use their calculations and attach to our specification data sheet with the requisitions. I belief that the process calculations are DIERS based.
42

 

[MikeClay]

API Standard 521 / ISO 23251 "Pressure-relieving and Depressuring Systems" January 2007 paragraphs 5.15.3.3 and 5.15.3.4 indicate that two phase flow is not normally considered unless certain conditions exist.

--Mike Clay--

 

[don1980]

The need for 2-phase relief design is very rare for any non-reactive relief scenario, and that's especially true for fire. The exception is when you have a fluid that is known to be foamy, such as a process in which surfactants are added.

For non-reactive and non-foamy processes there might be some initial 2-phase material released from the relief device but liquid disengagement will almost always occur before the system reaches the relieving pressure (peak pressure). Therefore, 2-phase relief design isn't necessary for those cases. That's typically even true for vertical vessels that operate liquid-full.

Unless the vessel has a mixer that continues to run, a fire scenario will result in wall-boiling but not "volumetric boiling" (uniform boiling throughout the cross-section of the vessel).

In a reactive scenario you have uniform volumetric boiling, thus the high risk of 2-phase relief. Foamy mixtures casue 2-phase relief because you just don't get disengagement until the vessel is nearly empty. But for other, more common, cases it should be rare that you need to size the relief device for a 2-phase relief scenario.

 

[babelfish]

don1980 you have got my full agreement on that, but how or what do I quote that stops people believing that they should consider 2-phase flow in the fire case. I've tried Fauske's work from CEP, ICI's PSG8, even the foreword of the HSE Workbook (UK's DIERS report) but all to no avail. They just say, but you have to do the level swell calcs.

One specific case we are debating I have looked at two-phase flow and I'd need a new nozzle on the vessel. Assuming the intial two-phase flow rate required cannot flow out my vessel I get some pressure rise, [its a tempered (vapour pressure) system (no reaction)] whilst I'm losing mass out my relief stream. As I lose mass the heat input from my fire is falling (assuming the wetted wall is for liquid full), so if I let the pressure increase a bit the relief system goes to vapour flow and I can handle it with my relief system.

The problem comes when you calculate the temperature rise (heat you're not removing by boiling liquid heats the bulk liquid). The vapour pressure exceeds the allowable 21% overpressure(we are working with a set press. of 0.41barg), again this will be transient as now my flow rate is higher than required [end point was assumed when flow rate caused backpressure equal set+accum press.](>than heat input) and the pressure will fall.

Now there is an obvious possibility of mechanical failure during this "pressure excursion", but how dangerous is it really? Is the risk high enough to justify major vessel modifications? Anyone have practical experience in this?

"Are we speaking the same language?...engineering - it's the appliance of science."

 

[Latexman]

In latex production units we add lots of different surfactants in lots of vessels to stabilize the emulsions that are made, so we have to consider two phase flow in a fire scenario. I've seen the aftermath of a burst rupture disk on these vessels. The best description I have of it is the vessel "throws up" it's contents.

Good luck,
Latexman

 

[don1980]

Refer to API-521/ISO-23251. You'll find this same explanation of why 2-phase relief is rare. Do a search for "two phase".

In the vast majority of cases, the DIERS test calculations will indicate single phase relief for non-foamy fluids (disengagement occurs before the peak relieving pressure is reached) however there are conditions that will result in the need for a 2-phase relief design. The ideal solution is to use a PSV sizing program that includes this screening test. There are some good commercial products that do that.

 

[babelfish]

Gents,

Nice comments. Our client has their own variation on the two-phase flow calc. Which does cause problems. One hardly every says two-phase flow but the other usually does. However, they both disagree with DIERS/API521. Not to mention their own standard says you should design for two-phase flow.

As for two-phase flow, etc and runaways. I worked at the place where the "incident" involved ripping an 8" ANSI150 off a heat exchanger and turned it through 90deg unitl it hit structural steel and then launched the contents at sonic velocity in to the field acrocess a major freeway. Much fun investigating

I'll try API521 in detail and see if that gets any clearer guidance.

"Are we speaking the same language?...engineering - it's the appliance of science."