Liquid formation in flare header 1

[sjackson]

I am tring to determine if a drip leg on a flare subheader has sufficient holdup volume.

The flare subheader is 24" NPS and the drip leg is located at the low point on the flare subheader prior to elevating about 2 m in order to tie into the main flare header (upstream of KO drum). Our client has asked that we produce calculations to show how long it would take for the drip leg to fill with liquids during a relieving event.

There are approx. 10 PSV's located upstream of the drip leg on the subheader. I have checked several releiving scenarios for each PSV to determine if any liquids will form downstream of the valves. None of the relieving scenarios produce any liquids. I have also checked to see if the mixing of multiple relief streams would produce liquids and once again no liquids appear to form.

My knowledge of flare systems is still fairly elementary and I am concerned that I have not considered all possible causes of liquid formation. In my experience I have never come accross a situation where liquids form downstream of a PSV or other releiving device. In every case it seems that the saturation temperature downstream of the PSV is at least 20 to 30 C below the releif stream temperature. If anyone can offer some suggestions on different events that may result in liquids forming in flare headers, I would very much appreciate it.

P.S. I should mention that I have taken into account cooling of the releif stream from cool ambient cnditions and have still not encountered liquid formation.

 

[MortenA]

Based on you description it seems as if you covered the subject.

Since no liquids forms then the drip leg wont fill.

Is there any chance of carry over of liquids and such?

Best regards

Morten

 

[sjackson]

Liquid carry over does not appear to be an issue. There are several separators that relieve into this subheader but they are equipped with mist eliminators and I have confirmed that they are sized correctly.

A co-worker of mine has suggested that I consider the possibility of vortex formation in a separator during a releiving event. I'm not sure if this is a credible scenario and if so, how would I begin to determine the quantity of liquids that would be drawn out of the vessel by the vortex?

 

[Flareman]

sjackson

It is possible to have hot gaseous relief from a PSV on a process vessel, which condenses as it cools and travels in the line at ambient conditions. This can occur even downstream of a disentrainment vessel.
I have come across a condition in which a high liquid level of light material in a KO vessel was depleted by evaporation into a hot flow through the drum, only to recondense again in the downstream line and be blown out of the flare as droplets.

If your reliefs are all from storage at ambient conditions and don't generate any liquid at the coldest temperature, the only thing you will have to drain is rain water (or steam condensate) coming in at the top of the flare.

A drip leg is actually not good for much else.

You should check some two phase flow literature for final numbers but, as a very rough guide, you can figure that liquid drop out is only a realistic proposition if the flowing velocity is less than (about) 20 fps (10 fps if you want to get a fairly dry stream). Above 20 fps any liquid tends to be re-entrained and carried along by the moving gas. Some of it may still be pushed into a drip leg at the lower velocities but as the stream picks up to (about) 40 fps the liquid will be moving steadily along the pipe with the gas and by the time you're up to 80 fps you can't keep anything from moving out of the end (unless you do some fancy centrifugal-type separation).

My very arbitrary 20 fps should also be a guide for the maximum level which can be allowed in a KO vessel. Too often, I come across drums sized "economically" using API-521, with no thought to the superficial velocity across the liquid at max level, which could be re-entraining liquid constantly (or at least not really disentraining any). This is most common in Oil field separators which can dump saturated gas with large droplets back into the flare stream leaving the drum with subsequent heavy smoking at the flare tip due to the high (and hidden) liquid load.

David

 

[UmeshMathur]

sjackson:

Another possible reason for liquid separating out from pure relief vapor, even as pressure reduces in the direction of flow, is called "retrograde condensation". This happens in cases where the saturated vapor is at a pressure above the pressure corresponding to the cricondentherm temperature. The possibility of this phenomenon becomes obvious when you look at the complete phase diagram for the mixture in question. The phase diagram can be obtained directly using, e.g., the Hysys simulator with the proper equation of state such as Peng Robinson for hydrocarbon mixtures.

For a discussion of the retrograde condensation phenomenon, see T. E. Daubert: "Chemical Engineering Thermodynamics" (McGraw-Hill, 1985). Another good discussion is found in S. M. Walas: "Phase Equilibria in Chemical Engineeering" (Butterworth, 1985).