Flare seal pot sizing

[TD2K]

I'm trying to size a flare water seal pot and am getting rather larger numbers that I might expect (on the other hand, I've never had to size one so larger is subjective).

The approach I used is based on preventing droplets over a given size being carried overhead by the rising gas stream in the seal pot. The seal pot is vertical, the incoming line enters through the top head and extends down into the water. The outlet gas line exits from the shell near the top.

1. For my relieving conditions, calculate the drag coefficient for liquid droplets using the method in API 521. I used 600 um, that being the larger end of the 300 um to 600 um a flare KO drum is sized for (and that may be my problem).

2. Calculate the volumetric flow of gas through the pot at relieving conditions. Add in the area occupied by the incoming flare line to the water seal. Convert the area to a diameter.

Sounds right? Am I missing something obvious? Comments appreciated.

 

[monaco8774]

If I understand the aplication correctly, you are talking about a seal in the flare line after the flare knock out drum intended to stop back flow of oxygen into the flare drum. As such I assume you have a vent not a flare ?

If this is the application, you do not size the seal to avoid emptying during full flow relief, as this will normally be close to 0.8 Mach velocity on a properly sized vent line. what you do is design the seal to refill after a relief event and ensure that it is tall enough to prevent the liquid being lost in normal operation (i.e. 10m of water for every bar operating pressure in the flare.)

The diameter of the seal is the same as the vent line.

If it is indeed a flare and not a vent then seal sizing is based on a continuous gas purge to flare through a propriety device such as a coanda seal which is designed to ensure that there is no reverse flow in normal operations.

 

[TD2K]

Thanks for your comments. This is on a 30" flare line, not a vent line. The intent of the seal pot is to hold a backpressure on the flare for day to day operations for the flare gas recovery compressors, about 3' of water.

Your suggestion that the seal pot is the same size as the flare line is roughly the approach suggested in API 521. However, I've also seen at other plants where the seal pots (also downstream of the main KO drum) are more in line diameter wise with the approach I've considered.

 

[Flareman]

TD2K
Sorry I've been out of circulation for a while and just noticed your post.

I have a number of comments, for what they're worth. Your basis is logical and correct but destined to give you enormous numbers because anything over (about) 20 fps is moving most of the water along with the gas flow. 45 fps definitely is heavily laden with water. (2 phase flow principles) Most flare lines are sized at about 350 - 400 fps for design flow so 5% flow hits the 20 fps mark. To handle 100% flow and stay at 600 mu, you're looking at a drum perhaps 6 or more times the header diameter. Could be big !!

You have to decide how the vessel is going to operate once it blows.
Are you happy to poke all the water down the line to the flare. Stay small diameter.
Do you want to catch all the water, big diameter.
In the long run it's the $$ which catch up with you and you compromise.

Another issue, if you hold a back pressure of 36" of water you need to look at the presure design for the flare itself.
Let's suppose that under a 100% design case, the flare generates a back pressure at your seal of 6 psig (= 166 ins wg) When the seal blows, regardless of what you want to happen relative to flow, it's likely to run off at a rate of sqrt(36/166) = 47 % of the design rate which is more than 5% so all the water goes down the line.
[And the flare smokes like a S..B because theres not enough steam for a 47% rate. (just a thought)]

When the seal blows, and all the water dissappears, the back pressure goes to zilch and the GRC shuts down.

I find that one has to examine all these issues of dynamic as well as static water movement.
If you have a big head to hold and you want to keep it in place, you need a big volume (diameter) immedately down stream of the water. YOu can sometimes fool the water by making it surge into a separate volume which is not in the line of the flow and feed it back slowly, or use the upstream pressure to gradually displace the water into an adjoining volume so that there's not much to shift when the pressure goes high (not actually so easy but possible). You can use baffles inside the drum which direct the gas flow into a path which guarantees to clear the water from the path but doesn't entrain so much that you lose it all in 2 seconds. You need to be a little inventive. You have a side mounted discharge so you may have the opportunity to wheedle the flow into a cyclone which separates the water before it gets to the discharge (like a steam drum drier).

Cheers
David