Vapor Disengagment: Sump volume vs self venting pipe?

[gte682n]

Hello,

I am currently evaluating options to stabilize a sidestream draw off of a vacuum distillation tower. The sidestream gravity drains to a buffer tank before being pumped to the next unit. The engineering firm has proposed to increase the residence time at the draw sump inside the column by an order of magnitude to assist with vapor disengagement. Current sump residence time is ~5 sec. As part of this option, the liquid level will be raised in the sump above the nozzle by installing weir plates around an over flow pipe that goes down to the packed bed below.

In addition to this option, the sidedraw nozzle and piping will be increased from 6" to 8" to meet self venting flow criteria, Fr# <0.3.

My questions are:
1) does submerging the sidedraw nozzle in liquid defeat the purpose of designing a pipe for self venting flow? how will dissolved gases vent back to the column if the line is full of liquid?
2) if I simply increase the sidedraw nozzle and piping to 8", do I need to increase my liquid residence time for disengagement? Isn't that what the self venting pipe is for?

As always, your knowledge is appreciated.

 

[georgeverghese]

Cannot imagine how you'd get decent vapor disengagement on a tray side draw no what what you do, given the violent aeration that occurs on trays - the residence times are in any case too low.

What happens at this buffer tank ? Can you not install a vapor return line from this tank back to the column tray. Else a external vapor liquid separation drum will be required at the tray elevation. Not clear what you mean by " stabilising the side draw " - is this liquid side draw feeding a booster pump ? Or find a pump that has a greater tolerance to vapor in the liquid feed stream.

 

[gte682n]

Hello georgeverghese,

The buffer tank does have a 2" vent line back to the column. I am unclear the basis for the sizing.

We have difficulty maintaining a steady side product flow rate. When on level control in the buffer tank, there are occasions when the flow out will decrease to maintain level because the sidedraw from the column is fluctuating. We would like to ensure there is nothing restricting free gravity flow from the column to tank. From the buffer tank it is pumped to another larger tank. There are no issues with cavitation on these pumps. We would just like to maintain a steady flow out of the column.

 

[georgeverghese]

Think I have only a partial picture of the setup here, so can we post a sketch of this setup with process controls if you think it will help.

Is this buffer tank at grade, so you'd have level control at this side draw sump that operates an LCV on the feed to this tank ?

Else, for a submerged liquid draw, the Froude number criterion is usually applied for coarse vapor liquid separation. See Perry Chem Engg Handbook 7th edition page 6-28 for narrative and formula for downflow pipes.

 

[gte682n]

Hello, I have attached a sketch of the existing side draw configuration. The buffer tank is not at grade. It is about 15 ft above grade. The LCV is on the outlet of the buffer tank, feeding the next tank.

 

[georgeverghese]

Okay, the sketch makes things clearer now..

From the fig in Perry and the side draw exit orientation on your sketch, you've got an overflow pipe with L/D at 60 for 8inch downflow pipe and 80 for 6inch pipe. Check that you meet the min required h/D criterion for nozzle submergence in fig 6-30 and the limit on Nfr < 0.31 for downflow velocity.

So you've got the level in the buffer tank reflecting level in the overflow sump on this tray. Would suggest we tune this LIC to enable the buffer tank low level alarm to meet the min submergence as indicated in fig 6-30. Also the control volume in this buffer tank should be adequate to allow for a reasonable control response time from this LIC. Check also the LAH setting at this LIC to avoid hydraulic flooding on this tray.

The downflow line and the vapor return lines should be piped up to avoid any local pockets i.e these lines should be free draining.

I dont see any other issues from what you've said so far.

 

[georgeverghese]

Ideally, the control volume between LAH and LAL in the buffer tank should enable a residence time of 2minutes (with controller setpoint at midpoint). At the very least, 1minute between LAH and LAL may be okay if you've got good controller tuning and there is little fluctuation in the side draw flowrate.

 

[gte682n]

Hi georgeverghese,

Thank you for your responses. I was finally able to get my hands on a copy of Perry's and look at figure 6-30. One item was unclear to me: h in h/D refers to total liquid height or only the height above the top of the nozzle?

Thank you again for your replies.

 

[georgeverghese]

To me, h would be the minimum permissible liquid height above the top of the exit nozzle.

 

[georgeverghese]

From what you've said so far, it doesnt sound like the level in the buffer tank is identical to that in the side draw sump in the column. Some static head builds up in the side draw line within the dip pipe in the buffer tank to account for the diff pressure created, which is the sum of the friction pressure drop in the return vent pipe + the draw tray dp, and this is not enough to back the liquid up into the sump even at design flow.

If I've guessed right, then the min liquid submergence fig 6-30 wont be applicable, since the draw sump is dry and the draw pipe is not notionally liquid full. You'd only use the max permissible Froude number criterion then to enable adequate self venting in this draw line.

 

[aje]

It looks like the vent from the buffer tank goes below tray T3 when the drawoff comes from above T3. There is a pressure difference (equal to the pressure drop on the tray) that may hamper liquid flow. A dip leg in the buffer tank would help.