Two Phase Feed Distributor for Multi-pass Tray Tower 2

[molechaser]

Are there any good references for two-phase (C3/C4 hydrocarbon) feed distributor design for multi-pass tray towers?

We have a multi-arm perforated pipe distributor proposed but air/water tests and CFD simulations indicate severe liquid maldistribution with one arm (first one) getting little flow because of the high velocity/momentum of the inlet fluid.

 

[Process879]

hope.... you can refer UOP design specification. this is give you good iede for choosing the feed distributor

 

[molechaser]

No, I don't have access to the UOP design specficiation. Can you describe teh features of the 2 phase distributors?

 

[CJKruger]

molechaser, my view is people overcomplicate tower feed distributors, and that actually creates problems.

Assuming you have a 2 pass tray.
- Feed at a location where you have side downcomers.
- Feed pipe comes in over central downcomer.
- Then split in Tee and go to each side.
- A few inches before the side downcomers, split each branch in a Tee again.
- Now you can just stop at these Tees, or provide a distributor (4 in total) for each end.
- The key is that 2-phase flow will split okay in a right angled Tee.
- If you direct feed against downcomers, you have to insulate the downcomers.
- Feed pipe bot is about 2/3 tray spacing above tray, and tray spacing is increased (say by pipe ID+6"), or to allow enough room for a manway.

 

[molechaser]

Thanks for the response

The situation is actually a little more complicated becasue it's a multiple downcomer type tray. I have attached some sketches.

Obviously I'd have to post a lot of information to allow anyone to get involved in a detailed resolution of the problem. I do have some talented fluid dynamics people working on the issues but I wanted to give them some basic concepts for two phase tower feed distributors that have been used in the past. There is some guidance in Henry Kister's book but it is more focused on conventional trays and doesn't quite correspond to this situation.

 

[CJKruger]

molechaser, I have no experience on multiple downcomer trays (have never seen one), so recommend you contact the vendor.

If all else fails, I would consider a H shape distributor.
- Enter till center of tower, split in Tee.
- After a few feet, each side then split in Tees again.
- Install 4 distributor pipes with holes where appropriate.

 

[tkdhwjd]

I am a bit puzzled by the fact that your simulation is indicating maldistribution because of high velocity/momentum... unless you meant the main header velocity.

Based on your drawing, I suspect that you have very low pressure drop from the distributor inlet all the way to the exit. Suggest that you check the pressure drop across the last two legs. Too low exit pressure drop does cause maldistribution. If it is vital to have this configuration because of your tray design and process reasons, then try to lower the perforation/open area of the last two legs.

 

[davefitz]

On subcritical , multi pass once thru boilers, the 2 phase distributors were of teh following type:

a)downcomer to end in a "spider bottle"- if this is spliting to 10 feders, then there would be 10 internal ( pie shapeed) radial dividing plates , and the 10 smaller feeders would be mounted to extract in a horizontal direction.Typially used for 0-10 % steam by wt

b) for higher steam by weight, would use a horizontal distributing header. The risers leaving the horizontal hdr to be located on opposing sicde centerline of theheader, and the inlet feeders to discharge verticlaly down thru multiple nozzles on the top or bottom centerline. Diameter of tis hdr to be minimized - perhaps 1.5 times diameter of inlet feeders.

 

[molechaser]

tkdhwjd

Yes, I was refering to the main header velocity/momentum. The liquid jets toward the end of this header, builds level and forms a wave that rebounds back towards the inlet. Becuause level builds at the end of the distributor the arms at that end do tend to distribute most of the liquid.

Your suggestion of lowering the perforation area in the last two legs is one option that we are exploring. Another is a progressive reduction in diameter of the main header. There is plenty of pressure drop available so it's a question as to whether there's a suitable way of using it to get the right distribution.

Thanks for the response