Continue to Site

Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations KootK on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Multiple Y-fitting to split the flow 1

Status
Not open for further replies.

Bozo_Sam

Aerospace
Aug 22, 2019
32
Hi there,

I would like and opinion what would be the easiest way to split the flow coming form a single pump (160 m3/H) and even flow in 8 pipes (each 20 m3/h). the split into smaller pipes is needed for the water treatment that can only be made on the 20m3/h flow and we would like to avoid having 8 pumps but only one.
I made a 3D concept below but would need the advice of what standard parts to use. I am especially having trouble finding Y-split fitting that would split the flow at such high flow rates.

PIPE1_d9gi9h.png
PIPE2_p5mt5b.png
PIPE3_s7zmvg.png
 
Replies continue below

Recommended for you

Just a remark: Shouldn't you avoid installing an elbow right in front of the suction side? You've give no diameters for the pipes, but at 160 m^3 / h I would guess from your model, that there's no uniform flow at the suction side. I stand to be corrected here, too.

Can't help with the rest right now though. But I'm sure someone will chime in.



 
You could try looking at a large header and branch system.

So long as the header is about twice the square area of all the offtakes or bigger, then the pressure loss along the header becomes very small and essentially the start of the branches are all at the same pressure. It helps if there is quite a large (and equal) pressure drop in each of the branches.

Depends how "equal" all the pressure drops are in the 8 tubes and how accurate you need to be. If 5% difference is Ok ( so a range of 19-21) then you should be ok with a header system.

Allowing for some sort of "trimmming" is always a good move, e.g. being able to insert an orifice plate or maybe a slightly smaller ID section pipe into each one to equalise the flow if it gets out of your required accuracy range. Are you metering flow anywhere? Do all 8 all work at the same time?

But yes, the elbow before any splitting is a bad idea. I would include an ISO type flow straightener - e.g. before trying to split anything.

A few more details would be good.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
 
Turn the pump 180[sup]o[/sup] and eliminate two elbows and a bit of pipe.

Eliminate all the Y-fittings and use a single pipe manifold with 8 distributors. Similar to:

Capture_n4jl3h.jpg


Good Luck,
Latexman
 
Nobody is thinking about how you will support all that pipe in such a tightly packed layout. Or how hard it will be to do maintenance on the pump with such limited access. I'd think that the pump mfgr won't allow such weight to be supported by the pump nozzle. A 180 directly in front of the pump isn't qreat either. Is this inside a submarine or something?

Flow control using orifice plates isnt easy. Neither is getting equal flow distributions from a non-symetric pipe arrangement. Inlet should be in the center of a header, not at one or the other end.

Nice drawing though. I think I saw that on the Empire's Death Star.

PS Y's are hard to find, as you already know, and more expensive.
 
I agree with LittleInch and Latexman. Trying to use y-fittings to obtain equal flow won't work as well as you think it will without some other controls. A single large header with 8 takeoffs makes a lot more sense. Do you have any valves in this system? If it's operating under consistent conditions, you could put some manual butterfly valves on each takeoff so the flow could be manually adjusted. Of course, if you might have changing demands in each branch, it would be more complicated, or if you want to avoid introducing additional pressure drop.

In terms of -fortyfive's comments, I don't think the header inlet necessarily needs to be in the middle, as long as the bulk of the pressure drop in this system is in each branch. But certainly you'll need to think about support and maintenance access.
 
Distances are so short between offtakes such that f/ft x ft apx = nothing and it can be considered as a Bernouli only problem. As flow is taken off, velocity in the next section slows, maybe causing pressure downstream to actually rise more than the friction loss from the reduced flow in that section causes it to decrease. An unsymetrical shape makes that very difficult to get balanced. The different elevations of the takeoff pipes may even have a noticeable effect. Lesser flows in the highest 4 pipes, more in the lower pipes. Its hard enough to get balanced flows when the shape is symetric. You tend to get symetricly equal flows, but unequal pipe flows. If you buy 8 valves, can you even fit them into that spaghetti can. Maybe locating them far far downstream might work.

You may get a lot of prerotation from that 180 going into the pump. And a lot of potential leaky points with so so many flanges. I would hate to be the guy that has to tighten all those nuts on the bottom row.

 
Maybe pump some thru'the system, might make a great pipe organ 😉

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)
 
For single phase flow, Latexman is corrrect, use a properly sized header. For 2 phase flow, the wyes may work if internal vanes are included to split the flow prior to entering each wye.

"...when logic, and proportion, have fallen, sloppy dead..." Grace Slick
 
IMO, with the header proposed by Latexman the flow will assume a preferencial way along the time, it will not be perfecty splited or equal distributed through the pipes A,B,C and D.

luis
 
Hi there and thanks to everyone for such extensive help.

I really like the solutions By Latexman and I would be happy to go with this solution. I have a few questions though.

1. How are such parts manufactured? Welded sheet metal?
2. How do we assure the even flow through each of the pipes? Maybe a conical shape of the main pipe of some kind?
3. Would you say that each pipe should have a separate valve?

Thanks in advanced!

 
Well for me,

1) It's pipe and branch fittings or tees
2) you need to make the header much bigger than the branches and keep its length a short as possible. So e.g. 8 off 2" branches you would use something like a 12" or 16" header that's no more than 2m long.
3) So that you can "trim" the flow if you need it equal within a very small amount.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.
 
Yeah. 1 big D header and the same centerline elevation of all pipes would help. Config symmetry and no elbows nearby. Not sure how you'd trim the flow though without a lot of valves.

 
Use Search, under the thread title and between Forum and FAQs, to research the terms "flow distribution" and "maldistribution". Without controls, for the flow in each branch to be about equal, the resistance of all 8 branches must be about equal. There are threads you should find that show how to estimate the maldistribution. 5% maldistribution is usually acceptable.

Good Luck,
Latexman
 
Hi Sam,

I agree with others to employ simple large header with smaller take offs. You have to start and think about process and controls first before having piping system model.
I would recommend to use PipeFlow calculation software, where you can model your piping system with all pipe, valves, fittings, pump and see what kind of flows you can achieve using different size header, branches, elbows, equipment, etc. This exercise will not only give you pretty accurate results, but also how sensitive your solution to piping system variables. You can actually get piping layout equal branch flow without any flow control valves.

Thanks,
Curtis
 
All those flanges are going to disrupt the flow. Not to mention that you might not be able to use conventional tools to install those nuts.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor