air flow through into network of pipes 4

[opticsman]

hello , I am interested in air flow through pipes but as it isnt my field I would appreciate your comments. I am trying to understand what happens if I start off with a large diameter input pipe and split it into say 10 smaller pipes. The flow through the large pipe is say 10 litres per second ( air velocity is 10 metres /sec) . If one ignores practical losses (friction etc) will the flow through the smaller pipes be 1 litre /sec (velocity 1 metre /sec)? Would it make a lot of difference If the smaller pipes are a) end-on (in line) with the larger pipe or b) they are perpendicular to it, with holes in large pipe and with it blocked-off at the end? In the latter case would the holes/pipes need to be progressively larger as they get the end ( of the 10 holes) to obtain the same air flow through each small pipe? Again I am concerned about ball park values ignoring small losses etc. Also what would be the practical difference if the large pipe was tapered? Many thanks in advance. Opticsman

 

[BigInch]

The best way to get an even split is to run the main into the center of a "T" shaped header, with the main line entering the header as does the leg of the "T", and all outlet lines branching from the cross bar of the "T" at equal distances. The crossbar of the T should be of a larger diameter than the leg to try to get all outlet pressures into the branches as equal as possible.

Trying to take equal flows off a progressively decreasing diameter of pipe becomes more complicated than the T "header" arrangement, since flow from each outlet would be roughly dependent on the square of the pressure at each outlet. Pressure loss within each progressivley smaller diameter would tend to increase considerably with each reduction of the main and outlet flow into each branch would also drop in proportion to the square of the pressure at each individual outlet. It can be done, however usually a "T" header arrangement will give the best of equal flows and also provide a more efficient use of piping material as well, since you might have to buy up to ten different pipe diameters and fittings, and it is quite possible that one of the sizes you need will not be in the standard table anyway. For example, if the lead in main was 10", the next size down would be 8" diameter, but since the 8" only has 64% of the area of the 10, that would be too much diameter reduction for stripping only 10% of the flow from the main. That 10 to 8" diameter reduction might try to force 30% to 50% of the flow off to the first branch. Trying to hold the stripoff to 10% might require a 9.5" diameter, a 9" or an 8.5" diameter, but those are not standard pipe sizes and won't be available. Even if you could do it with something like 10-9-8-7-6... its a lot of work for the engineer to design, the draftsman to draw, the purchasing department to buy, the warehouse to receive and stock and keep track of all the material certificates, the constructor to install and the maintenance department to stock the spare parts for all the different size valves, etc. It would be far more practical and efficient use of material to keep the configuration simple and use a 10" leg, into a 12" header and have all the rest, 2" outlet branches... for example. Anyway, that's how'd I'd do it.

http://virtualpipeline.spaces.msn.com

 

[BillPSU]

What kind of an environment are you talking about? Is the usage of air a continuous process or is it intermittent such as driving impact wrenches, grinders, and other various air tools? Typically in a general manufacturing environment the concern is air pressure drop and variability in usage becomes a problem. The current air system plant designs create a loop around the building and drops from the loop. The loop effectively creates a reservoir of air allowing the usage variable to not impact the delivered air pressure. The loop design allows greater usage further away from the compressor. The problem with this approach is the initial higher installation cost. Your proposed system of reducing the size of pipe the further you move from the compressor makes excellent sense if the usage is consistent and you properly size the pipe. My warning would be if new equipment is installed and is a large consumer of air and close to the end of the run you would have to replumb your air system because of the new equipment.

 

[davefitz]

If the inlet header is of constant diamerter and the small pipes discharge to atmosphere, then the pipes near teh dead end will have more flow than the small pipes at teh feeder end. This efffect is due to the fact that the static pressure will be higher at the dead end.

If the small tubes are dischrging into another outlet header, then these static pressure effects cancel out if the in and out headers are arranged in a "u-shape".

If you want to cancel out the static pressure effect in the case of the small pipes discharging to atmoshphere, then you would continuoulsly neck down the header diameter to hold a constant header axial velocity. This solution was used in the design of the NASA rocket engines- you can see this in the engines on display in the Wash DC air and space museum and also on display in the Houston space center.

 

[Ashereng]

This type of question is comes up fairly often in Eng-Tips. And usually, the answers are "T"s and reducing diameter pipe distribututions. Since this is an air problem, and no mention of pressure, I will assume low pressure. So, here is a left field thought:

How about a wagon wheel distribution header. The main feed pipe comes up through the hub of the wheel, and then the take-offs are the spokes of the wheel. All the spokes can be idientical in diameter since the pressure drop is equal at each spoke.

Just a thought.


"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[BigInch]

"Reinventing the Wheel" Ashereng, that fitting might get named after you.

http://virtualpipeline.spaces.msn.com

 

[Gator]

I seem to recall seeing something similar used for filling high pressure gas cylinders - I think they were called spiders.

 

[Ashereng]

BigInch,

Actually I borrowed the idea from a circular barn that I saw. The feed goes down the center from the upstairs chutes, and the farmer then puts it into the pens that radiate out from the center. He saves walking time since it's not linear. Of course, this is an old barn. Wouldn't really work well with today's barns and hundreds of cows.

I am not sure what it's called (I'm sure I forgot the answwer). Old age sucks.

"Do not worry about your problems with mathematics, I assure you mine are far greater."
Albert Einstein
Have you read FAQ731-376 to make the best use of Eng-Tips Forums?

 

[BigInch]

Humm... cows on a big rotary table... maybe one chute....

http://virtualpipeline.spaces.msn.com

 

[opticsman]

Hello All
Thanks for all the useful comments. I am still not clear on the relative air flow volumes and air flow speeds through the small pipes (Just ball park , ignoring real loss effects , lossless interface etc) where the total combined area of the 10 pipes is the same area as the input pipe. This is for a system in and out into the atmosphere. many thanks Opticsman.

 

[gerhardl]

...when theory,knowledge and consultants fail to give the answer: build it and see if it works!

 

[BigInch]

Well it should be more or less the same, in a T header arrangement, or a hub and spokes, but each branch flow depends on the exact pressure in front of each outlet and the branches outlet pressure. If inlet pressures are equal, outlet is atomospheric, they're all equal, so flows would be +/- equal.

The case of the progressively decreasing diameter would require a lot more care in controlling pressures between each outlet, via diameter or some other means which I see as possible, but requiring much more care in pressure and flow control. Might be good for NASA, but NOT ME.

http://virtualpipeline.spaces.msn.com