pipe full of water: opening to the air/opening to water - different results

[Gorostidi]

Hi to all,

We are involved in a project where we are making a first prototype and we found the following:

We have a pipe under pressure full of water and small circular openings. the diameter of the pipe is 50mm and the holes 5mm. we have 5 holes, pointing downwards. The water goes out to the air.
The problem of calculating flowrate per opening and the pressure loss in the pipe, is easy to solve by bernouilli equations tanking into account the adequate discharge and contraction coefficients. we have checked it measuring the pressure drop and everythin is ok.

But now we move the pipe to a tank. And we put the pipe so that the openings are in contact with the water (we insert them a few mm so that there is no signigficant pressure increase "outside" the opeinings). In this case the pressure drop goes down (not much about 8%) and the flowrate increases.

Is there any change of the velocity or in any coefficent that should be taken into account in this case?- when we put it in contact with water instead of coming out to the air?

thanks a lot

gorostidi
phd mechanical engineer

 

[LittleInch]

Clearly there is. Try using an orifice calc for the "flooded" condition and see if that gets you closer to the test results rather than the nozzle approach valid for the air discharge.

There's probably some other coefficient to use for this but try that one and see if it works.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[zdas04]

The Cv of the holes changes with different exhaust media. I don't know of a way to predict these changes mathematically. Typically people who publish Cv values do experiments and publish the values to people who purchase their kit. The interaction of viscosity, surface tension, hole geometry, and relative pressures is quite complex. It is likely that (since the flow increased when submerged) that the controlling factor is the surface tension of the exiting fluid, but I don't know that, just conjecture.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.
The plural of anecdote is not "data"

 

[BigInch]

I'm not much when it comes to very tiny systems like this, but perhaps increasing the resistance to outflow in the immersed condition also increased the diameter of the vena contracta more than enough to compensate for a slightly less exit velocity. Also conjecture...



Independent events are seldomly independent.

 

[hydtools]

There may be the vena contracta, a reduced flow area, resulting in lower flow. Compared to no reduced area vena contracta when flow exits into water. May be worth considering.

Ted

 

[racookpe1978]

You have 5x holes, each hole at 5 mm in a 50 mm pipe may seem like a small difference, but all of that water is flowing out the 5x holes, right?

Your resistance to flow - as noted above correctly - is changing based on the difference between effect of air and water, even if there is very little back pressure involved. It's a spray nozzle type effect: entraining just a little bit of air into a flowing stream through a different shape hole a different place in the nozzle will greatly change the speed and shape of the output spray, even if all of the nozzles on your garden hose have nearly the same area for flow.

You will also have a effect of the flow down the pipe: the final hole - if the same size as the first flow, will have notably less pressure behind it. Each hole will not get 20.0% of the flow.

 

[Gorostidi]

Hi and thanks to all comments:

All comments are very wellcome.

Our idea of what is really happening is very close to what zdas04 and racookpe1978, suggest, that the Cv is for some reason in this case changing. Cc should be the same, as i understand the vena contraction is because of the internal flow inside the pipe. But what BigInch and hydtools say is also reasonable. The fact is that when inserting into water the flow increases instead of decreasing (which is what we expected). In any case it seems complex to prevent.

From the mathematical point of view we will also try LittleInch suggestion. Seems to be the right one.

Our REAL AND COMPLETE experiment is the following:

1) experimentally obtain the cc cv and discharge coefficients. Water inside, air outside. this is very close to any reference found in books.

2) Then put the pipe inside the water (as i described). the results are similar but as i said they change a bit, and we did not undersand the physicall reason.

3) now go the other way round. air inside the pipe water outside, so we inject air into a water tank. (in fact the pipe will be in a tank of a filtering system, where sometimes we insert air, sometimes water). If we take the results taken from point 1) the results are quite approximate to the experimental ones.

We have done it this way because it is very easy to get a very precise data of the flowrate with water, and all the coefficients as described in point 1). On the other hand, with air, we need (and we use) a flowmeter, which is ok, but has some small error. In any case if we use this data in points 2) and 3) we obtain acceptable results but not completely precise.

An error of 5% is acceptable in our development, its ok, but we think that it will be worth getting data from 2) and 3) experiments in real conditions, this is with water outside.

We will go for it and let you know. in any case, new ideas will be wellcome and reference books or articles that have been involved in something similar, i guess we are not the first ones.

in any case from the comments we will make the experiments and the the coefficients experimentally. but it is not easy sometimes to find the real reason of why they change depending on the experiment parameters

thanks

gorostidi
mech eng phd

 

[Gorostidi]

obviously we took into account for the calculations the density of water and air on each case in bernouilli equations... as i said if any has any references to compare to our results will be very usefull for us

thanks

gorostidi

 

[katmar]

I understand that your experiment is not the same as the usual problems of determining the orifice coefficient as a function of the d/D (orifice bore over pipe bore) ratio, but the very complex function of the coefficient relative to the d/D ratio and Re makes it more acceptable that the coefficient should behave differently depending on the downstream conditions. A typical graph for the coefficient as a function of d/D and Re is available at

http://webwormcpt.blogspot.com/2007/10/how-to-apply-valve-equation-in-hysys.html

as Figure 10-20. The curves on this graph cross each other several times.

I suspect that it is mostly related to the formation of the vena contracta, but I agree with David that surface tension is probably a factor too.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[Gorostidi]

thanks for the link.
The experiment looks simple but the coefficents behind are not that easy to get (or find!) in books or articles . We will use this graph as a reference to compare with.

gorostidi

 

[Gorostidi]

Hi to all,

we have made a ver simple experiment:

we take one pipe of 25mm and install put a manometer to measure the pressure just before it. after the pipe we put an elbow open to air. the flow is big enough (the manometer measures about 2 bars) so that the pipe is completelly full

now we just insert the system (just the elbow) a few mm in a tank. the pressure shown in the manometer decreases, from 2 bars to about 1.6 bars. it is the same effect as we described bellow but in a much simpler system.

but the flowrate at the two cases remains the same.

does anyone have a clue of what we are misssing?

thanks a lot to everyone

gorostidi

 

[BigInch]

That only measures static head. There is a velocity head to consider which will probably make a difference at these small scales.

Independent events are seldomly independent.

 

[Gorostidi]

Thanks, that should explain it. But we did another try with a tee instead of an elbow. The water goes out from the mid opening of the tee, and at the other end of the tee, after a piece of pipe we installed a manometer. Same result.

Will do again the tests...completelly lost...

 

[BigInch]

Lost me on that configuration.

Anyway, try calculating the velocities, then including velocity head. At least you might not be totally lost.

Independent events are seldomly independent.

 

[katmar]

It seems unlikely to me that in the experiment with the 25 mm pipe and the elbow that a 0.4 bar change in pressure can result in no flow rate change. What system are you using to deliver the water - pump (type) or header tank? How do you measure the flow rate? What is the flow rate? How long is the 25 mm pipe? Are you holding the pipe at the same angle to horizontal when in air and when the end is submerged? A dimensioned diagram would probably be helpful.

Katmar Software - AioFlo Pipe Hydraulics

http://katmarsoftware.com

"An undefined problem has an infinite number of solutions"

 

[Gorostidi]

Hi to all

finally solved the mystery. everything was ok, except that in we had some air in some parts of the circuit. this caused a wrong measure in the pressure value. when the holes are in contact to the water, the pipe gets rid of the air and the pressure value is ok. as soon as we have made a "closed circuit" then everything goes fine.

really thanks a lot for all your help.