Small diameter water nozzle efficiency versus nozzle length

[Andreas Gjestvang]

Hello!

We are 3 students working on a bachelor thesis rigth now, with the purpose to determine the efficiency for different nozzle designs. We have done several practical tests, where we have logged the water pressure and the flow out from the nozzle, and calculated the nozzle efficiency based on the following equation:

Three of the nozzles we have tested is as following, all with nozzle diameter 1 mm:

Nozzle 1:

Nozzle 2:

Nozzle 3:

The tests gave us surprisingly results. All three nozzles is tested is tested with 200 bar water pressure, and the measured average speed is about 120 m/s. That gave us the following efficiencies:
Nozzle 1: 37 %
Nozzle 2: 47 %
Nozzle 3: 63 %

That will say, the voumetric flow rate would increase as the nozzle leangth increase.

We thougth that the nozzle efficiency would decrease as the nozzle length increase. But as you can see, that did not happen, the results says the opposite. We have also done the same tests with 2 and 3 mm nozzle diameter, and found that this effect would decrease as the nozzle diameter increase. For the 3 mm nozzle, the efficiency is decreasing as the nozzle length increase, which is opposite to the result from nozzle 1.

We are currently a bit stuck on this. Is there anyone out there who have a explanation or a theory around this phenomenom?

Thank you in advance for all replies, we appreciate all inputs!Andreas Gjestvang
NTNU Gjovik

 

[Kevin McHugh]

Probably has to do with shear. The longer nozzle creates more back pressure as resistance to flow is increased. The more contact with the inner wall of the nozzle increases shear, especially with the longer land length ans smaller ID.

 

[Andreas Gjestvang]

EDIT:
The average speed for nozzle 1 is 120 m/s
The average speed for nozzle 2 is 137 m/s
The average speed for nozzle 3 is 159m/s

Also, the water is flowing from left to rigth, with the small diameter bore as an input.

Probably has to do with shear. The longer nozzle creates more back pressure as resistance to flow is increased. The more contact with the inner wall of the nozzle increases shear, especially with the longer land length ans smaller ID.

So it is resonable that increased shear will increase the speed out of the nozzle, and thereby the nozzles throughput?
Thanks, that will be studied further!

 

[JBA]

Are you discharging to the atmosphere or a down stream fluid?

 

[Kevin McHugh]

EDIT:
The average speed for nozzle 1 is 120 m/s
The average speed for nozzle 2 is 137 m/s
The average speed for nozzle 3 is 159m/s

Also, the water is flowing from left to rigth, with the small diameter bore as an input.

So it is resonable that increased shear will increase the speed out of the nozzle, and thereby the nozzles throughput?
Thanks, that will be studied further!

Actually, it is the increased back pressure that increases the velocity, and the increased velocity increases the shear.

 

[CarlHauk]

Are you discharging to the atmosphere or a down stream fluid?

We are discharging into the atmosphere.

Actually, it is the increased back pressure that increases the velocity, and the increased velocity increases the shear.

Thanks we will study that closer!

Carl Helge Bacus Haukås
NTNU Gjøvik

 

[Andreas Gjestvang]

Actually, it is the increased back pressure that increases the velocity, and the increased velocity increases the shear.

Thanks! By the way, do you know any good litterature about this subject, essentially about smaller diameter fluid streams or nozzles?

Andreas

 

[Kevin McHugh]

Not really, my knowledge is based on 20+ years of plastic extrusion. Different material, but the physics are the same. BTW, there is a point of diminishing returns. Too much back pressure can cause non-laminar flow which will ruin efficiency. Also, you run into problems forcing the material through smaller and longer nozzles. For plastics too much back pressure will destroy physical properties.

 

[Andreas Gjestvang]

Not really, my knowledge is based on 20+ years of plastic extrusion. Different material, but the physics are the same. BTW, there is a point of diminishing returns. Too much back pressure can cause non-laminar flow which will ruin efficiency. Also, you run into problems forcing the material through smaller and longer nozzles. For plastics too much back pressure will destroy physical properties.

Thanks!
Another thing i forgot to say:
The water pressure on the nozzle inlet is 200 bar, and we are discharging it into air with atmospheric pressure.

 

[Nidum]

(1) The length of the 1 mm bore is not the only thing that is different in each of the three nozzles .

(2) Calculate the theoretical flow velocities and see if there is any correlation with your measured ones .

(3) Please describe your test set up and in particular describe how and where you are doing the velocity measurements .

 

[Andreas Gjestvang]

(1) The length of the 1 mm bore is not the only thing that is different about the three nozzles .

(2) Calculate the theoretical flow velocities and see if there is any correlation with your measured ones .

(3) Please describe your test set up and in particular describe how and where you are doing the velocity measurements .

(1) The plan was testing nozzles with three different lengths: 3 mm, 10 mm and 20 mm. In the product the nozzles is being used, the material thickness is 20 mm. We wanted to test how the nozzle length would affect the efficiency with the lengths described above. Because of design issues in our test equipment, the total length for the nozzle had to be longer than 20 mm, it ended up with 34 mm length. Our theory was then that the biggest bore (Ø12mm) would give the same result as if the nozzles was placed in a 20 mm thick steel sheet.

I know nozzle 1 and two have 3 different bores, but nozzle 3 only have 2. Like mentioned in last pharagraph, we thougth that the big bore would compensate for the extra length of the nozzle.(2) We can also calculate the theoretical average velocity out from the nozzle with Bernoullis equation, let us take nozzle 1 at 200 bar water pressure as an example:

The average measured velocity with nozzle 1 and 200 bar water pressure is 120 m/s.


(3) The test system mainly consists of the nozzle, a double acting hydraulic sylinder, a hydraulic accumulator and a hydraulic pump. The accumulator is being filled with oil at a desired pressure.
The hydraulic cylinder contains a oil-chamber and a water-chamber. The nozzle is mounted in the water-chamber outlet, and the oil chamber is connected to the accumulator with a hydraulic hose with a ball valve between.

Before the test starts, the water chamber is being completely filled with water. Then, the ball valve on the oil site is being opened. The oil pressure inerts a force to the cylinder piston, that makes the piston start moving. The nozzle is limiting the water flow out from the cylinder, and a water pressure is being build up inside the cylinder.

During the test sequence, the water pressure and the movement of the piston is being recorded by a datalogger module. With the cylinders movement we calculate the actual cylinder velocity in an actual time intervall. The volume being moved by the cylinder piston is equal to the volume being pressed out of the nozzle, and thereby we calculate the average water velocity out through the nozzle in the actual intervall. This velocity, together with the actual water pressure in the same time intervall, is being inserted into the equation for efficiency:

Then we have the efficiency in each very small time intervall (0,02 s). This efficiancy at each intervall plottet with time as x-axis is showing that the efficiency is approximate constant through the whole test sequence, and therefore is the average efficiency through the whole test sequence presentable as a result.

 

[CarlHauk]

We also have a small webpage for our thesis. Some Pictures are a bit irrelevant, but look at image 6 and 13 of the testsystem http://dysetest.weebly.com/bilder.html



Hope Andreas and my reply is any help!

 

[Nidum]

Interesting to look through your website and Video on this project . Nice to see that you are actually allowed to do some hands on engineering yourselves !

Problem itself though comes down to two questions :

(1) Is there something wrong with the test procedure .
(2) Is there something odd about the flow in the nozzles .

As regards (1) Try testing the flow rate using a very direct method to confirm your data logger results . Simple discharge into a measuring cylinder maybe .

As regards (2) Try doing a more complete theoretical analysis of the flow in the nozzles taking into account fluid friction and the changes in flow area .

Come back if you do not know how to do this .

 

[CarlHauk]

The thesis is to be handed in 18.may, so right now I don't think there will be time for discharging into a measuring cylinder. Although we will discuss it on monday.

About more complete theoretical analysis fluid flow. Without too much calculation, just reading litterature we don't see friction as the main cause. But we are discussing changes in the flow area. Could you please elaborate more around how we could study the flow changes? Books and articles would be great! besides that, we're setting up a ANSYS simulation to study the flow changes.

 

[256bits]

Have you looked at the flow pattern from the discharge?
The spread pattern could be a cause for this equation,

to be in error, due to a pressure greater than atmospheric being produced in the middle bore.

 

[CarlHauk]

I don't have a image, but there was a protection plate covering the nozzle

. So we couldn't see the spread close up while testing. Afterwards we could see the water had scraped off the painting around a small spot on the cover though! But there's not to hide that when we tried one of the last nozzles water had scraped off the paint slightly to the side of the earlier spotlocation.

Afterwards we agree that there should have been a see through cover, but we're working on what we got :) Grateful for all answers that can help us!

 

[Nidum]

to be in error, due to a pressure greater than atmospheric being produced in the middle bore.

I've had this thought but it has not been easy to come to any definite conclusion .

How difficult would it be to make three new nozzles with the test bores on the exit side and with large diameter entry bores ?

or

Modify the existing nozzles so as to have single large diameter exit bores in each ?

 

[Nidum]

You mention 'in the product' in one of the earlier posts . This suggests that you have to take the nozzle design as given ?

Two possibilities :

(1) If you are constrained to use the given nozzle design and you are confident that your measurements are accurate then really you have obtained a valid set of calibration data .

(2) If you can alter the nozzle design then you might be able to get better flow characteristics .

 

[256bits]

I've had this thought but it has not been easy to come to any definite conclusion .

How difficult would it be to make three new nozzles with the test bores on the exit side and with large diameter entry bores ?

or

Modify the existing nozzles so as to have single large diameter exit bores in each ?

Well, I don't know nozzles like the back of my hand, but the discharge coefficient for the L/d ratio seems to have something going on.
Is it that P2 is not atmospheric?
Does cavitation play role, but one would think that effects should extend to longer L also?
Or hydraulic flip ( air entering the bore ), which might throw the pattern to one side, or alternate?
Are vena contracta effects become more predominant for a low L/D where A2 is not the bore area but less than that?

Throwing ideas out here that may or may not be applicable.

Such a simple device, but yet complicated.

 

[CarlHauk]

You mention 'in the product' in one of the earlier posts . This suggests that you have to take the nozzle design as given ?

Two possibilities :

(1) If you are constrained to use the given nozzle design and you are confident that your measurements are accurate then really you have obtained a valid set of calibration data .

(2) If you can alter the nozzle design then you might be able to get better flow characteristics .

The thesis is too calculate energyloss for a subsea shock absorber where they use very simple nozzles. The designs are given and I guess we could just pick a number and say that's the answer, but we ofcourse want to explain the reasons behind the results. Better flow designs might be of interest, but production costs are relevant. 10 of the 13 nozzles which we tested is just done too the performance of different nozzle designs. We thought analyzing wouldn't be to timeconsuming, hopefully we have cleared up all the unknown soon.

Well, I don't know nozzles like the back of my hand, but the discharge coefficient for the L/d ratio seems to have something going on.
Is it that P2 is not atmospheric?
Does cavitation play role, but one would think that effects should extend to longer L also?
Or hydraulic flip ( air entering the bore ), which might throw the pattern to one side, or alternate?
Are vena contracta effects become more predominant for a low L/D where A2 is not the bore area but less than that?

Throwing ideas out here that may or may not be applicable.

Such a simple device, but yet complicated.

Throwing our ideas is great, we are reading up on vena conctracta and hydraulic flip and we think it's relevant. Not quite sure how much of an impact the cavitation and hydraulic flip gives though, do you have any sources we could read? We're waiting to do some ANSYS simulations, just our schools server is down at the moment.

P2 should be atmospheric, we are discharging into air.

 

[Nidum]

For the nozzles with three bore diameters two different flow conditions could exist in the intermediate bore :

(1) Jet doesn't expand much and goes clean down the bore centreline without touching the walls .
(2) Jet expands rapidly and contacts the bore walls so that some part of this bore is full of water .

Different calculations needed for each case .

 

[256bits]

Throwing our ideas is great, we are reading up on vena conctracta and hydraulic flip and we think it's relevant. Not quite sure how much of an impact the cavitation and hydraulic flip gives though, do you have any sources we could read? We're waiting to do some ANSYS simulations, just our schools server is down at the moment.

P2 should be atmospheric, we are discharging into air.

It is an area of study for nozzles, as the couple of sites show.

http://miedept.mie.uic.edu/lab/aggarwal/papers/2010-12/Som_Cavitation_JEGTP.pdf
http://www.iust.ac.ir/ijae/files/site1/user_files_62fop6/eng/hakim-A-10-63-93-3b847ef.pdf
http://www.aem.umn.edu/people/faculty/joseph/archive/docs/955-Orifice.pdf
where some of the information is new to me also.
How applicable is all that to the nozzles in question.
I am not sure if you have the time to do something really thorough for your study.
And that I do not have tunnel vision, and are leading a path down a rabbit hole.
In any case, you should now be aware that high pressure nozzles have some interesting flow features.

Since we do not have a high speed camera, an x-ray machine, or a see thru nozzle, at this point one would have to rely upon a simulation and calculations.
Is the difference in pressure great enough for cavitation to be a problem, or can it be ruled out completely.

That is one of the reasons I asked if you could see the flow from the nozzle, if it was misty or a jet.

Which by the way seems to tie in quite nicely with the post from Nidum.

 

[CarlHauk]

Thanks for all the replies, yesterday we finished more tests with different nozzle diametres hoping to be able to explain how efficiency varies with increasing diameter. If you want to see the flow there's 2 new videoes on our webpage: http://dysetest.weebly.com/blogg nozzle type in youtube titles