(Somewhat unexpected) Pressure Rise 2

[zaphod1]

I have a parts washer system that uses a vertical centrifugal pump to deliver a heated, 97% water solution up to the top of the washer through a 3” pipe, then distributes (branches) into two 3” pipes. One of these pipe branches delivers water to 21 nozzles, with an equivalent orifice size of 0.203” each, that spray down on to parts passing under the nozzles. The other pipe delivers a similar system that sprays the water on the under side of the part. The pump pressure on the down-stream side of a filter is around 12psi. On the branch that feeds the nozzles spraying downwards the pressure is consistently around 0.25 to 0.5 psi higher. Coming from the pump the water raises around 5 feet, then after passing through a horizontal filter, the pipe drops approximately 18” then branches and drops a further 18” to the downward nozzles and approximately 36” to the upward spraying nozzles. The upward facing nozzles are typically around 10 psi. What could cause a pressure rise, other than gravity, in the branch that feeds the downward facing nozzles?
Thanks,

 

[katmar]

A rough calculation makes me believe that the difference you are measuring is too large to be explained by velocity head effects (Bernoulli) and I cannot think of any other reason. Were all the measurements done with the same gauge, or could it simply be a calibration problem on your gauges?

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[zaphod1]

Hi,
I have installed 0.25% FS accuracy (30 psi span) transmitters, I also see this effect on the less accurate dial-type gauges that are installed at these locations.
Thanks,

 

[danw2]

Interesting. I'm not surprised at the higher pressure on the downspray header, but at the lower pressure on the upward spray header.

The inaccuracy of the transmitters is only 0.075psi. Mechanical gauge accuracy depends on model and can be as bad as 3% (0.9 psi on a 30 psi gauge)

It appears that your 12 psi measurement is at the highest elevation point.
The 12 psig is downstream of the filter. The filter location must be at the top point of elevation since "coming from the pump the water raises around 5 feet, then after passing through a horizontal filter". Everything else drops from that point.

The downward spray nozzles are 36" below (18" drop + further 18" drop) that top elevation measurement point.

The upward spray nozzles are 18" below the downward nozzles.

Gravity adds a psi for every 27.7" of elevation, and you have a 36" elevation drop to the downward spray nozzles. So an increased pressure makes sense.

But the upward spray nozzles are 18" lower in elevation than the downward nozzles, yet are 2psi lower than supply.
I'm still scratching my head. Have I missed something in the sketch?

Dan

 

[25362]

Flow distribution in dividing (as well as in combining) branched manifolds may be nonuniform, affecting static pressure measurements due to the different impedance characteristics of each branch.

 

[KiwiMace]

i'm with 25362, in that your piping network balancing is causing the difference. Put a header between the two nozzle branches sized to give negligible pressure drop and feed it centrally.

 

[katmar]

I did not pick up that the first pressure tapping point was at the highest point - thanks to Dan for the excellent sketch which enabled me to see the true setup.

The pressures read at the two spray headers are lower than they would be if static head was the only consideration. Obviously there are friction losses after the first pressure point, and these losses explain why the header pressures are lower. Without knowing in detail the pipe lengths and numbers and types of fittings it is impossible to comment on whether the losses observed are in line with expectations.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[BigInch]

Hydraulic equations are not as accurate as many people think. It is not unusual to find differences in the predictions vs actual readings of up to 10% in some cases, even when the exact dimensions, pressures and flows are known in each branch. Here you only give a few pressure points, so I wouldn't be too worried if I could not explain 2 psi at this stage.

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[prex]

If the velocities in your system are quite high (well over 2 m/s or 7 fps), what you see could be explained by a velocity head effect, but of course you should have about the same pressure in the two sides of the branch, if the flow rate divides more or less equally (this point is not clear in your description).

prex

http://www.xcalcs.com

Online tools for structural design

 

[zaphod1]

Thanks all for your excellent posts. Thanks danw2 for the great graphic. The graphic is essentially exact, except for the locations of the gauges. They would be located in the pipe before transitioning to the nozzle assemblies. I also failed to mention that in the upper assembly (nozzles spraying downward) there is a 3x2" reducer and 2" valve that is very near the transition to the nozzles. This pipe remains 2" until the transition to the nozzles and the PI and PT are located downstream of this reducer. In the lower branch there is a 3x2-1/2" reducer a 2-1/2" valve and then a second 3x2-1/2" reducer (transitioning back to 3") that is around 36" from the transition to the nozzles. The PI and PT are located downstream of this assembly. I'm not sure if the oem was trying to use up spare valves or if this is all by design.

Thanks

 

[Latexman]

Please confirm or correct.

The upper PI and PT are in 2" line prior to any spray nozzles, and the lower PI and PT are in 3" line prior to any spray nozzles.

Good luck,
Latexman

 

[BigInch]

Uh...
2" valve 2-1/2" valve
2" pipe 3" pipe
3x2 reducer 3x2-1/2 reducer,
and we're looking for 2 psi? Its in there somewhere.
Adios muchachitos.

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[zaphod1]

There are actually three PIs and PTs. PI 103 and PT103 are directly after the filter and PI 104, PT104 is upstream of the upper nozzles, PI 105, PT 105 is upstream of the lower nozzles. I hope this helps.
Thanks,

 

[Latexman]

Picking a 0.19" free passage Bete spray nozzle at randon, it'll flow 21 gpm at 10 psi. 21 nozzles like this would need about 400 gpm. 400 gpm through 2" line would have significant higher velocity head and lower static pressure than through a 3" line!

This may not be exactly what you have, but it's probably in the right order of magnitude.

Good luck,
Latexman

 

[katmar]

This is starting to make sense now.

The gauge after the filter reads 12 psi, and if the 36" drop down to the downward spraying header added its full static head the gauge on this header should read 12+1.3 = 13.3 psi. However, it actually reads between 12+0.25 and 12+0.5 psi. Call this pressure 12.3 psi. We expect 13.3, but actually get 12.3, so we have losses of 1 psi.

If I calculate the flowrate through a 3x2 reducer and a 2" globe valve to give a 1 psi pressure drop I get about 45 gpm. If the valve is a ball valve rather than a globe the flowrate would have to be 170 gpm to give 1 psi loss. Both of these estimates are way lower than Latexman's 400 gpm, but I suspect that Bete's definition of "free passage" is the largest particle size that can pass through, rather than the orifice size. You would never get 21 gpm through a 0.2" orifice at 10 psi.

Unless we know exactly what the length of 3" and 2" pipe is, the number and size of all the fittings and the type of valve we cannot give a definitive answer. But we can say that the answer looks to be in the realms of the possible.

Katmar Software
Engineering & Risk Analysis Software

http://katmarsoftware.com

 

[zaphod1]

Thanks all for your replies. I have suspected gravity or static head, but since I am no expert I value your assistance.

Best Regards,

 

[Latexman]

Katmar's right. A 0.22" orifice /0.13" free passage spray nozzle will flow 4 gpm at 10 psi. It's starting to add up now.

Good luck,
Latexman

 

[quark]

I think I am missing a point here. My basic understanding is that, in liquids, the pressure gauge reading indicates the pressure drop that will occur after the point where the gauge is installed.

The pressure gauge after the filter should indicate the resistance offered to the flow by the pipe section after the gauge. If the elevation is supplementing the the pump pressure, that should replicate in the gauge reading after filter, unless there is siphoning effect. Otherwise, I am not getting any reason for the filter gauge reading to be lower than that of header gauge.

May be I am wrong.

 

[25362]

Indeed the pressure readings on the branches are supposed to show the friction drops through their downstream spray nozzles.

Since the lower upward-spraying header shows 10 psig while the upper downward-spraying header shows 12.25-12.5 psig, assuming type, degree of cleanliness, and number of nozzles, are equal, it could mean that the flow on the upper branch is greater. Which would confirm the original assumption of flow uneven distribution.

 

[25362]

Quark, on gas-free liquid flows, beside the downstream friction resistance effects, the gage reading after the filter could be greater or lower than those of the downstream gages depending on whether the flows are upward or downward (due to gravity), and the pipe diameters don't change velocity heads into pressure heads. I think we talked about this in past threads.