Centrifugal pump curve falls short of factory test
Centrifugal pump curve falls short of factory test
(OP)
against the factory testing that is..
This is a drop of a substantial amount. close to 20 gpm difference at 30ft of discharge head which is the lowest our piping system will go.
pump shutoff measured here and at the factory is 47 ft.
NPSH is 2ft, 2.0hp, 1750rpm, pump is a 6.125" impeller, 1" discharge dia, 1.5" suction dia. factory BEP Flow is 55 gpm. our field tested rpms during operation have all been better than 1750rpm. we've changed out the suction piping to a few different configurations tank + valve to a stand pipe and 3" continuous suction piping leading up to the pump. flowrates and pressures have been double and triple checked with calibrated equipment onsite and several different gauges and styles of flowmeters.
I though a pump curve can only change due to loss of rpm or a suction side starvation. We've also vibration tested the pump and checked its impeller clearances. Any help is greatly appreciated.
Sam Boswell
This is a drop of a substantial amount. close to 20 gpm difference at 30ft of discharge head which is the lowest our piping system will go.
pump shutoff measured here and at the factory is 47 ft.
NPSH is 2ft, 2.0hp, 1750rpm, pump is a 6.125" impeller, 1" discharge dia, 1.5" suction dia. factory BEP Flow is 55 gpm. our field tested rpms during operation have all been better than 1750rpm. we've changed out the suction piping to a few different configurations tank + valve to a stand pipe and 3" continuous suction piping leading up to the pump. flowrates and pressures have been double and triple checked with calibrated equipment onsite and several different gauges and styles of flowmeters.
I though a pump curve can only change due to loss of rpm or a suction side starvation. We've also vibration tested the pump and checked its impeller clearances. Any help is greatly appreciated.
Sam Boswell





RE: Centrifugal pump curve falls short of factory test
Mike Halloran
Pembroke Pines, FL, USA
RE: Centrifugal pump curve falls short of factory test
RE: Centrifugal pump curve falls short of factory test
I noticed strange figures in bhp (test 1)
delta pressure=discharge pressure-suction pressure
it is delta pressure or delta head you want to use in a plot
RE: Centrifugal pump curve falls short of factory test
Second guess: Measurement uncertainty. I've collected a fair number of pump data sets in the field, well actually in boats that were rolling and pitching, and it never looks particularly good, though rarely as funky as what you've got there.
Third guess: Database problems. I hope the problem is limited to Caterpillar. Their Web-based software retrieves what are clearly the wrong pump curves for their raw water pumps. It appears to be one of those problems where the data was stored in a zero-based 2D array, but retrieved as if it were a one-based array, or complementarily. So you get _a_ pump curve, just not the right one, except for the ones on the ends of the array, in which case you get ... random data from other fields. I have despaired of them ever accepting a bug report.
Mike Halloran
Pembroke Pines, FL, USA
RE: Centrifugal pump curve falls short of factory test
The figures for bhp in test 1 should be omitted. I was playing with some equations for water flow and trying to get horsepower numbers to match the factory test data.
Delta pressure is what I want to use for the pump curve? I'll use this number, except the factory graph does not seem to use delta head pressure. its using the "total head" figures which seem to only factor a combination of the velocity head plus the discharge head. I'll plot everything again using this method to see if I get some kind of an agreement between the two data sets.
Mike Halloran,
Temperature was checked as well, several different times during the day. we were unable to see major differences between the range of 60-80 degree's f. 1-2 psi differences were somewhat normal from two back to back tests. Though we have never reached the 66 gallons per minute at 30 psi that the factory was able to achieve. in fact our best curves are plotted in test 1 and 2. these are far away from the factory's test. the end point of the factory being 66 gpm and our field testing being 44 gpm the factory has a 150% increase. very unusual numbers they've gotten...
should I be concerned about my piping arrangement or turbulent flow characteristics as the factory is telling me?
I can post the P&ID or mechanical drawings if necessary. its basically a 100-150 ft length of pipe without rise, around 8 2" cpvc 90deg elbows, 2" mag flow meter, two 2" diaphragm valves and a 2" ball valve at the very end going into a tank inlet. Upstream is a 8" stand pipe connected to a 3" header that falls 6 ft below the pump centerline and rises up to meet it again 10 feet away. bell reducer from 3" to 1.5" and into the pump. certified gauges and flowmeters downstream. gauges on both sides are within 1' of the pump and have proper air bleed off. heheh i'm at my wits end... Factory is standing by their testing and has insisted it is not a warranty issue err.. basically they've said there is nothing they can do.
RE: Centrifugal pump curve falls short of factory test
The brake horse power - bhp - for a pump or fan can be expressed as:
Pbhp = ( γ Q h / 33000 ) / η (2)
where
Pbhp = brake horse power (horsepower, hp)
Q = volume flow rate (ft3/min, cfm)
η = overall efficiency
RE: Centrifugal pump curve falls short of factory test
Are your flow measurements accurate? What are you using to measure the flow?
RE: Centrifugal pump curve falls short of factory test
TenPenny, the factory curve is I'll admit a little bumpy but its understandable that the accuracy can be smudged so i'm not knocking that. we have tested the flow with a fill test of a known volumetric cylinder as well as a calibrated rosemount magnetic flowmeter, in addition to these we've used a pitot tube instrument and a 2" inline horizontal flowmeter purchased from mcmaster. We've taken similar readings from all of these (usually they agree within 3-5 gpm). and pressure gauges, we've been through calibrated and uncalibrated, new old and everything in between. I think the interesting bit is that the test data towards the end of the factory test is so far and away from our field tests (1.5 times) that i'm starting to think the error might be in the form of an HCL to H20 conversion. though that would only make up for 1.2 times the difference.
RE: Centrifugal pump curve falls short of factory test
RE: Centrifugal pump curve falls short of factory test
RE: Centrifugal pump curve falls short of factory test
The way the data is presented in the factory test and in your field tests don't make sense. An example:
Factory Test Point 2 - if they measured 42.7 ft absolute that means the suction was flooded to 8.74 ft [42.7 - 14.7 x 2.31)]. The velocity head was 0.2 ft. The discharge head was 45.3 ft. Therefore the total dynamic head is 36.76 ft (i.e. 45.3 + 0.2 - 8.74) and not 45.5 ft.
Field Test #1 Point 3 - the flooded suction pressure is 2.84 ft (i.e. 1.23 psi). The velocity head is 2.86 ft. The discharge head is 38.8 ft. Therefore the total dynamic head is 38.8 ft (i.e. 38.8 + 2.86 - 2.84) and not 41.65 ft
I assume all of the pressure readings have been corrected to the centerline of the pump. Also i note that the pump speed for the factory test is different and the test results have to be corrected to a baseline speed.
Without sounding condescending...did the people who did the factory testing and the people who did the field testing have experience in this field?
Also, i don't know what type of pump this is but the efficiencies sure seem low
RE: Centrifugal pump curve falls short of factory test
Appended is some more information from the Fairbanks Morse hydraulics handbook that will help you out.
In my last email i did not answer your question directly but it is difficult to say who is right and who is wrong. The BHP calculation is a proven equation. If the data that is entered is incorrect then the BHP results are incorrect
RE: Centrifugal pump curve falls short of factory test
also the pump is a centrifugal 6.125" dia horizontal end suction pump. 1" outlet 1.5" inlet
Thanks for your comments
RE: Centrifugal pump curve falls short of factory test
Your mission now is to figure out why your setup isn't giving you what you think it should. Perhaps you could post a piping schematic, there's obviously something wonky.
RE: Centrifugal pump curve falls short of factory test
If I understand it correctly you have massaged the values in the factory test results table???. If so, I would rather see how the pump factory presented the data.
TDH for a flooded condition = reading on discharge pressure gauge - reading on suction gauge + delta v2/2g
TDH for a suction lift condition = reading on discharge pressure gauge + reading on suction gauge + delta v2/2g
I would agree with Tenpenny that the factory test confirmed the catalague curve. It would be best to post a schematic of your pump setup showing where the pressure gauges are with respect to the centerline of the pump. In a table, post the suction pressure gauge reading the discharge pressure gauge reading and the corresponding flow for every point you throttled. Do not do any corrections to the readings. I assume you have calibrated pressure gauges and flow meters and they are reading accurately
RE: Centrifugal pump curve falls short of factory test
It would be nice if you could post a photo of the installation.
Is it possible that with your piping setup that you do not have enough NPSHa, at for example, 45 usgpm? Do the calculation if you have not done it. The tested pump curve will follow catalogue pump curve as long as you have enough NPSHa. If you don't have enough NPSHa The field tested pump curve will start to droop at the point where you just don't have enough NPSHa. The droop will will get worse as the NPSHa becomes more and more negative.
RE: Centrifugal pump curve falls short of factory test
FACTORY TEST
QualityTime, apparently you assume that "Suct. Pres. (ft ABS)" actually means 42.7 psia, but the units are given in [/color red]FEET[/color]. The title of that column should be SUCTION HEAD FT ABS
ALL FACTORY TEST CALCULATIONS APPEAR TO BE CORRECT
however "disch psi" should be "disch press psig".
"disch head" should be "differential head ft"
"total head ft" should be "total differential head ft"
All the numbers are correct, if you assume my definitions.
THE CURVE LOOKS WIERD because you are used to looking at curves plotting differential head, but this is plotting TOTAL DIFFERENTIAL HEAD. Most pump curves do not plot total differential head, they ignore velocity heads. Therefore where the velocity becomes relatively high, you see its squaring conribution to total head as keeping the head at the higher flow ranges than what you are probably used to seeing. When the velocity component is not included in the pump curve, the curve falls off much faster at the higher flowrates.
FIELD TEST 1
I agree with QT and think somebody didn't know what the heck they were doing on this one.
FIELD TEST 1 POINT #3
suction head from 1.23 psiG is 2.84 ft,
differential head from (16.8-1.23) psiG is 35.96 ft
discharge velocity head I take as 2.86 ft,
total differential head is 35.96 + 2.86 = 38.82 ft
(I get the same value as QualityTime there)
Your power calculation is not BRAKE power its HYDRAULIC power. Divide by efficiency to get BHP. I used efficiency values from the mfgr's curve for similar flowrates and found your power consumption is extremely high.
Correcting for HCL assuming it has a SG of 1.2 would make your field test's pressure conversion translate to heads that would be lower than an equivalent water head by 20%, thereby making it worse, ie, not going in the correcting direction and making your power consumption 20% higher.
NOW, what's with the low suction pressures? Perhaps you are cavitating the hell out of this pump, causing the low performance figures you are getting? Nowhere do you ever say what NPSHR is, so I'm not sure, but it looks like a possibility. Your suction pressures are very much lower than the factory's values. A significant portion of NPSHR could be made up from velocity head, so if your's is slower than the factory's there's more of the problem there too.
If this is a new installation, be sure the pump is not running backwards.
"We have a leadership style that is too directive and doesn't listen sufficiently well. The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward CEO BP
http://www.youtube.com/watch?v=hpiIWMWWVco
"Being GREEN isn't easy." Kermit
http://virtualpipeline.spaces.liv
RE: Centrifugal pump curve falls short of factory test
When I did my calculations for Factory Test Point 2, I definitely used absolute ft and not absolute psi. The table showed the suction pressure was 42.7 ft absolute. Therefore the suction was flooded to 8.74 ft above the centerline of the pump [i.e. 42.7 ft - (14.7 psia x 2.31 ft/psia)]. The velocity head was 0.2 ft. The discharge head was 45.3 ft. Therefore the total dynamic head is 36.76 ft (i.e. 45.3 + 0.2 - 8.74) and not 45.5 ft. I think if you look at the one of the questioner's posts he advised that he changed the appearance of the factory data. The way the data is presented does not make sense. If I remember correctly even the velocity head calculation of 0.2 ft is incorrect.
I would agree with you that velocity head for pumps that have high discharge pressure won't make much difference to the TDH calculation. But for pumps with low head it could be very significant. Any pump testing company will include the velocity head calculation when a witness test is done especially because it is the proper way to do the calculation especially if there is liquidated damages that could be applied
RE: Centrifugal pump curve falls short of factory test
Yes I agree here too. Usually Vhd wouldn't make much difference, but here it is an important proportion of total head, so yes it should be included. I just meant that as a comment towards explaining what some have noted as the strange shape of this curve, as it makes for an apparent hump where velocity head increases rapidly, especially in relation to the fall off of the other head component.
"We have a leadership style that is too directive and doesn't listen sufficiently well. The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward CEO BP
http://www.youtube.com/watch?v=hpiIWMWWVco
"Being GREEN isn't easy." Kermit
http://virtualpipeline.spaces.liv
RE: Centrifugal pump curve falls short of factory test
Sboswell has advised that he has manipulated the factory pump test results....that may explain the questionable results that are in the factory test data table. They are questionable because the numbers just don't add up properly. You and I are both in agreement in questioning all of the test data results.
You and I are both suspecting that there is NPSHa/cavitation problems here. Even though the test data is suspect the drop in the performance curve is an indicator of lack of NPSHa which leads to cavitation. Sboswell should do some calculations to check it out as i previoulsy suggested. Furthermore he should have been able to have heard the classic sounds of cavitation when he was doing his field tests.
RE: Centrifugal pump curve falls short of factory test
"We have a leadership style that is too directive and doesn't listen sufficiently well. The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward CEO BP
http://www.youtube.com/watch?v=hpiIWMWWVco
"Being GREEN isn't easy." Kermit
http://virtualpipeline.spaces.liv
RE: Centrifugal pump curve falls short of factory test
"We have a leadership style that is too directive and doesn't listen sufficiently well. The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward CEO BP
http://www.youtube.com/watch?v=hpiIWMWWVco
"Being GREEN isn't easy." Kermit
http://virtualpipeline.spaces.liv
RE: Centrifugal pump curve falls short of factory test
the people on this board are a credit to humanity. I should have joined a lot sooner, thank you guys so much.
-Sam
RE: Centrifugal pump curve falls short of factory test
RE: Centrifugal pump curve falls short of factory test
"We have a leadership style that is too directive and doesn't listen sufficiently well. The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward CEO BP
http://www.youtube.com/watch?v=hpiIWMWWVco
"Being GREEN isn't easy." Kermit
http://virtualpipeline.spaces.liv
RE: Centrifugal pump curve falls short of factory test
BigInch,
we've modified the suction piping with a few different designs at the manufacturers recommendation. we piped the entire line with 3" with a bell reducer going into the pump, where it was 2" before, you can see this in the attached picture on the far pump. the bell could've been set further back but i think we have more than enough distance from the elbow on the original setup (shown in background)
I agree with you that there shouldnt be anything wrong with the pump. would you bet its a factory test mixup.
RE: Centrifugal pump curve falls short of factory test
RE: Centrifugal pump curve falls short of factory test
There is an arrow in the bearing frame, near the coupling guard to confirm.
RE: Centrifugal pump curve falls short of factory test
RE: Centrifugal pump curve falls short of factory test
Factory Performance Tests
I can't figure out the factory performance test results. The TDH calculation does not make sense. Also I notice that the performance test was at SG=1 and the as built performance at rated speed the SG=1.2. What is the SG of your product? But let's assume it is correct.
Filed Test #3 and #4
I corrected your results and you can see it on the appended spreadsheet in yellow boxes. You can see the formulas in the spreadsheet. The equation to compute total dynamic head is as follows:
TDH for a
Flooded condition = reading on discharge pressure gauge - reading on suction gauge + delta v2/2g
The shut in head for the factory test, test 3, test 4 is 46.5 ft, 44.04 ft and 45.43 ft is close enough using the instruments that you have. Both site test curves start to drop off on increasing flow. Have you checked your suction piping for debris or checked inside the pump impeller for debris? The shape of the curve is a classic curve for something blocking the flow in the suction piping
Have you checked the performance for the other pump?
Can you see any vortexes at the pump suction at different water levels
Appended is my spreadsheet that shows how i did the calculations
RE: Centrifugal pump curve falls short of factory test
If it is the one in the foreground with the concentric reducer on the inlet this could be causing some problems with performance.
RE: Centrifugal pump curve falls short of factory test
I am resending this with the spreadsheet and graph. The last version did not have the graph
Factory Performance Tests
I can't figure out the factory performance test results. The TDH calculation does not make sense. Also I notice that the performance test was at SG=1 and the as built performance at rated speed the SG=1.2. What is the SG of your product? But let's assume it is correct.
Field Test #3 and #4
I corrected your results and you can see it on the appended spreadsheet in yellow boxes. You can see the formulas in the spreadsheet. The equation to compute total dynamic head is as follows:
TDH for a
Flooded condition = reading on discharge pressure gauge - reading on suction gauge + delta v2/2g
The shut in head for the factory test, test 3, test 4 is 46.5 ft, 44.04 ft and 45.43 ft is close enough using the instruments that you have. Both site test curves start to drop off on increasing flow. Have you checked your suction piping for debris or checked inside the pump impeller for debris? The shape of the curve is a classic curve for something blocking the flow in the suction piping
Have you checked the performance for the other pump?
Can you see any vortexes at the pump suction at different water levels
RE: Centrifugal pump curve falls short of factory test
RE: Centrifugal pump curve falls short of factory test
What liquid are you pumping? Does it off gas?
RE: Centrifugal pump curve falls short of factory test
RE: Centrifugal pump curve falls short of factory test
Quality time, could the strange data from the factory be explained by the TDH was total dynamic head rather than total differential head? I did not see them taking into account their suction side head (ie they were not subrtracting the number from their discharge head. They only added the velocity head to the discharge head and went ahead and plotted that against their capacity.
Yes, our test medium so far has only been water. we will eventually be running hydrochloric acid (SG 1.2) through the pump. No gassing.
Thank you for weeding through my results, I should have posted them in an editable format. I see what you did there with the delta v2/2g, that spreadsheet is excellent.
Pump #2 in the background runs roughly the same with no difference to the pump in the foreground. I should have stated that earlier. all the changes we have made have done nothing to help our situation out. Once I've answered all the questions I'll explain what did help.
The impeller was clean and clear as was all the piping going into it. The concentric reducer was just a temporary idea suggested by the manufacturer it was changed to that from the setup shown in the background. I could see why it would be a better idea to have an eccentric reducer there. We'll probably do that in a final design if one is needed along that horizontal run. The vortice issue was one that the factory had brought up as well. My coworker and I had tossed around the idea of using clear pvc pipe and injecting food coloring.
At some point yesterday I decided that I wanted a pressure gauge even closer to the pump discharge outlet, so we removed the rubber expansion joint and fitted a hard piped flanged nipple with a gauge. We ran the pump for a few minutes and noticed a drop in pressure from this newly placed gauge to the original gauge. Its clear to me now that the discharge pressure gauge cannot be downstream of any major drag inducing elements, the rubber expansion joint we installed was constricted to .85" diameter and had a rough internal layout. As the pictures show, downstream of the flex coupling we have a 90 degree elbow, 1" Tee, a 1x2 bushing & 2" tee where the old gauge is located.
After the new gauge location was fitted the pump was started; after settling for awhile the new gauge read 17 psi at the fully open condition. The gauge situated in the old location read only 10psi. We've since taken 4 points and they all graph far beyond our old tests, lying directly on top the factory curve. As soon as I get back to the office I'll post the results.
Thanks for all your help in this matter!
RE: Centrifugal pump curve falls short of factory test
RE: Centrifugal pump curve falls short of factory test
A standard mass produced pump that has been round for years and properly tested over and over will perform near curve everytime provided it is installed correctly and any testing is undertaken with correctly positioned and accurate measuring equipment.
In my 30 years in the pump industry, not once have I seen a factory tested pump which "doesn't meet the curve" on a site test and returned for re-test fail the re-test. On a number of occassions where liquidated damages etc were involved I had pumps tested on other test rigs just so there wasn't any bias with results.
RE: Centrifugal pump curve falls short of factory test
It seems that the discharge velocity head was calculated for the field tests based on a 1" pipe, but the static pressure was measured in the 2" section of pipe. You can't add the two to get the total head. This is adding apples and oranges.
The original pressure gauge was mounted on a reducing tee, with the inlet flow via a reducing bush. You could make a crude venturi ejector this way and I wonder if the gauge wasn't actually seeing something of a venturi effect? But the expected pressure drop through the 1" elements upstream of the gauge pretty much account for the "missing" head and this venturi effect is probably small.
Katmar Software
Engineering & Risk Analysis Software
http://katmarsoftware.com
RE: Centrifugal pump curve falls short of factory test
One reason why it is unusual for a pump company to guarantee pump hydraulic performance based on site testing.
RE: Centrifugal pump curve falls short of factory test
"We have a leadership style that is too directive and doesn't listen sufficiently well. The top of the organisation doesn't listen sufficiently to what the bottom is saying." Tony Hayward CEO BP
http://www.youtube.com/watch?v=hpiIWMWWVco
"Being GREEN isn't easy." Kermit
http://virtualpipeline.spaces.liv