Why Would Fire Pump Over-Perform at 150%?

2/5/2024

I have a unique quandary that I haven't run into before. We have an overperforming fire pump.

It was factory tested, we have a factory performance curve of 68 psi churn, 52 psi at 500 gpm (100%), and 35 psi at 750 gpm (150%).

Once it was field installed, we were 1-3 psi over on Churn, 50%, and 100%, but +10 psi on the 150%. We had a net pressure of 45 psi at the 150% test when, based on the factory curve, the net should have been 35 psi.

We checked the nameplate for the match, nameplate for the expected pressure at 150%; they all match. The supply for the test in the field got pretty low (about 25 psi suction at 150% flow).

The only unique thing I can see about this setup is that we have a very tight room and had to meet a military spec to flow through the flowmeter and both run outside as well as back to the recirculation. As a result, the path from the pump discharge to the outside has to navigate through nine (9) elbows in order to get enough clearance upstream and downstream for the flowmeter in this very tiny pump room.

Could a test header with that many bends be affecting the net pressure on the 150% test?

Is this considered a failure?

As this is military, it'll be by the book and I'm concerned that an overperforming pump might set up future tests for failure if I can't identify why it would be overperforming at the 150%.

Thanks in advance!

Sent in anonymously for discussion. Click Title to View | Submit Your Question | Subscribe

19 Comments

Dan Wilder

2/5/2024 08:02:00 am

If you're setting up the pump efficiently, flowing through the flow meter to the test header is preferred as you can calibrate the flow meter onsite without having to take equipment out of service and sending it off.

If this is a diesel pump (assuming so), they run rich to start anyway so being over a curve is also pretty typical on the first two points, the third - ehhh...

You did not mention what the RPM's were...do these align with the pump MFG sheets or is it going into overspeed?

Is it occurring when flowing water, using the flow meter (closed loop), or both?

Has your pump rep checked the programming for both the engine speed (if going into overspeed) and fuel system?

Random but....it's not using bio-diesel is it?

As this sounds like a Cx activity, the pump meets the curves.
However, per 25, if you meet at least 95% of each of the points for the pump, it's considered a pass. Not meeting the rating of the pump or not meeting system demand are the fail points, variations from year to year more than 5% should be reviewed (in a nut shell) so you are correct in thinking this may setup an issue for next year. Not 100% on any military requirements on the ITM side.

2/5/2024 08:11:23 am

Dan - great input, thank you!

It is an electric vertical inline pump, and it is a commissioning exercise. No diesel or bio-diesel.

The RPMs on the 150% test aren't terribly off (around 3560, if I remember correctly).

The flows we measured on the 150% were from actual water flowing outside, using handheld pitot gauges (all gauges calibrated with recent certificates).

We ran the 150% twice, at different times of day with about a 5 psi different in coming suction pressure, but the net pressure was identical in each of these two tests.

2/6/2024 07:18:20 am

Just a follow up with having no viable answer....

You said the water supply suction was at 25PSI when the pump was flowing at 150%, does this align with the water flow test curve for the supply? I know you were specific for pump NET pressure, but just filling in the boxes for me.

Was the test header using all the available nozzles or the minimum? I ask this because I have had test that were failing using only 2 nozzles at the higher pitot, but when we opened the third and checked that against the lower pitot, the test passed.

2/6/2024 07:43:37 am

Dan - that's interesting. Flow test was 32 psi static, 30 residual at 987; so not terribly far off. That makes the pump suction 3-6 psi lower than the flow test, which is probably accounted for with the backflow preventer loss.

We did the 150% with both (2) 2-1/2" hose valves flowing with a handheld pitot measurement of 17 psi (two hoses).

Bill Jones

2/5/2024 08:16:22 am

So being an electric, what are the data plate numbers and what are the actual reading numbers of amps and volts?. Are you using the controllers reading for AMP's and volts or taking them from the lugs and wires inside the controller?

On the RPMs many VILs are driven by 3550 rpm motors. What are the RPMs at each reading? Mainly Churn, 100 and 150% Id be interested in.

2/5/2024 10:13:49 am

Sure - thanks Bill.

The RPMs were 3577, 3569, 3565, and 3561 RPM for the Churn, 50%, 100%, and 150% respectively.

RPM was read with a meter right at the stem of the pump itself, not from the controller.

Voltage was 499/498/499, 497/498/498, 498/497/498, and 496/496/496 for the Churn, 50%, 100%, and 150% respectively.

Amps were 19A, 22A, 24A, and 26A.

Voltage and Amps were read off the controller.

Bill

2/6/2024 09:20:04 am

This is where I as a rep would turn to the sprinkler guy and say....someone owes us a change order for some more money. We gave you a better than required performing pump.

The one thing I was looking for was the amps to that of the motor plate. The amps given did not show a massive rise so I doubt its a incoming power issue.

I'd tend to look to prior comments about downstream pumping stations kicking in. Have had that occur several times on military facilities.

Glenn Berger

2/5/2024 08:55:34 am

I would recommend checking the water supply condition. You might have kicked on a supply pump with the higher flow rate.

To your other question - NFPA 20 does have a clause that if the pump test outlet is over a specific distance from the pump then the test pipe needs to be upsized. The quantity of elbows can present an issue.

2/5/2024 10:15:19 am

Right - the water supply at the suction was 39 static, then 27 psi (50%), 26 psi (100%), 23 psi (150%).

The pump curve though is based on net pressure, so even a little bump in the 150% shouldn't (I wouldn't think) impact the net pressure performance of our pump.

Greg

2/5/2024 09:28:28 am

The military base may have a water supply system that is similar to some municipal arrangements, where an elevated water supply is supported by booster pumps or a refill - booster pump.

My first thought with the condition you described is that at the 150% flow gpm, the test is receiving a benefit of an extra booster pump within the piping network.

This additional pump may not kick on until either ; the flow rate, the pressure drop , or the level of the elevated tank decreases sufficiently.

Double check with the water system operator to understand all of the supply sided inputs and their sequence of operation; elevated tanks, level of refill, tank refill pump, booster pumps, well pumps, and occasionally, interconnection to another water supply through a check valve, that will also kick-in for very low water pressures.

2/5/2024 10:17:38 am

Thanks Greg - I would certainly agree that it could be a culprit, however the suction pressure at 150% was still pretty consistent with the 100% test.

Our 100% test has 26 psi at pump suction, while the first 150% test had 23 psi suction, and the second (hours later) was 27 psi suction... so not a huge spike one way or another that would make me think that a base pump is kicking on.

It's the Net pressure that is 10 psi over the factory curve; and that's what's throwing me for a loop here.

Pete

2/5/2024 10:29:05 pm

I've seen a similar set up in a municipal water purveyor network, where the stage pump is designed to turn on and off based on a pressure switch. However, because the stage pump could potentially create its own feedback loop and that could lead to short cycling, a time delay was added to the start and stop after the threshold pressures are achieved. We went out to do a second flow test on a hydrant, and the water purveyor met us there. He said, "just wait". 5 minutes later we met demand pressure.

Todd

2/6/2024 06:46:09 am

Contact the pump manufacturer and provide them with the details of your test. You have a good data set and they should be able to review along with their results and advise you of the ANSI/HI 14.6 (American National Standard for Rotodynamic Pumps for Hydraulic Performance Acceptance Tests) allowed performance variances.

Also,
The NFPA piping limitations on the discharge size (size/length of run) have been established to prevent creating a system curve that would prohibit the ability to test a pump to the necessary flow rates. The 'system' curve is essentially a representation of the piping system response to flow; starts of low a smaller flows and increases as flow increases - where the two meet it the limit of the system and no matter how many valves you open the pump will not proceed beyond that point. Essentially the 'system' becomes a flow restriction. If you think of a pump curve as \ and a system curve as / where the two X is the furthest you are going to get on a performance test.

2/6/2024 07:47:50 am

Thanks Todd - this is good advice. We'll send it back up the fire pump rep chain for their take.

And absolutely on the system vs. pump curves. I'm not 100% certain on the impact that having so many elbows between discharge and the test header has -

In theory, if there were 50 elbows back to back, then there's be so much pressure difference between the test header flowing 750 gpm and the pump discharge pressure in order to make the test work. That MIGHT result in an artificially high discharge pressure, because that pressure is "allowed" to build up. That could be the cause (or at least a theory) as to what's happening here.

Essentially - if it were a clean, short and open run to the outside with little resistance, then the pump performs 750 gpm with no resistance. But since there's so much resistance, perhaps the discharge pressure is allowed to build up and it results in a higher discharge pressure AND a resulting higher NET pressure for the pump.

Again, just theory, not quite sure if that is correct here.

2/6/2024 01:45:32 pm

Not quite what I'm saying. The fact that you can get the pump out to the 150% mark (and perhaps beyond ???) indicates that you don't have a system curve that resembles a partially closed valve. The pump is a machine only capable of producing outputs related to inputs. If the friction loss through the discharge piping after the pump is too high it will artificially throttle the pump back to a specific flow rate on the pumps performance curve just like closing a valve.

https://www.engineeringtoolbox.com/pump-system-curves-d_635.html

2/12/2024 07:09:59 am

Once you have an identified culprit; it would be great if you could let us all know what it was... Thanks.

2/6/2024 08:56:23 am

Reading through your question again, two final thoughts;

With 9 elbows, are you measuring turbulent flow , which may give you an errant reading (even with a steady needle) ?

Is the discharge side pressure relief valve creating an errant reading , possibly due to the exacerbated discharge configuration ? I think there is a term for it , the throttling phenomena ?

Jack G

2/6/2024 10:16:08 am

The pump you ordered was on an assembly line. Commonly, the factory is short parts, most typically frames, sometimes engines. They test the pump and the curve is “ off”.

Dave Foley

2/7/2024 01:04:46 pm

Has anyone considered that If the flow meter is dumping the pressurized water back to the supply, that this will increase the suction pressure but make it look as though the pump is adding more pressure than its rating. the amount of added pressure recorded would depend on the location of the gages you're using to record the suction and discharge as well as the amount of pressure that bleads back into the city supply pipe.

Why Would Fire Pump Over-Perform at 150%?

Leave a Reply.

ALL-ACCESS

SUBSCRIBE

COMMUNITY

YOUR POST

PE EXAM

FILTERS

ARCHIVES

PE PREP SERIES