Booster Pump Running Off of Curve

[dchrist21]

OK, I know that a pump can never run off its curve, but I'm not sure how to address this issue. The pump is a booster that serves a branch of a facility heating loop. Suction is connected to the hot water supply, discharge pumps through a heat exchanger back to the return line. I've attached the pump's curve and the system curve based on the original design operating point.

The pump has been going through seals at a faster than normal rate, so the client contacted my firm to address the issue. They approached us with the following data:

Booster Pump Off: Flow: 81 gpm, dP: +2 psi.
Booster Pump On: Flow: 221 gpm, dP: +2 psi.

My first task is verify the dP numbers provided. There is no way that the pressure should increase as 81 gpm flows through an idle impeller. For now I am assuming that they confused the suction and discharge gauges and the actual value is -2 psi. Second, I presume the seals are wearing because the pump is WAY right of its BEP (possibly at run-out) and the abnormal impeller and shaft loads are causing all sorts of trouble. The problem is that this operating point is nowhere near the published curve. Can anyone offer an explanation of what is going on here? I will get updated pressure numbers as soon as possible.

Thanks,

 

[DubMac]

Guess I'm a little confused about your system, but how is 81 gpm flowing when the pump is off?? Is there another pump in the loop and if so, what are it's characteristics.

And don't make any fun of my confusion, I am in a crotchety mood today and your youth and skill are no match for my old age and treachery.

 

[dchrist21]

DubMac:

Yes, there is another pump in the system. The primary flow through the loop is provided in a separate part of the facility. This branch loop tees off of the heating supply line (into the suction of the pump) and discharges into the return line. The pressure differential between the supply and return is enough to induce 81 gpm of flow. In that respect, the two pumps are in series, but only a (small) portion of the flow is available to the booster (second) pump. I don't currently know the specs on the primary pump.

Dylan Christopherson, PE

http://www.intechen.com

 

[JJPellin]

I agree that you have probably been given bad data. There is really nothing you can do until you get good data. Can running at high flow lead to increased seal failures? Yes

What is the pump configuration? One volute or two? What is the suction specific speed? How much NPSH margin do you have? More data and better data will result in better answers.

Johnny Pellin

 

[DubMac]

DP = 2psi??? From the curve, the pump should be giving about 70' head to the system. 2 psi doesn't make sense.

Any way you could provide a simple sketch of the piping layout w/ respect tot he pumps as well as any of the info JohnnyP has asked for?

 

[dchrist21]

I will get some data this morning. Might be Monday before I can post it.

Thanks for the help/insight.

Dylan Christopherson, PE

http://www.intechen.com

 

[1gibson]

I know that a pump can never run off its curve

If you are talking about the manufacturer's curve, this statement is not accurate. How often do the laws of physics obey a curve drawn on a piece of paper? If you meant "should" instead of "can" then ok, you should never run off the curve. But it does happen. Look into 4 quadrant curves. Too big of a topic to get into here.

There is no way that the pressure should increase as 81 gpm flows through an idle impeller.

Why do you say "idle" impeller? This tells me there is a non-reverse ratchet, and flow is coming from the discharge, passing backwards through the pump to the suction. If there is no NRR, impeller is not idle. If flow is coming from suction to discharge, even with a NRR, the impeller will spin and is not idle. In the case I described (with NRR, flow backwards) the pump is acting as an orifice, so +2 psi higher pressure on the disch side for the "booster pump off" condition is perfectly reasonable.

Cutting to the chase, I will guess that you either don't have a NRR and the pump is spinning backwards. And/or, the thrust is reversing in the "off" condition and separating the seal faces, then smashing them back together when conditions change. Either of these could damage the seal (even if the seal drawing specifies the direction of rotation, and max allowable axial movement, which one might think would prevent these scenarios from occuring!)

I'm sorry but the language is really not precise enough for me to understand what is being described. Hope I guessed right, and this helps anyway.

 

[dchrist21]

If flow is coming from suction to discharge, even with a NRR, the impeller will spin and is not idle.

This is what I believe to be occurring. By "idle," I meant "not driven by the motor," and did not mean to imply stationary. There is a check valve on the discharge of the pump to prevent reverse flow, so I am confident that is not the case.

Dylan Christopherson, PE

http://www.intechen.com

 

[1gibson]

If direction of flow is suction to discharge, pump will spin in the same direction as operating. The flow could lift the rotor and separate the seal faces, depends if force is sufficient to lift the rotor (pump + motor stator) completely. If it is, then you could slip the set screws on the collar, not good. Not very likely that the upthrust from that flow will lift the entire rotor.

If the RPM is sufficient, you could be heating up the seal faces enough to cause flashing, or a slip-stick condition. I'm leaning towards this, considering that these are "hot water" pumps. You won't have sufficient RPM to create a good fluid film at the faces, and the friction will just continue to add heat to the fluid in the seal/seal chambe area.

 

[EnergyMix]

There is a check valve on the discharge of the pump to prevent reverse flow, so I am confident that is not the case.

Are you sure the check valve is intact? Won't be the first time that a check valve has failed.

 

[JJPellin]

Now that you reveal that the 'idle' pump is actually spinning forward from the flow of the other pump, I have some comments. How fast is it spinning? What is the seal flush plan?

The most likely flush plans (11, 21, 23) will not function properly in this freewheeling condition. The seal is spinning at an unknown speed without proper seal flush. Premature seal failure should be expected.

Johnny Pellin

 

[JJPellin]

1gibson,

In my industry, we would tend to define a pump curve as all of the operating points there that specific pump can operate. Using this definition, no pump can ever run off of its curve. The common response would be that you have bad data or you have the wrong curve. Obviously, pumps can operate differently that the OEM published curve (the book curve). I suspect that this is what the OP meant.

In the original post, the fluid is described as hot water. But, the curve shows a SG greater than 1.0, so this would not be hot water. I assumed it was some sort of heat transfer fluid.

I also didn’t see a description of the pump configuration. I assume that the pump has a true thrust bearing such that the seal faces cannot open up from the force of forward flow.

Perhaps this points out one problem. We have too little information and too many assumptions. I will wait until the OP comes back with more information before I comment further.


Johnny Pellin

 

[1gibson]

JJ, the definition of "can" (manufacturer allows it, vs. it being physically possible) is important to the discussion. Lacking so much critical information, the only real chance we have is trying to make assumptions based on what is there: a curve and a few paragraphs.

A centrifugal pump can operate off either end of the curveand, and will do so if given the chance. This includes negative flowrates, negative differential heads, you could even operate most centrifugal pumps as turbine/generators if you felt like it.

Agree 100% with the rest.

Can we email the OP an invoice? Sure.
May, will, or should we? No, probably not.
Will, or should he pay it if we do? Nope.
Can, or may he pay it? Certainly!

 

[dchrist21]

I have had a chance to collect additional operating data for this pump and speak with the mechanic responsible for its maintenance. I will try to answer as many of the questions you have posed, but at this point I don't require any further assistance. I am just posting this information out of courtesy for all who have offered their advice. Thank you, to those of you who have helped.

The original data I presented was collected during the summer, and I am collecting data during the winter. As I mentioned, this is a heating loop and the system curve varies seasonally with the connected load. Therefor the data I collected today may not be representative of the summer operating conditions. For that reason I am deferring any further analysis until I can observe the system this summer. But, for those interested, here is the most recent data.

Mechanic today said the maintenance issue is with the pump bearings, not the seals. However, they do change the seals while the pump is apart so there may be seal issues that have not yet surfaced because of frequent replacement.

Pump: B&G Series 1510, 2-1/2" BB, 1750 RPM, 8.625" Impeller.
Design Point: 200 GPM @ 72' of Head
Design NPSHr: ~6'
Fluid: 50/50 Ethylene Glycol/Water, 150 deg. F.
Suction Specific Speed: ~5735
Single Volute/Impeller


Suction Pressure: 28 psig
Disch. Pressure: 58 psig
Flow: 274 gpm
Check valve is intact, so no reverse rotation should occur under any conditions. At this point I am unable to shut the pump down to determine the "idle" rpm.

Again, thanks to all who offered their assistance, and you may see this topic come up again when it warms up in these parts.



Dylan Christopherson, PE

http://www.intechen.com