Pump Curve vs Performance

[Clucky]

Hello, question here related to what I'm seeing on a pump curve vs. what the established system is displaying. I have a 2.5 x 3 x 8 centrifugal 3500 rpm three phase that is either pumping less than designed or being effected by the system. The pump curve for the 7 3/16" impeller displays a deadhead pressure of ~220 ft with a very gradual slope down to 300 gpm at 200 ft. The fluid is ambient temp water and there is about six feet of static head pressure on the inlet of the pump with very minimal inlet piping. The system consists of roughly 30 ft of 2" line with a handful of 90 degree elbows and a few restrictions in the form of a few ball valves, a pressure reducer (fully open), a flow meter tee which is 1.5", an eductor throat (0.75"). I'm getting about 75 gpm of flow through the meter and a pressure of about 210 ft directly off the pump. I guess I'm looking for help on how the sytem effects where the pump will operate. I'm obviously way below the curve-suggested flow but I'm not sure if that is due to system limitations or the pump. Thanks for any info.

 

[SNORGY]

Just reading the book, Cameron Hydraulic Data, if you have SCH 40 pipe you can expect to incur a friction head loss of 149 ft / 100 ft; for SCH 80, 208 ft / 100 ft. Without doing more detailed analysis and quantifying the rest of the L/D, the system may be behaving correctly. The 2" pipe is a tad small...

Regards,

SNORGY.

 

[SNORGY]

Sorry...the head losses I am giving you are those that would have been predicted for 300 GPM; at 75 GPM, they would be 10 ft/ 100 ft and 13.9 ft / 100 ft, respectively.

Regards,

SNORGY.

 

[TenPenny]

It looks to me like the flow rate is not at all surprising, given the head (210 ft).

Need to lower the resistance of the system - larger pipe and/or full port valves and/or less elbows etc.

With a 'gradual' curve, it doesn't take much to move very far on the curve - there's only 20 ft diffence in head for a range of 0 to 300 gpm.

 

[JJPellin]

If you are measuring the pressures right at the pump suction and discharge, then the system should have no affect on where you run relative to the curve. It will have a very great affect on where you run on the curve. In other words, If the system has excessive downstream restriction, it may run at a very low flow rate. But, plotting that flow rate and head, you should still be on the curve.

A have made up a number of tools to help troubleshoot pumps that are (or seem to be) underperforming the curve. The most common causes are these:

1. Blocked on the suction
2. Blocked on the discharge
3. Alternate flow path
4. Pump wear
5. Running backwards
6. Assembly error
7. Wrong product
8. Driver speed

In your example, you cannot rule out #1 unless you have measured the suction pressure right at the pump inlet flange. There could be a dropped gate on a valve or a plugged suction strainer. In cold water, you would probably be able to hear the hard cavitation in either case, so it may not be likely.

You can rule out #2 if you have a pressure gauge right at the pump discharge. If there is anything in between the pump and the pressure gauge, even a wide open valve, this could be the problem. We find at least one gate valve with the gate dropped each year.

Alternate flow path usually involves flow that is not passing through the flow meter. A minimum flow spill-back, a leaky check valve on a companion pump, a warm up line or any branch connection that would be diverting flow such that it does not pass through the flow meter will introduce error. Block in any other lines to see if the flow you register on the flow meter changes.

Pump wear could involve impeller wear, wear ring clearances, volute lip wear, etc.

Running backwards, some centrifugal pumps will put up as about half as much flow, at a given head. Your three phase motor could be wired for the wrong rotation.

Assembly error has mainly affected us for double suction impellers which can be installed backwards. But, if a wear ring was left out or some other serious assembly error occured, you could have a major loss of performance.

Wrong product usually just affects that accuracy of the flow measurement and the calculations you did to convert pressure into head. But, for ambient water, this is unlikely.

If the curve is drawn for a speed different that your driver is running, this will show up as a loss of performance. This is more serious with a turbine, but can be relevant with a motor, as well.


Johnny Pellin

 

[Pumpsonly]

From the info given about the shut valve head and the the head at 300GPM, your pump curve seems relatively flat. Is the 75 GPM at 210Ft on the pump performance curve?

You should post the curve for better study.

 

[DubMac]

Just for a back to the basics comment:

Centrifugal pumps will ALWAYS operate at the intersection of the pump's performance curve and the system head curve.

In answer to your question: How does the system affect where the pump will operate?
Any increase in friction, or flow resistance, will cause the system curve to become steeper, and thus move the intersection point with the performance curve to the left.

 

[Clucky]

Thanks for the info thus far. The pump curve can be found at the link below. It's the C curve for the 7BF model on page 5. Update - I bypassed the eductor throat will a 2" hose, flow increased to ~98 gpm. I then replaced a solenoid ball valve and the adjacent Honeywell pressure regulating valve with a spool piece and the flow increased to 190 gpm. I've reinstalled the solenoid valve and put a spool piece in place of the prv to assess that flow but it looks like in all the excitement I got water in my flow meter,its no longer reading. Letting it dry and hoping it start working.

http://www.pumpshop.com/pdf/3656M/36-3756M.pdf

 

[TenPenny]

Ah, a 3656. Those curves are so flat that it takes very, very little change to your piping to make a fairly significant difference in flow.

If you're looking for 300 gpm, you'll have to do something to reduce the friction in your system.

 

[Pumpsonly]

The curve is so flat, your flow will be fluctuating like hell with slight change in the system friction.
If you can not reduce your system losses,the next easiest step is to upgrade the impeller to the B size which is still within the 25HP motor.

 

[CaracasEC]

This is the problem on flat curves. A little changes on system resistance, would mean a big movement to the left of the BEP. One head with many flow. A minimum 5% head-rise-to-shut-off is recommended.