Centrifugal Pump Flow issue 1

[mech711]

On a chilled water system I have installed a pump which is designed to 65 ft head @ 137 GPM on the pump manufacturers curve. I am getting 46 Ft head 91 GPM and a pressure differential of 20 psi (34 In 54 Out) from the pump. It's a Taco 3009 KV series pump @ 1760 RPM and the inlet outlet piping sizes are as specified at 3". Does anybody know any possible problems that may be causing this?

 

[JJPellin]

My first guess would be that the motor is running in reverse. However, there are several things that can cause similar performance issues.
Resticted on the suction - If the suction pressure gauge is right at the pump this is unlikely or you would see low suction pressure.
Restricted on the discharge - Also unlikely. You would expect high discharge pressure.
Alternate flow path - Flow is passing back through a spill-back line or backwards through the spare pump and not passing through the flow meter. Check the motor amps to see how that compares.
Wrong impeller - The pump could have delivered with the wrong impeller trim or pattern.

 

[IRstuff]

Can you bypass the output and just dump the water to find out if it's the pump?

TTFN

 

[mech711]

A water balancing report was provided by a certified balancer showing actual amps to be 9.0 and running at 230 volts. Required amps for the pump are shown @ 230 volts to be 13.2 amps by the manufacturer. This is a two pump configuration where the two pumps are piped in parallel. One acts as a standby and they are operated in a lead lag automation. The report shows that both pumps are showing the same conditions when tested individually. Both under flow by 46 GPM. The suction line entering the pump is piped with the required 10 pipe diameters as needed (I was sure to check this first) and that is where the balancer has taken his flow reading with his ultrasonic device which strap on the pipe. This 5 HP pump should well be able to handle the conditions according to manufacturers pump curves. I don’t believe cavitation could be an issue but haven’t ruled it out. As for the location of the pressure gauges, each pump has inlet/outlet gauges about a foot from each side of the pumps. Any further suggestions with this new information? All help is very much appreciated.

 

[mech711]

I can not bypass the outlet. The system is a 40% propleyne glycol solution.

 

[BigInch]

I have heard that Taco pumps are one problem after another. I had one guy tell me they were easy to find the problem ones at night, cause they were glowing red hot just before they exploded.

I know its not much help. Just to say "expect the unexpected".

Going the Big Inch!

http://virtualpipeline.spaces.msn.com

 

[Yorkman]

I know you have probably already check it but is there a check valve at the discharge of each pump and are they functioning? Have you tried dead heading the pump at the discharge and finding out what your dead head is, it should calculate out to your 65ft head. By the way running the numbers
GPM1²xFtHd2 = GPM2²xFtHd1
I get 115 GPM.
One other thing to check, make very sure that you've vented the evaporator bundle of air. If air is trapped in the water side of the evaporator it can affect the discharge pressure of the pump, air locking the evaporator if you will.

I'm not a real engineer, but I play one on T.V.
A.J. Gest, York Int./JCI

 

[JJPellin]

You didn't mention if you verified rotation. The last item that came to mind later was impeller setting. I don't know this pump manufacturer. But if they use an open face impeller, the setting of the impeller is key to proper capacity. Some impellers are set to a certain clearance to the front-side. Others are set to a certain clearance to the back-side. Are you sure that the impeller was set correctly? If it was vertical volute pump there is another problem I have seen. Some of these pumps will not meet performance at full diameter impellers because the bottom of the case is not machined to a flat surface to a large enough diameter. The face of the open-face impeller is hanging out in space since the coresponding surface is not machined to match it. I believe it was an old Ingersoll Rand HOC pump that had this problem.

 

[25362]

Additionally, from a checklist in Yedidiah's Centrifugal Pump User's Handbook:

[•] air in the system
[•] excessive leakage through wearing rings
[•] viscosity of liquid higher than originally designed
[•] impeller OD machined to a smaller dimension than specified
[•] partially plugged suction strainer
[•] NPSHA too low

 

[25362]

Actual thermophysical properties of the pumped solution may negatively affect the NPSH requirements often determined for water by the pump maker.

Besides, when comparing the volumes of 1 gr of saturated water vapor developed by an equal heat input by the pump, it shows at 0 Celsius, 206300 cm³ vs 44730 cm³ at 25 Celsius (assumed testing conditions by maker).

 

[pmover]

to supplement previous postings, always question, verify, and confirm the data when equipment performance does not meet mfg design performance (especially if new equipment). a gauge not properly calibrated, a plugged gauge connection, data not taken at proper location, units of measurement on gauge (yes, encountered this one before), etc. are some matters to investigate.

good luck!
-pmover

 

[walkes]

Did somebody accidentally put pumps in that are to operate at 50% load instead of duty standby? Did something change towards end of design. The flows seem to match that situation. ie. 2 pumps operating in parallel 132 GPM @ 65ft if one fails you will get approximately 70% of the flow or ~91 GPM.

 

[BigInch]

Not likely. The OP says he has "A" pump curve showing the one pump is designed for 137 gpm @ 65', so if there was one pump not operating that was supposed to be running in ||, the system design flow would be 274 gpm at 65'. 70% of that would be 191 gpm and the pump differential head would typically rise from the 65' it is supposed to be at BEP as flowrate decreased. Diff head has decreased to 45'.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com

 

[walkes]

I meant two pumps each with 94 gpm design flow. I checked some Taco software it indicated that particular model would have an 8" diamter impeller for 94 gpm and an 8.3" impeller for the 137 gpm. The power consumption/amp draw seems to coincide with the situation.

 

[BigInch]

That's prob. a long shot, given the OP description. Stranger things have happened I guess.

BigInch-born in the trenches.

http://virtualpipeline.spaces.msn.com