series centrifugal pump motor mystery 1

[larryp001]

We have a pump station with four sets of paired centrifugal pumps operated in series to produce the necessary head for the force main application. Any set of pumps will run for a period of time and then system flow will begin to drop off from 2250GPM to 1500GPM. The weird thing is the motor current remains the same. We have monitored voltage, current, frequency, harmonics, and performed thermal scans and don't find anything outside specified parameters. If we stop the motor and wait for three or four minutes the pump will restart pumping 2250GPM and then drop after about 40 minutes of runtime. We have recently cleaned the wetwell to eliminate any grit and we have raised our level setpoint to avoid the ingestion of any vortices into the pump. Our increased runtime rules out flowmeter anomolies. My question is: What conditions will allow a motor to pump less fluid and maintain the same full load current when the amount pumped was as much as 30% less than full load? We have used recently calibrated Fluke instrumets for all of our measurements. These are 200 HP dry-pit submersible motors connected to the pumps. Any ideas or suggestions would be greatly appreciated. Thanks.

 

[Skogsgurra]

Motor current can be constant even if shaft lod changes. It depends on what power factor you have. But it is not usual when running close to full load. You can usually see it at partial load.

Nonetheless. Take a P.F. reading just after start and another one when flow has decreased. If your other observations are correct, then your power factor probably is less the second time you read it.

Gunnar Englund

http://www.gke.org

 

[itsmoked]

[green]"Our increased runtime rules out flowmeter anomalies."[/green]

I don't understand this comment.

I would still suspect the flow monitoring. I would add a pressure sensor temporarily too.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[larryp001]

Thanks for the information. We have the ability to feed the motors with A/B Soft Starters or A/B VFD's (we have dual capability for redundancy). We have the same occurance with either starter. The load is constant, +/- 150 feet of hydraulic head, to the top of the force main. I know the power factor will fluctuate with a change in voltage and load, both of those factors are static. When we feed the motors with their VFD's the supply system does not see an inductive load due to the nature of the VFD's. When operated with an SCR soft start, we see the same results. I don't think power factor is the issue. What do you think?

 

[Skogsgurra]

You have a VFD? Then you should be able to read not only motor current but also power - perhaps even torque from the inverter. If P.F. frequency, voltage and current are constant and the load is changing then there must be something wrong with one (or more) measurement either electric or other.

Gunnar Englund

http://www.gke.org

 

[larryp001]

With regards to the flowmeter question: The flowmeter is of the magnetic flow meter type. If the flowmeter were malfunctioning, we would expect to see normal pump times. By comparing our flow chart recorder readings, we see that we have substanially longer pump run times that seem to indicate the validity of the flowmeter information. We adjust this for daily flow changes and additional system loads and it seems the meter is working properly. It cannot be calibrated in the field but we have the ability to verify the calibration and it appears okay.

Thanks for your suggestions.

 

[larryp001]

Thanks, Skogsgurra. If the load were decreasing, I would expect to see the current and power factor decreasing. They appear to be remaining stable. When we are testing this system to find out what is going on, the VFD is set to minimum and maximum speed based on 60 Hertz. That way we can compare it with the Soft Starter. They both react the same. I would expect that the torque shouldn't shift unless the voltage, load or frequency changed.

 

[BJC]

Whats the piping system look like? Are ou filling it up or otherwise "packing" it? If so the water may not have anywhere to go and the flow drops off.
Pumps can heat a lot of water under the right conditions, have you check the temperature?

 

[aolalde]

For balanced and constant Voltage and Frequency,the current change follows the load torque demanded on the motor shaft. The power factor is inherent to the motor parameters and it will not change unless the load, voltage or frequency change.

This means the motors have the mechanical load.

Each set has two pumps in series, so the discharge of the first pump is the suction of the second?
If so, could turbulence be the reason for the extra power demand?
I am not expert on hydraulics, have you any oher experience of performance with a similar pump array?

 

[larryp001]

Thanks, BJC. The water is discharged into a common header that charges a force main. The main crests at an elevation of +150 feet above the pump. From there it gravity feeds into a sewer main that is at the bottom of the grade a couple miles away. If we stop pumping the checkvalve on our pump closes and the force main remains surcharged. So basically, it is pumping into an open channel, although raised in elevation.

 

[larryp001]

Thanks, Aolalde. We utilize the two pumps in series to produce the desired head required to pump over the hill. We have been monitoring both motors and pumps and everything looks fine. The inlet to the first pump is submerged and has +/- 15 feet NPSH. The discharge of the first pump is about 75 PSIG and travels through +/- 8 feet of pipe to the suction side of the second pump. The discharge of the second (identical) pump is +/- 150PSIG. This equates to 346 feet of head which is adequate for the 310 feet of head in elevation change. I mis-wrote earlier that the elevation was 150 feet. It is about 310 feet, which accounts for the 150 PSIG discharge. We utilize series pumps in a couple of locations and have never run into this problem. Thats why I am scratching my head so much.

 

[waross]

What conditions will allow a motor to pump less fluid and maintain the same full load current when the amount pumped was as much as 30% less than full load

An increase in dynamic head.

Whats the piping system look like? Are ou filling it up or otherwise "packing" it? If so the water may not have anywhere to go and the flow drops off.

Most likely there is a slight restriction in the line down to the sewer. As you are pumping the line is slowly backing up. When it finally fills up it increases the dynamic head. Put a pressure guage on the pump discharge this shopuld show an increase in dynamic head.
respectfully

 

[larryp001]

Thanks Waross - The discharge shows +/- 150PSIG. As the flow drops off there is a corresponding duplicate drop-off in the discharge pressure of both pumps.
Larry P.

 

[larryp001]

Also, the line to the top of the grade is always surcharged (filled) so there is no "backing up". The decrease in discharge pressures leads me to believe that we are pumping less. My question is why? We have eliminated air entrainment as a possible culprit. The piping layout exactly follows the pump manufacturers layout. They have no idea why things are operating the way they are. We are busy collecting data and thermal image scans to send to them for help.

 

[quark]

Power curve of mixed flow impeller pumps is not a straight line and there is possibility that you can have same power consumption at two different flowrates. The peak load is some where in the middle of the curve.

When you say full load at 2250 gpm, is it the maximum power with in the operation range or limiting power of the motor?

What happens to the level in the well after every 40 minutes?

What is the suction pressure of second stage pump when the flow is 2250 gpm and also when it is 1500 gpm?

What are the discharge pressures of first stage pump for the two flow conditions?

 

[itsmoked]

I predict based on your topography and the puzzling results that everyones suggestions are essentially occurring.

The downhill side of the line is longer than the uphill side.
With the pumps off the uphill side tries to drain back.
The check valve stops this.
The uphill side remains filled and you see the pressure sit basically static, pumps on or off, probably about 20PSI less than pumps on.
Meanwhile the slack side actually drains down the long distance to the main.
When the pumps start the water starts to tip over the top of the hill and then runs downhill rapidly in a partially empty pipe.
This leaves the pumps essentially pumping to the top of the hill and out of an open pipe. You get full flow and expected HP requirements.
Meanwhile the downhill pipe fills from the bottom of the grade at the main back towards the hilltop. This takes 40 minutes. The downhill side never fully fills entirely leaving a large air space that cannot pass water. This leaves your downhill pipe essentially restricted at some fraction of its theoretical diameter. This does not occur immediately but takes the forty minutes.

Mess with syphons and clear hoses and you will soon see this phenomenon. I have seen it several times! You will reach a steady-state static situation where the system just sits there acting like a pipe 1/2 or 1/3 the pipes diameter depending on the entrained air bubble.

This is not a packing but a dynamic choking state.

You could prove this by venting the line somewhere like 10 to 40 feet down the hill from the summit on the downhill side after the 40 minutes.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[waross]

Hi Keith
That was what I thought but you said it much better. The problem is that that would result in an increase in discharge pressure and larryp001 is reporting a drop in discharge pressure when the flow drops.
larryp001;
Does the suction pipe enter the pump in a straight line or is there a 90deg. bend at the pump? If the pipe runs straight in I'm wondering if you may be getting an internal vortex after awhile. With two pumps, the second pump may be pulling water past the vortex, but with unexpected operating conditions.
Re the temperature measurements; Can you get an instrument that will read very small differential temperatures? With the amount of water flowing and removing heat, differential measurements may be needed to locate internal heat generation.
respectfully.

 

[larryp001]

It is great to get such good input from all of you. The downhill side of the force main gravity feeds to a sewage lift station where the water is collected in a wetwell and then pumped to its final destination, a wastewater treatment plant. The flows into the station are roughly constant on a day to day, or weekly basis but the hourly flows fluctuate with time of day. (morning showers, lunches, dinners evening showerws, etc.) The station operates in a fill and draw mode rather than trying to maintain a specific level in the wetwell. We keep the low level shut down setpoint high enough to avoid the creation of vortices that could be pulled into the suction line. When inspecting the manhole at the top of the forcemain, it appears that water level never raises. It just flows in the channel without surcharging the manhole. Most importantly, as we close the discharge valve from the pump while it is operation we see the expected decrease in current (not moving water, not doing work, lower current). Applying that observation to your suggestion, I would think that we should see a decrease in current as the water was prevented from flowing. We keep seeing approximately the same current, but our flowmeter shows a gradual reduction in flow of up to 30% without a reduction in current. That is what is really puzzling me.

 

[rbulsara]

Is it possible that the density of fluid changes after initial flow? More sludge gets sucked and moved so more friction, work remain the same but less flow? Or on other words higher head and lower flow. HP=head x flow rate.

 

[rbulsara]

Is there a way of monitoring the head in addition to flow?