minimum allowable flow for centrifugal pump with inverter 2

[amazing azza]

Hello friends, is there any rule of thumb for calculating a minimum flow for an inverter controlled centrifugal pump?

How could one determine this value? My confusion comes from this: given a frequency you can probably find the minimum flow by restricting the outlet valve and watching for noise/temperature. But then again, you could just reduce the frequency without closing the outlet valve and achieve the same flow reduction...

 

[controlnovice]

You also have to be careful of the turn-down ratio of the motor. Running it to slow will cause it to overheat, and you won't know it by just trying it for a few minutes.

If the motor is inverter duty, it should have a turn-down ratio on the nameplate. That is the minimum speed you want to run the motor.

______________________________________________________________________________
This is normally the space where people post something insightful.

 

[GroovyGuy]

Howdy ya'll,

Actually since the pump is centrifugal type, and hence exhibits a variable-torque load profile to the motor, the motor turn-down will not be as issue. ie you will be able to run the motor continuously at any reduced speed without the motor over-heating.

If the pump were a constant-displacement type, then the load profile would be constant torque. Generally speaking a motor can only provide a turn down of 2:1 with a constant-torque profile, but you would need to confirm this with the motor OEM. Sometimes an aux blower fan can be attached to the motor to extend the turn-down well beyond 2:1.

GG

"I have not failed. I've just found 10,000 ways that won't work." Thomas Alva Edison (1847-1931)

 

[amazing azza]

controlnovice, what would be a reasonable guess for the turn down ration, if it is not listed? Or does it vary greatly?

edit: hehe, I should really refresh the page before posting Thank you GG

 

[djs]

In general TEFC motors should not be run below 50% speed. Below that speed, they usually do not have enough air flow over them to keep cool.

 

[GroovyGuy]

Howdy djs,
Take a peek at the torque capability curve of a NEMA motor under VFD control;

http://www.weg.net/catalog/weg/US/e...-60-Hz-IC411---TEFC---Foot-mounted/p/12603919

From the above curve, for a WEG 10hp 460V motor, you will see that the motor cam produce full-load torque down to 5hz. Although I would not recommend that you operate a motor at full-load torque (At 5hz), the concept is as I stated above.
GG

ps For what-its-worth, I see no reason why the OP could not run his centrifugal (variable-torque) pump from synchronous speed down to 5hz, although I am sure that he will not require that low of speed. (ie usually a centrifugal pump seldom is ever called upon to run lower than 75% speed).

pps Again I would recommend that anyone wanting to operate a motor under VFD control, ask the motor OEM to provide the torque-speed capability curve for that motor.


"I have not failed. I've just found 10,000 ways that won't work." Thomas Alva Edison (1847-1931)

 

[roydm]

Would you not reach a point where the pump is not producing enough head to move the fluid.
It might sit there in the pump and just get hotter and hotter until it boils.

Just asking

 

[ScottyUK]

You certainly would. That's why GG's observation that you don't often see a centrifugal pump operating below 75% speed is fair, although it's not universally true.

VFD's driving pumps operating into a constant discharge pressure aren't always a great idea - the frequency modulation range between zero flow and full flow can be very narrow, especially in high head applications like boiler feedwater pump applications where the majority of the energy goes into producing head, not flow.

 

[LittleInch]

No there is no rule of thumb because every system is different.

Without first calculating or thinking about a pump system curve you can't know whether your pump will continue to flow at your reduced speed or not.

Affinity rules apply here for centrifugals
Flow is proportional to speed / RPM
Head is proportional to speed^2
Power is proportional to speed ^3

So say your pump flows at 100 units per time constant at head of 100 units at a speed of say 50 Htz
At 40 htz the pump could pump at 80 volume, but a head of only 64 units. If to flow 80 units you actually need say 75 units of head, then you won't get 80 units, you will get less than this as the pump backs up its pump curve.

However if you do only need 64 units of head then you will get 80 units of flow.

sometimes your system curve changes over time, e.g. filling a tank is easier when it's empty compared to when it's full. Sometimes the system your pipe is connected to changes in pressure.

That's why there is no such ROT, but note that at 75% turndown you only have half the pressure/head that you do for 100% for the same point on the curve.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[djs]

What I meant in my post is about motor cooling, not torque. TEFC (Totally Enclosed Fan Cooled) motors are cooled by an external fan mounted on the armature shaft. Reducing the motor speed reduces the fan speed which reduces air flow over the motor. Below a certain speed, usually 50%, the motor will over heat due to insufficient air flow over the motor. There are solutions to this issue such as externally cooled motors.

Another issue is heating inside the pump itself. The very action of a spinning impellor generated heat. This heat needs to be dissipated or removed somehow. Usually the flow of liquid through the pump removes this heat. However at zero flow, this is not the case.

 

[amazing azza]

Thank you all kindly for the valuable responses!

 

[jraef]

To the point raised by ScottyUK and LittleInch, I once did a project in which we replaced 3 x 600HP 2300V pump starters with VFDs. The pumps drew water from an artesian well and pumped it up a hill to a tank about 5 miles away. At commissioning, they decided to test the system at low flow and we started all three of the pumps at 50% speed. An operator with a radio at the other end reported no output into the tank. We kept increasing the speed until we got to 95%, at which time he could SEE the water in the pipe, but it stopped about 5ft short of the top! So as it turned out "minimum speed" on those pumps was 98%... making the entire $1 million project of retrofitting them with VFDs a compete waste of time and money (good for me though). The Civil Engineer lost his job over that...


" We are all here on earth to help others; what on earth the others are here for I don't know." -- W. H. Auden

 

[itsmoked]

Great Jeff. 2% adjustablility!

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ScottyUK]

jraef,

That's really bad! That kind of story is why VFD's get a bad reputation in some industry sectors, when in reality it's usually a good technology applied badly by people who don't understand their process properly.

 

[GroovyGuy]

Jraef, Somebody did not review the pump curves... oh bother!

"I have not failed. I've just found 10,000 ways that won't work." Thomas Alva Edison (1847-1931)

 

[LittleInch]

jraef,

That's a bit of an extreme example where it seems 98% of the head is required fro static lift versus friction, but does illustrate very well why you need a pipeline engineer to figure these things out....

Any idea what the elevation change was? That must have been some "hill".

I would say they reviewed the pump curves quite well, but didn't really understand how to superimpose a system curve on it. Or maybe they didn't work out that m or ft head of the pump versus the static elevation change.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[waross]

It is possible that reduced speed will rduce the head to the point that the flow stops. If the pump is left churning with no flow, the churning of the fluid may cause the fluid to boil in the pump.
I have seen churning and boiling against a fairly high head burn the paint off of pumps.
If you had left the pumps running with no flow you may have seen this after a while, Jeff.
I have also seen pumps broken, that is a piece of casting the size of my hand, blown out of the side of the pump housing when reduced speed caused the pump to run at a forbidden frequency.
The rule of thumb may be:
1. Do not reduce the speed to the point that reduced head causes a no flow condition.
2. Investigate the system to determine if there are any forbidden frequencies that may be encountered at reduced speeds.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[LittleInch]

"forbidden frequency"??

Please explain - it's not something I've come across before, at least for centrifugals.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[ScottyUK]

Usually something that excites a resonant frequency; I've always known them as critical speeds. The VSD can be programmed to make a rapid transition from one side of the 'forbidden frequency' to the other without lingering, and will not allow operation at the critical speed. I haven't seen the problem first-hand on a pump but I can imagine that a pump impeller might excite a resonance at a critical speed. Operation near a critical speed on a turbine gets a little exciting. ;-)

 

[waross]

Thanks Scotty.
Each impeller blade passing the discharge port sends a pressure pulse at the speed of sound in that particular fluid down the discharge piping. When this pressure pulse encounters a restriction such as a 90 degree bend, part of the pulse is reflected back to the pump.
At the critical frequency of the VFD or the critical speed of the pump, the returning pulse coincides with the next leaving pulse. These pulses may become additive and the pressure pulses will increase in pressure until they may be limited by the elastic expansion of the pipe or in some instances actually blow a section out of the pump casing.
Think "Water hammer on steroids".
A short distance to the first 90 degree bend will generally raise the critical frequency about the operating range of the pump.
In the installation that I saw damaged, the pump was in a tunnel under the center of a large settling pond. The pump discharged straight into a long straigh pipe down the tunnel. The first 90 degree bend was some distance away, so the two way transit time of the pulse was able to match the period of the pulses and meet the next impeller blade.
And what Scotty said. note: Due to the slip in an induction motor, there is a slight difference between the critical speed and the critical frequency. The critical frequency is the slightly higher frequency that causes the critical speed.

Bill
--------------------
"Why not the best?"
Jimmy Carter