Minimum flow for a VFD pump running in close loop 4

[cholopa]

Hi:

I've been searching on this topic a lot and i can't find a reason why a pump can't operate at 1 Hz in the following conitions:

A system in a big AHU(cooling coil) with one VFD pump and no three way or two way valves.

Let's say at a given point a need 10% of the cooling power of that ahu. A tipical coil for water entering at 44ºF "would need" around 5% of the flow.

As the circuit is a close loop, the head needed would be extremlly low, so the power of the pump would be around 1 or 2%. I think i would probably need the VFD to run the pump at 5Hz or less.

I know the efficency would be horrible, but even then , the power consumption would be much much less that a system running with three or two way valve.

All the cons i have read (overheating, Radial hydraulic thrust, Flow re-circulation in the pump impeller...), i think are nonsense, as the speed of the pump is so slow that they are not of a consern. Take into account that the slip of the motor at those Hz is so high, that the real rpm is extremlly low (maybe 60 r.p.m), or in other words, just like if you were turning the pump with your hand at a rate of one rev per second.

What do you guys think of this?

Thanks for your time and help,

cholopa.

 

[DubMac]

if you cut the flow by a factor of 20 (from 100% to 5%), the head will drop by a factor of 400. In other words there will essentially be NO head developed.

Instead of talking theoreticals, could you show what the system curves would look like at 100% flow, 50%, 20% flows?? In other words, what heads would your system actually require at the low flows? It would make it easier to overlay some actual pump curves and look at actual performances with .

 

[BigInch]

As long as you can get some circulation at that low speed, i.e. you have essentially no static head; the system losses are comprized entirely of friction from flow, you could most likely get away with it, although it would most likely be better to simply shut off the system. Who needs water at 44F. When it's 83F, I shut down my AC, open the windows and listen to the birds sing. Will you ever turn it off? I'd think it would essentially be doing nothing at around 30% anyway.

We will design everything from now on using only S.I. units ... except for the pipe diameter. Unk. British engineer

 

[bimr]

Not so fast.

A VFD-driven general purpose motor can overheat if it is run too slowly. (Motors can get hot if they’re run slower than their rated speed.) Since most general purpose motors cool themselves with shaft-mounted fans, slow speeds mean less cooling. If the motor overheats, bearing and insulation life will be reduced. Therefore there are minimum speed requirements for all motors.

The motor manufacturer should be consulted for operational limitations. Most variable frequency drives restrict the minimum continuous speed to some percentage of the nameplate speed. Below this minimum speed, temperature rise may damage the motor. Every 10oC rise above the rated temperature in a motor reduces the insulation life by half.

"For lubrication purposes a pump should have a minimum speed of at least 18Hz."

http://www.danfoss.com/NR/rdonlyres/ABBA5C76-36B6-4513-99DA-28D64C1DF3B0/0/vfdlesson1.pdf

 

[BigInch]

The vfd should be put into sleep mode at 10% and solve that problem, but as I said, it should be put to sleep at 15% or so anyway. Who would run a system at 5% load, is what I'm asking.

We will design everything from now on using only S.I. units ... except for the pipe diameter. Unk. British engineer

 

[1gibson]

"All the cons i have read (overheating, Radial hydraulic thrust, Flow re-circulation in the pump impeller...), i think are nonsense, as the speed of the pump is so slow that they are not of a consern"

That is an interesting philosophy... If it is existing equipment, you could always try it and see how long before the motor overheats.

If you are considering this for new equipment, it goes from interesting to irresponsible.

You also need to consider that most of the inefficiency of the pump will go towards heating the fluid. What is your motor efficiency at such a low speed? VFD losses? These will also heat the components. So while you may save power, it could affect your process in ways you hadn't anticipated (or in ways you had anticipated, but deemed inconsequential.)

A complicated way to maintain efficiency for very large pumps at fraction of design flow, is to use a turbine/generator in place of a throttling valve. Obviously a large initial investment and not likely to be cost effective. Just a thought.

To be honest, it sounds like you want a small auxiliary pump. Cost for that, vs risk of damaging your main pieces of equipment, should easily be justified.

 

[DRWeig]

Hi cholopa,

Welcome to eng-tips!

I'm going to jump on board with bimr -- the pump probably won't be a problem, as BigInch allows -- but how will the motor cool and lubricate itself? It needs to keep both air and lubricant moving... All the waste heat from the motor inefficency at that low speed has to go somewhere, even though it's a small amount.

You might consider flagging this thread and re-posting in electric motors, generators, and control systems forum for a better analysis from the motor side.

Good on ya,

Goober Dave

 

[cholopa]

Thanks for your welcomes and replays.

Let me put this in perspective. I will change the numbers so it makes more sense (sorry i'm not used to use I-P units).

For keeping and industrial place at comfortable temperature (72 ºF), an AHU has got a water coil that works at a AT of 10ºF (44-54 ºF, of entering-leaving cool water.

Lets say the place at a particular moment needs 30% of the cooling power. Following the rules of coils, if the water is entering at the nominal temperature, in that conditions, the flow trough the AHU would have to be around 10% of the nominal flow.

As the pressure loss-flow curve in the system follows (more or less) an exponential(to two) path starting at zero (no static pressure need it, as the system is a close loop), at 10% flow, the pressurre need it would be:

P1: Pressure at nominal conditions
F1: Flow at nominal conitions.
P2: Pressure at 10% nominal flow.
F2: Flow at 10% nominal flow = 0,1 * F1

P2=P1*0,1^2*F1^2/F1^2 = 0,01 *P1

That means that P2 is 1% of P1.

As power need is:

Power 1 = (constant) * P1 * Q1 / (water to wire efficency)
Power 2 = (constant) * 0,01 * 0,1 * P1 * F1 / / (water to wire efficency)

So considering efficency is more or less equal at the two points (i know is not, but just for now lets consider it):

Power 2 = 0,001 * Power 1

Or in other words Power 2 is 0,1% of Power 1

Lets say that the pump at nominal conditions needs 4 kW.

Following this numbers, at 30% flow, the power needed would be 4 W.

So, now lets consider efficency at 30% flow is horrible, and lets think power consumption would be 40 W (thats and efficency of 10%).

You think a motor prepare to disipate around 400 W (10% of 4 kW), would have any problem dissipating 40 W at very low speeds?

Heat build up in the water is not a problem, as if too much heat is added, the control would speed up the motor, as i wouldn't have enough cooling at the coil.

Now that i have made the numbers, i have realize that there could be problems, notice this link, page 2, figure 1:

http://www.itrc.org/reports/vfd/r06004.pdf

Efficency seems to go well below 10% at 0,1% Nominal power.

Anyway, i would like to know your opinions.

best regards,
cholopa

 

[cholopa]

Sorry, the two last 30% flow, should be 10% flow.

regards,
cholopa

 

[cholopa]

I have realice that the graph i've point out, is for "fixed" r.p.m, wich is not the case when using VFD's.

What happens to that graph when reducing r.p.m's?. The effects can be seen in the same report, figure 12. In one hand, you can see that reducing r.p.m's reduces the efficency, but on the other hand, reducing the load, incrises the efficency!!!

So really don't know what to think.

regards,
cholopa.

 

[CaracasEC]

Hi Cholopa,
I consider that the pump involved in the discussion is a centrifugal pump. To lessen your power consumption is you're proposing a vfd control for the pump. From affinity law power is directly proportional to the speed raised to the power of 3. So you are right that decreasing the frequency will decrease the speed and the power of the motor. However please take note that a centrifugal pump characteristic curve clear tells us that it has one point on the performance curve that is considered as the BEP or the best efficiency point. Operating to the left of this curve would mean a loss on pump efficiency and hydraulic problems will arise when you are away about 50% from the BEP to the left which will include suction recirculation, cavitation, etc. You will notice that the efficiency of this would decrease as low as 8% if you go far from the BEP and this 92% energy that was created by the pump thru the driver will be converted to heat which would mean that the stable flow of the pump at low flow will be at risk.

 

[BigInch]

CaracasEC, sorry but I have to disagree. It is generally considered that efficiency values will closely track with the change of speed, meaning that a plot of efficiency values at the full speed curve will scale down very closely matching the head vs flow curve predicted at any speed. For example, a pump running at 1750 rpm, BEP at point 200 feet at 1000 gpm, efficiency 76%, will have the same efficiency at its scaled down to 50% speed version, i.e. at 875 rpm, the flow = 0.5 * 1000 = 500 gpm, head = 0.5^2 * 200 feet = 50 gpm, but efficiency remains more or less the same at 76%. All full speed efficiency points scale down in a similar manner, meaning no change in efficiency at any speed (within the range of normal usage), thus the percent of heat generated would not be any more, or less, and absolute quantity of heat generated by the pump, would fall with the reduced total power being consumed at lower rpms too.

Additionally, the various danger ranges for bearing stresses and thrust forces caused by pressure imbalances when operating outside BEP also scale with the variable speed curves. It would seem that the only things that still might be problematic are operating frequency ratios that might shift towards resonance frequencies of the shaft natural frequency. I think it might be possible to set a variable speed where shaft vibrations might increase, if the shaft was not stiff enough to avoid reasonance frequencies falling within the 0 to full speed operating range. Maybe someone else might want to comment on that.

We will design everything from now on using only S.I. units ... except for the pipe diameter. Unk. British engineer

 

[CaracasEC]

Yes, i agree with big inch information. If you look on the pump efficiency, this will not be affected if your reducing the speed by using Vfd's but the head will be affected. For these type of pump a flexible shaft design is not necessary as this will operate below the first critical speed as contrary to the flexible shaft which will operate between the firt critical speed and 2nd critical speed. Synchronous vibration will not be possible with these type of pump however assynchronous vibration is possible. In actual practice, the affinity laws provide an approximation between flow, head and horsepower as the pump impeller diameter or speed is varied. The actual values vary somewhat less than predicted by the affinity laws. That is, the actual exponents in the affinity equations are slightly less than their stated values and are different for each pump. This fact is a result of friction in hydraulic passages and impellers, leakage losses, and variation of impeller vane angles when diameters are changed. Pump manufacturers should be contacted to confirmed actual impeller changes and in this case the speed changes to meet new duty requirements.

 

[Pumpsonly]

bimr.

"For lubrication purposes a pump should have a minimum speed of at least 18Hz."

This has no meaning as at 18Hz , a 2 pole motor will run at 1080 RPM, 4 pole motor at 540 RPM and 6 pole motor at 360 RPM synchronous speed.

 

[bimr]

Pumpsonly

Refer to "Function #4 – Limits" of the link posted above. I will repeat it since you appear to have a problem with comprehension.

Also note that the following quote is taken directly from Danfoss Drives literature since the poster is discussing VFD's.

"For lubrication purposes a pump should have a minimum speed of at least 18Hz.

"

http://www.danfoss.com/NR/rdonlyres/ABBA5C76-36B6-4513-99DA-28D64C1DF3B0/0/vfdlesson1.pdf

Since you seem to think that you know more about VFD's than the Danfoss Drives organization, I suggest that you take this matter up with Danfoss Drives (414-355-8800). Call Danfoss Drives and inform them of your expertise in VFD's and that Danfoss Drives has made some kind of mistake in the design of their equipment. The Danfoss Drives website states that Danfoss has only been making the drives for 43 years and maybe you can teach them something.

Let me know how this works out for you.

 

[Artisi]

bimr:
Lubrication of what? a very vague reference which doesn't make any sense.

It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. (Sherlock Holmes - A Scandal in Bohemia.)

 

[cholopa]

Danfoss makes VFD's not pumps.

VFD´s has no problem running at low Hz's.

Pump manufacters also say that minimum Hz's for a VFD pump is around 20 HZ.

But for me, the point is that a pump running in a loop with no static pressure has to make a much less effort than a pump running in a loop with high static pressure, and so, the heat build up is much less. As far as i know no pump manufacter distinguish between this two situations and I thnik they should.

regards,
cholopa

 

[cholopa]

If the system with high static pressure, it can be posible, that running at 20Hz means you have no flow. That's not the case with system with no static pressure, so the two situations are completlly different.

cholopa

 

[cholopa]

Heare you can see grundfos literature about a pump with an integrated VFD:

http://net.grundfos.com/Appl/WebCAP...&typecode=NBE00&appcode=AIRCON&pdfid=2667276#

Page 24, figure 33, min. curve 12%, thats about 6 Hz in a 50 Hz nominal pump, or 7,5 Hz in a 60Hz nominal pump (if i'm not wrong).

regards,
cholopa

 

[bimr]

Artisi

I assume that the article is referring to lubrication of bearings.

Oil film thickness in motor bearings is directly influenced by the speed of operation of the motor. If the bearings are not driven fast enough, the bearing lubrication is not adequate enough for the bearings. For a general discussion on bearing lubrication:

http://www.sea.siemens.com/us/inter...esigning slow speed sleeve bearing Rev 8.pdf

Obviously, without knowing the details on a particular application (electrical motor and pump), 18 Hz should be considered as a general guideline.

"Most pump failures with VFD systems are caused by the pump running with no flow passing through. The typical failure modes are:
• Seal Failure
• Bearing failure due to over temperature
• Cavitation due to operation at the first 10% of flow on the curve"

http://www.techsys.com.au/downloads/techbull/VFDVsPumpControlSystem.pdf

cholopa, there is a good reason not to operate at low Hz. Oil film thickness in motor bearings is directly influenced by the speed of operation of the motor. At low speeds, the bearing will not be properly lubricated and will fail prematurely.

"The motor must never operate below 30 Hz. This is the minimum speed required to provide correct bearing lubrication."

http://www.franklin-electric.com/aim-manual/page-41.aspx

"Programming is also pre-set to 30 Hz minimum to maintain motor bearing lubrication required by some motor manufacturers."

http://completewatersystems.com/wp-content/uploads/2011/08/BRSDRIVE2.pdf

http://machinedesign.com/article/how-to-make-bearings-last-in-electric-motors-0427

"Operation of the driven motor at less than 25% of the motor full speed rating may cause the motor to run with inadequate cooling. Inform the user that such operation may affect the motor life, and that they should verify with the motor manufacturer’s representative that the driven motor is rated for such operation. Note that the minimum speed may be set at either the drive or
at the energy management system (EMS).

If the driven motor is to operate at less than 50% of its name-plate speed, the thermal over-load protection may not properly
protect the motor. The motor may overheat due to the reduction in ventilation at reduced speed even though it is operating at well less than its rated full load current. A thermally responsive overload protection device that responds to actual motor winding temperature may be required in this case. Advise the installer to check this requirement with the motor manufacturer’s representative.

For motors that are equipped with oil sleeve bearings, operation of the driven motor at less than 50% of the motor full speed
rating may cause the bearings to receive inadequate lubrication. Inform the user that such operation may affect the motor life,
and that they should verify with the motor manufacturer’s representative that the driven motor is rated for such operation.

http://www3.lacdc.org/CDCWebsite/up... - Functional Test VFD Seattle City Light.pdf

cholopa

Obtain the specific manufacturer's information on the pump and electrical motor and follow those manufacturer's guidelines so you avoid violating the manufacturer's warranty terms.

You will be making a mistake following "seat of the pants" opinions made by persons not familiar with your application and equipment. A more professional approach is to follow the manufacturer's recommendations.