Centrifuge AC Induction Motor Requireing 30 Starts Per Hour

[rglassburn]

Hello All,

I have been asked to specify a motor for a new centrifuge. It needs to be able to handle 9.4 lb-ft^2 of inertia and spin at 3425 rpm. In order to cycle the number of parts that I need, I will need to be capable of cycling the centrifuge from 0 to 3425 rpm up to 30 times per hour. The centrifuge will be driven by a VFD. My plan is to direct drive the table using a flexible coupling.

As I have been studying the requirements, I have found the MG 1-40.40 table that seems to indicate 10 HP motor. Since a VFD will be needed to control the speeds, I will pick an inverter duty motor. However, looking around a bit more, I found MG 10-2001 which has table 7 "Allowable number of starts and minimum time between starts for Design A and B motors" which indicates that a 10HP 2 pole motor would not be able to start that many times per hour. It shows a little more than 1 start per hour?

Can anyone provide some clarity and/or point me to a source that can help me with this spec?


Thank you in advance for your insight!

Thanks,
Rob

 

[RobWard]

With a start every two minutes, could a flywheel/clutch combo offer any advantages to you to offset the additional complexity: smaller motor, energy conservation etc

"I love deadlines. I love the whooshing noise they make as they go past." Douglas Adams

 

[baxtersdad]

I believe that the tables to which you refer are based on full voltage starting with the attendant high currents/heating of that starting method. The VF drive should be able to operate the motor at 100% torque throughout the acceleration period with much lower heating than full voltage starting would produce. A motor manufacture should be able to model this for you and advise regarding the temperature rise of the motor. Within you two minute cycle you will probable have two high torque phases (accel and decel) which will contribute to the heating. You will also need to consider having a VF drive that has regeneration capabilities to dissipate the stored energy during stopping, otherwise you won't have much time left for loading/unloading of the unit.

 

[electricpete]

I concur with baxterdad - those starting limits are for direct on line start (DOL). VFD start will be much less stressful for the motor.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[jraef]

Depending on the torque required at startup. Centrifuges often are designed to utilize the high starting torque to accelerate quickly. A VFD can give you 100% torque without stressing the motor, but if acceleration of the mass in the allotted time requires 160% torque, the VFD is going to stress the motor as well. Maybe not as much as DOL, but it's not to be totally discounted either. You also must consider the effects on the VFD as well, it may overheat if it has to supply excess current every 2 minutes, they are typically rated for 150% current for 1 minute, but with a specific rest period in between of maybe 5 minutes (I'd have to look it up).

 

[electricpete]

The fact that the referenced tables apply to DOL rather than VFD start does not depend on anything.

"Much less stressful" was a poor choice of words on my part. The method of meeting the demands of this particular application are not so simple as described by Baxterdad and as you have further elaborated.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[cswilson]

Is a vector drive (sensorless vector should be good enough for this application) within your budget? If so, this should be an effective solution to this application. You will have high currents on acceleration and deceleration, and you must size motor and drive accordingly, but it's to do real work, not lost in excessive slip.

For sizing, think of it more as a servo application, cycling through four states:

1. Accel (high current)
2. Constant speed (low current)
3. Decel (high current if you want to stop quickly)
4. Stop (virtually no current as you change samples)

Curt Wilson
Delta Tau Data Systems

 

[controlsdude]

recommend a braking resister package for this application with the VFD
I thought baxtersdad said this but not sure about the meaning of regen?

 

[rglassburn]

Very interesting comments all. Thank you!

So using the equation relating inertia, speed change, time I get an average torque required of 3.5 lb-ft. This should allow me to accelerate my inertia to my speed in the time frame I need.

The way I understand this is that as long as my motor has a rated torque value in excess of the 3.5 lb-ft I calculated, the VFD should be able to drive the motor to deliver this torque thru the ramp up and ramp down of the centrifuge.

Does this also mean that since the across the line starting torques that cause the excessive heat is eliminated and that the number of starts per hours is a non-issue?


Thank you,
Rob

 

[electricpete]

Full load torque on this motor (10hp 3600rpm) would be around 15 ft-lbf, so that's down around 25% torque.

I'm not sure the torque tells the whole story. The slip is quite an important part of the picture also. If you immediately apply line frequency at reduced torque, you'd have a helluva lot more heating than if you ramped the frequency. Most everyone here knows that as well or better than me... just wanted to mention it.

Have you figured out how you will accomplish deceleration?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

In fact, I would venture to say that applying the reduced torque at immediate full frequency is almost as bad as a DOL start. (The total time-integrated rotor heating will be at least as much as the final kinetic energy of the system in either case). You need to ramp the frequency.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Please disregard my immediately preceeding post 19 Oct 07 13:14. Let me rephrase it as a question:

Does applying a specified reduced torque (for exaqmple 25%) cause the VFD to ramp it's speed to maintain that torque (controlling slip down somewhere well below 5% the whole time) ?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

a correction to the correction to the question:
"Does applying a specified reduced torque (for exaqmple 25%) cause the VFD to ramp it's frequency to maintain that torque (controlling slip down somewhere well below 5% the whole time) ? "

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[rglassburn]

Electricpete,

Thanks for the reply!

I was going to let the VFD control the deceleration.

For your other VFD question - I do not have an answer.

I am being led toward a 3HP (8.9 lb-ft full load torque) or a 5HP (15 lb-ft full load torque) 4 phase(1750rpm) inverter duty motor by a manufacturer. They indicate that since I need the faster speed the VFD will allow that. They claim the motor can move at up to 6000 rpm (I am assuming the bearings is the limiter). However, they initially only said 3-4 starts per hour! I have given them more info about my requirements and am waiting for a reply.

So I have effectively put it into their hands. However, I would love to know what they are using to determine the motor size I need. Short of making the torque rating of the motor larger than the required torque, making sure the bearings can handle the speed, and using an inverter duty motor, I don't know what they are using to determine the size. I would prefer them to teach me how to fish instead of giving me a fish. Any ideas?


Thank you,
Rob

 

[cswilson]

At a minimum for this application, you will need to control the commanded profile of an open-loop VFD output quite closely. (The profile means both the voltage magnitude and frequency, roughly in straight proportion to each other.) The profile includes an acceleration ramp, a constant-speed portion, and a deceleration ramp. Either the VFD must be sophisticated enough to generate this profile itself, or you must externally command it, probably through an analog input on the VFD.

There are two reasons you need this. First, it permits quick acceleration and deceleration, which it seems like you need. Second, it dramatically reduces motor heating compared to step changes in voltage/frequency. A stopped motor given full AC voltage/frequency is almost a short circuit, which is why these motors only permit you to do this type of start a few times an hour. Ramping the profile eliminates this issue.

(If you plan to do much low-speed motion, a fan attached to the motor shaft may not supply enough cooling for the steady-state heating, something you should check.)

For somewhat more money, a closed-loop vector drive (with or without a shaft sensor) will provide even a more optimized capability, because it will modulate the output profile based on what the motor is really doing. You would undoubtedly be able to accelerate and decelerate somewhat more rapidly with a vector drive than a ramped VFD. Whether this is worth it to you is a different question.

Because you undoubtedly want to decelerate more rapidly than a coast down, you will be regenerating and so must consider where this energy will go. (By "this energy", I mean the full kinetic energy of motor and load at the maximum speed.) You have two choices. Either make sure your drive's capacitor bank is big enough to store the full kinetic energy without excessive voltage rise (which is nice because you get to reuse that energy for the next acceleration), or you must dissipate that energy in a shunt resistor (probably lower capital cost).

Curt Wilson
Delta Tau Data Systems

 

[itsmoked]

1) You will probably need to provide extra cooling for the motor. An auxiliary fan that can add cooling while the motor is even stopped.

2) You may need extra cooling for the VFD. Another auxiliary fan.

3) You may not realize that 'just the VFD' may not be able to stop the the centrifuge as quickly as you you think by itself. As the VFD slows the motor, the motor generates power which returns to the VFD's internal capacitor bank. This raises the voltage on the capacitors. When it gets a little higher than normal the VFD either trips on error,(you're screwed), or you must greatly reduce the deceleration rate and your 'grand scheme' takes a BIG hit.
The required solution to this is the braking resistor. As the voltage rises on the capacitor bank it is shunted to the braking resistor allowing a much much higher deceleration rate. A better solution in your case may be a more advanced VFD that allows the energy to be shoved back out to the power source. This is far "greener" and can save on power bills. Likely a large amount of regeneration is available on your application since you will be getting back most of your input energy since your system is not very lossy but rather just inertial.

4) You mentioned flexible shaft... Those can be very lossy. This increases the motor load dramatically by numbers like 20%, especially at higher speeds such as yours. All this loss shows up as heat in the flex drive cable of course which can actually limit your entire throughput. Also they don't do too well with accel and deccel which causes even more internal heating. I would personally jettison any signs of a flexible drive from centrifuge applications.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[cswilson]

itsmoked is correct. Returning the regenerated energy to the AC line is another option. This too adds to the cost of the drive. Sometimes I'm surprised at how many people prefer to waste it in a shunt resistor, but I haven't cranked the numbers comparing added up-front capital cost to ongoing electricity savings in quite a while. In my experience, the first reason most people choose regen to line is when they can't tolerate the heat generated by the shunt resistor.

Curt Wilson
Delta Tau Data Systems

 

[itsmoked]

Good point I didn't mention cswilson! Think about all the accel energy showing up in the area of the load resistor.

A ~3000w heater running all the time.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[sreid]

1) Could you tell us how the inertia is determined. lb-ft^2 can have different meanings. I would much rather see inertia in Newton-meter^2.

2)For a TENV (Totally Enclosed, Non-Ventilated) motor driven by a VFD, you can accelerate and decelerate continously as long as you stay at or below FLA (Full Load Amps). Even with a fan cooled motor, you are accelerating to top speed (good fan cooling) and holding speed for a large part of the duty cycle so as long as the accel decel are kept to FLA, you will not overheat the motor.

 

[HCBFlash]

I can't agree w/ sried on the claim "you will not overheat the motor". If production and reliability are critical, plan on some extra cooling. Of course all of us are assuming conditions of around 10-35deg C.

"flexible coupling" like a "Lovejoy"? Shouldn't be a problem, just don't plan on using a couplings flexibility to compensate for poor alignment.

good luck!