Vibration On AC motor with VFD

[mrr1671]

Hello,
I am not an Electrical Engineer but I need to measure the vibration on a motor that is controlled by a VFD for variable speed.
I have read that a VFD converts the AC current to DC and then converts again into AC but something that is called modulated (PWM), which gives a different frequency.
I have some questions regarding this:

First question is if the motor design (electrical) needs to be different than an AC motor running on a fixed frequency?

Next will be if minimizing all the mechanical sources for vibration, like rotor unbalancing, shaft runout, etc, the motor still vibrates on some frequencies?

Is there any relation or conversion between PWM and frequency (Hz)?

What is best to measure vibration on this case, Vel, Displ or Acc?

Any advice will help me a lot on this matter.
Thank you.

 

[thewellguy]

Is this a homework question?

 

[mrr1671]

No it's not, it looks like but probably because I am Mechanical Engineer and don't know too much about the electric motors.

 

[jraef]

Motor design does not HAVE TO be different, but it's a good idea (best practice) to use what are now referred to as "inverter duty" motors if at all possible. The differences are in the insulation material and cooling systems. Non-inverter duty motors are more prone to failure from a variety of issues related to being connected to a PWM power source such as an inverter drive. Further details on what those are can be found by searching these fora for "inverter duty motor".

Because the motor can now operate at any theoretical speed (within limits), it can begin to have resonant interactions with mechanical harmonics from any number of sources that may have always existed in a load system, but were never a problem before because the motor accelerated through them quickly. For that reason, most VFDs have what is called "critical frequency lockout" (usually more than one) where they can be told to NOT run within a speed range in which you have found mechanical resonances or harmonics.

Don't know what you are asking about with the term "relation" in that next question. Please elaborate.

Same for the last question; "best"? Best for what? Those 3 metrics are looking at different issues.

 

[waross]

Pulse width modulation is a technique for switching DC voltage to construct an AC current sine wave. There is a switching frequency, possibly 5KHz. and a resulting frequency that is variable from low to over 60 Hz. To compare PWM to frequency is a little like trying to compare apples to adverbs.
respectfully

 

[CJCPE]

In early PWM inverter designs, there were a specific number and placement of pulses per cycle of output. Therefore the PWM switching frequency was an integral multiple of the output frequency. Sometimes the PWM pattern changed at specific output frequencies and a distinctive "gear shift" sound could be heard when the pattern change took place. Now there is generally no relationship between the PWM switching frequency and the output frequency.

There is of course some relationship between the amount of modulation and the output frequency since the basic purpose of the modulation is to provide the optimum stator voltage which is roughly proportional to the stator frequency.

 

[CJCPE]

The harmonic content in the output waveform can cause torque pulsations ("cogging"). The only way that I know of to measure that is by driving the load through a shaft torquemeter. The resulting speed pulseations could possibly be measured with a tachometer. I believe that this rarely if ever occurs at a high enough level to causes problems.

 

[thewellguy]

mrr1671

What is the reason for wanting to measure vibration?

If you are doing it just for historical info then you need a plan. For example at what hertz/speed will you record? Does the system change alot or only run at 2 or 3 present speeds. Where on the motor are you going to measure? Can a vibration meter be used? This is the easiest

If you are trying to "lock out" frequency that are causing vibration then go through then, measure and lock

Hop this helps, sorry about the homework remark.

 

[Skogsgurra]

This might not be what the OP is after. But I have seen the problem several times. So, why not?

Bearings are monitored so they can be replaced when vibration levels start approaching critical numbers. The measurement is mostly done using velocity pick ups and level is mostly in millimeters/second or the corresponding imperial speed unit.

Quite often, especially when the measuring equipment is using the so called envelop technique (which also is called high frequency measurement) there are problems with variable speed drives. I have been to places where the apparent vibration level is up to tens of mm/s, but no vibration can be felt or heard. That is usually due to EMI from the motor cable that leaks into the vibration pick-up and cabling.

Sometimes there are also magnetoelestic vibrations from the PWM voltage that makes the stator iron vibrate at a high frequency (this you can hear quite well). This frequency is then picked up and the algorithms in the measuring equipment tries to transpose the HF energy down to "bearing frequencies" and the result is often very confusing since there is no correspondence between the resulting spectrum and what the bearing was supposed to do.

Solution is often to just ignore the spectrum - or go back to classic "measure in baseband" technique. Sometimes a motor filter helps. But du/dt is not enough. It has to be a sine filter - or close to one.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...