vfd malfunction results in vibration at vfd output fundamental frequency? 3

[electricpete]

I’m posting this question on behalf of “Rotate” from maintenanceforums.com. He’s knowledgeable in vibration but not very familiar with vfd’s.

The entire thread is here:

http://maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/59520075963/p/1

http://maintenanceforums.com/eve/forums/a/tpc/f/3751089011/m/59520075963/p/2

The attachment to the first post of that thread is attached.

The basic question: Is there a malfunction of a vfd that can cause the 4-pole motor to vibrate at roughly the fundamental frequency on the output of the vfd. This frequency is slightly higher than twice the speed of the motor (for example 2.02 times). For example 50hz vibration for a 4-pole motor running at 24.75 rotations per second. If we didn’t have a vfd in the picture, we would refer to it as “line frequency” vibration (note this is not the common twice line frequency).

What type of vfd is apparently unknown.


=====================================
(2B)+(2B)' ?

 

[electricpete]

Sorry, I forgot the attachment. Here it is.

=====================================
(2B)+(2B)' ?

 

[LiteYear]

From a theoretical point of view (I have very little practical experience), there's no fault on a traditional voltage source VFD that can cause a 50Hz component on the output. It's plausible a shorted rectification diode could transmit line frequency to the output, but that would quickly result in an input short circuit condition that would have other, much more significant indications.

There is a 150Hz component on the output that arises due to the cyclic referencing of the DC BUS to each of the phases. This can become much stronger with a phase to earth or star point to earth fault in the motor. However, it's hard to imagine how this would manifest as a 50Hz vibration.

I'd be looking for something a bit more telling - the particular VSD perhaps or noise coupled into the accelerometer measurement.

 

[Strong]

According to this reference:

http://www.industry.usa.siemens.com.../approach-to-solving-motor-vibration-prob.pdf

These faults could cause "line" frequency vibrations:
Interphase fault
Ground fault
Soft Foot/Eccentric Rotor

This M.E. is interested in motor vibrations!
Walt

 

[LiteYear]

Strong I'm pretty sure that table doesn't apply if a VFD is present. Certainly the only reference in the paper to adjustable/variable speed drives is in a brief discussion about the effects on imbalance.

 

[Marke]

Hello Pete

If you have a situation where one of the output IGBTs is not being controlled correctly, that will cause a vibration at the frquency of the output of the VFD. This is usually quite violent, however it depends on how much current limiting there is on that phase.

Mark Empson
Advanced Motor Control Ltd

 

[electricpete]

Thanks for all the comments. I'm on my phone....having a tough time reading Walt's link. Will definitely look at it soon.

Mark-can you offer any comments to help understand how that frequency arises?

=====================================
(2B)+(2B)' ?

 

[Strong]

I agree that the paper I referenced is pretty light on discussing VFD. If the purpose of the VFD is to provide a variable synthesized "line frequency" to the motor, then would all of the possible electrical faults for power to the motor apply? What do you have LiteYear that supports that it would not apply? I am a dry sponge waiting to soak up some electricl knowledge!

Walt

 

[Marke]

If one IGBT is not being properly controlled, you are effectively creating a DC component in the current which will be modulated by the ripple frequency on the DC bus.
This DC component would result in a stationary torque field component, plus there would be an imbalance which would result in negative sequence fields.
The exact nature of the problem would depend on the cause, if an IGBT is turning partially ON over the full conduction cycle, there could be considerable power dissipation in that IGBT resulting in heat and changes in current, desaturation protection on the gate drive could shorten the conduction pulses without the same level of heat etc, so there are a number of scenarios that could cause an imbalance in the positive/negative conduction and this may be very small as caused by an additional switch ON delay in the gate drive etc.
I expect that the torque fields could quickly become very complex in the frequency domain.

Mark Empson
Advanced Motor Control Ltd

 

[sibeen]

I'm not a VFD person, but don't these drives have a DC offset alarm / shutdown in the case where an IGBT is playing hookie?

 

[Skogsgurra]

There used to be a problem with the resistors in the output current measurement circuit. There are three Hall transducers and a burden resistor for each of them. The burden resistor sometimes consists of two parallel resistors and if one of them opens or gets disconnected (sloppy soldering or heat from the resistor itself), you will get a lower current in the corresponding motor phase. No alarms or other warnings since the controller is satified with what it gets.
This was not uncommon in older Siemens VFDs. Haven't seen it lately, though.
If one phase of the motor is fed a lower current than the other two, you will not get a circular flux vector path, but an elliptic one. That definitely produces a vibration with the VFD fundamental frequency.
It is easy to find out if this is the problem if you use a clamp-on ammeter to check motor current balance. The VFD display cannot be trusted (and seldom shows current in individual phases), so a clamp shall be used.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[electricpete]

Thanks for all the input.

I'd think that if one phase ends up weaker than the other two or missing altogether, then the frequency of vibration would still be twice fundamental frequency, PROVIDED the weaker phase was half-wave symmetric. In this case we simply have a superposition of fundamental frequency fluxes which is itself fundamental frequency and should always result in twice frequency vibrations. An example would be unbalance supply to non-vfd motor...we know it vibrates at twice line frequency.

On the other hand, if the absolute value of a given phase differed between the positive half-cycle and the negative half cycle, then I could see fundamental frequency vibration. Perhaps this occurs if there is effectively an open circuit on the link from one pole of the dc bus to one phase.

=====================================
(2B)+(2B)' ?

 

[electricpete]

According to this reference:
[Siemens' Analytical Approach to Solving Motor Vibration Problems]
These faults could cause "line" frequency vibrations:
Interphase fault
Ground fault
Soft Foot/Eccentric Rotor

I see the chart at the end does include the frequency 60hz (along with 120hz) for those faults.

I also see the article says this:

One Times Line Frequency Vibration:
Although not nearly as prevalent as twice line frequency vibration, one times line frequency vibration can exist. Unbalanced magnetic pull may result in vibration at line
frequency (one times line frequency) as well as the usual twice line frequency vibration. If the rotor or stator moves from side to side, the point of minimum air gap may move
from one side of the motor to the other. When the frequency of this motion corresponds to the frequency of the traveling flux wave, the unbalanced magnetic pull will shift from side to side with the point of minimum gap, resulting in vibration at line frequency. This line frequency vibration is normally very small or non-existent, but if the stator or rotor system has a resonance at, or near, line frequency, the vibration may be large.

From this statement along with the symptoms which cause it, we pretty much conclude this author believes that most of the faults which cause twice line frequency vibration can also result in one times line frequency vibration.

I do have a hardcopy of another article which describes the same phenomenon in detail: "Line Frequency Magnetic Vibration of A-C- Machines" by R.C. Robinson, AIEE Transactions Part III Power Apparatus Systems Volume 81, 1962. Here is an excerpt:

Line frequency vibration was first recognizedc in a wound rotor induction motor which was operated over a wide [input] frequency range. At a certain frequency, the stator vibrated violently from side to side with an amplitude sufficient to make the stator strike the rotor. At another frequency, the rotor took on an equally violent vibration. The frequency of the vibration was exactly equal to the frequency applied from the line. The phenomenon had nothing to do with the speed of the rotor, in fact it could equally well be produced with the collector rings open-circuited and the rotor at standstill. This type of vibration ha ssubsequently been found, in varying degrees, in other ac machinery. It was noted that he magnitude of the vibration was very sensitive to applied frequency. A shift of only two cycles per second from the resonant frequency would reduce the vibration by a factor of almost 100 to 1. It was alos noted that the vibration did not consist of a bending of the stator core, but rather the entire stator moved from side to side as a unit. By bumping the side of the stator core with a mallet and then recording thee resulting free vibration it was established that the resonance frequency corresponded to the latural natural frequency of the stator core on its supports.

=====================================
(2B)+(2B)' ?

 

[electricpete]

I'd have to say based on the articles that there must be something to this one times line frequency vibration. However, I don't recall ever seeing it recently in our machines where we can typically see a twice line frequency peak on any motor vibration spectrum if we look close enough (log amplitude scale). Maybe I'll go back and take another look for one times line frequency on log scale. At any rate, it seems to me like a relatively rare phenomenon for this to build up to any noticeable amplitude on fixed-speed machines. Since it puprportedly involves a resonance, it makes sense it would be more likely to show up on variable speed machines. It doesn't match Rotate's posted scenario since he said the condition was fixed by repairing the drive (which wouldn't alter this stator core support natural frequency). But still an interesting obscure phenomenon.

=====================================
(2B)+(2B)' ?

 

[LiteYear]

I agree that the paper I referenced is pretty light on discussing VFD. If the purpose of the VFD is to provide a variable synthesized "line frequency" to the motor, then would all of the possible electrical faults for power to the motor apply? What do you have LiteYear that supports that it would not apply? I am a dry sponge waiting to soak up some electricl knowledge![/strong]

Yep, all the potential motor faults still apply, except that they will occur at the VFD synthesized frequency and its harmonics, rather than at mains frequency and its harmonics. The rectification process in the VFD effectively isolates the drive from mains frequency. There are only two elements of mains frequency which persist - the DC bus ripple (6x mains frequency for a typically 6 pulse rectifier) and the common mode alternate referencing of the DC bus to each of the input phases (3x mains frequency). As electricpete illustrates - what then happens in the motor and in fault conditions is very complex. I'm willing to believe, though am not convinced, that some scenario could cause modulation of the 6x or 3x component and produce a mains frequency vibration.

I don't have any good references for you. This field is poorly represented in the literature at the moment. As I said this is a theoretical approach and pretty much comes down to lots of simulations and application of first principles to various VFD topologies - so apply the appropriate skepticism. I've been working recently on accurately modeling the spectral content of VFD driven motor currents, in particular in case of earth faults, and one strong and consistent phenomena is the lack of a mains frequency component in the output of the VFD.