High vibration @1000rpm of a 180kW Ac Motor 7

[CelsoSC]

Hello, I had to service a Siemens Simovert VC with all three input thyristors shorted.
This VFD is used on an Atlas Copco GA180 VSD compressor.
The fault history buffer showed consecutive overcurrent trips F011.

Suspicion was a mechanical issue, and it turned out that the bearings of the motor were defective.

At the shop, they exchanged the standard bearings with insulated bearings.

It's a Siemens 2-pole, 3x400V, 180 kW motor.

After putting everything in place, I noticed a very strange behavior of the motor-VFD combination.

On ramp-up or ramp-down of the motor, I have a high vibration of the motor at about 1000 rpm with a high current peak on the output of the VFD.

This happens with or without the motor attached to the compressor screwhead.
At other rpm's the motor seems to perform with normal current readings and no vibrations.

What can I check? What kind of motor defect can provoke this behavior?

Thanks for any help!

 

[electricpete]

Without opening the motor is there a trick to discover the phase starts/ends of the winding?

Here's my suggestion. Assuming you have access to the 6 leads in the figure and they are separated, find lead pairs using resistance measurement. Then hook up each lead pair to an oscilloscope channel (+ and - inputs to each channel). Rotate the rotor manually. The rotor residual magnetism should induce a sinusoidal voltage into each lead pair (maybe 100 millivolts ballpark). Look at the phase relationship between the induced voltages... if it is 60/60/240, then that's a problem and one phase needs to be swapped end for end. Once you get 120/120/120 then the polarities between two ends of each pair of leads matches. Then the final configuration will involve hooking positive of one phase to the same termination as the negative of the next phase for the delta configuration. And of course swapping 2 of power system leads changes direction of rotation.

edison may have some other ideas.

EDIT - I wanted to add that we do a similar procedure to this to try to verify hookup direction of rotation (so we can get it right on the first try without having to bump and potentially swap leads). In that case you look for which phase is leading which (are the displayed waveforms on a wye winding showing a sequence of t1/t2/t3 or t1/t3/t2). If the oscilloscope shows rotation of t1/t2/t3 and your power system is a/b/c, then hooking up t1/t2/t3 to a/b/c respectively will cause your motor to be driven electrically in the same direction that you rotated it manually during the test. If that's not what you want then you need a swap. We use that procedure only on the large motors with high inertia where it's tough to get a rapid manual acceleration to give a definitive initial swing of the meter using a standard rotation meter. For small motors it's a lot easier to use the rotation meter. Now to the point.... why am I telling you all of this? To explain that I've only done it on large motors and never on small. And that's important because the procedure requires you to spin the rotor at a relatively constant rate in order to attain a relatively sinusoidal waveform. That's easy to do on the large motors because the large motors have relatively low friction and high inertia so when you spin by hand it spins at a relative constant rate with the inertia keeping it going between pushes. For a smaller greased bearing motor if the grease is cold then it may decelerate quickly after you each push so it might be difficult to manually spin it at a constant rate (you don't know until you try). If you have difficulty spinning at a constant rate then you might have to make some adjustments to help attain a constant rate (maybe either run the motor first to warm the grease, or possibly couple it to something that has a high rotating inertia for the duration of the manual rotation).

 

[edison123]

Before we get into the rabbit hole of phase reversal identification, a few questions.

1. Are the motor currents in 3 phases nearly equal at all speeds?

2. Do the measured speeds match the VFD frequencies? (50 Hz = 1500 RPM etc)

3. Is the speed measurement accurate and counter checked with another handheld device?

If yes to all, then we can rule out phase reversal.



Muthu

http://www.edison.co.in

 

[Pandit1]

The motor is an inverter-grade motor or a normal motor. The normal motor when operated with VFD may cause vibration in a band of a certain speed, which may be due to resonance or other reasons. The solution to the problem is to skip that speed range in VFD. Also, check the motor for any looseness in the end shields, bearing clearance, and rotor dynamic balancing.

 

[electricpete]

The normal motor when operated with VFD may cause vibration in a band of a certain speed, which may be due to resonance or other reasons.

Yes, that's a well-known phenomenon, but it does not cause increased current.

The solution to the problem is to skip that speed range in VFD.

It may indeed be a practical solution for op to block the speed range which might enable problem free operation for a long time, even if op never figures out the cause of this strange behavior. But the other side of the coin is that you fix a symptom without understanding the cause, Murphy's law says it will come back to bite you at the worst possible time. So that decision (whether to block, or try to find the cause) is up to op. It sounds like he's trying to find the cause. For selfish reasons, I personally hope he finds the explanation for this strange behavior and shares his findings!

Sorry I wasn't trying to be argumentative with you in particular Pandit1. Just using your suggestion as a jumping off point to add some thoughts.

 

[iop95]

If all VFD parameters are ok and motor wiring/connections also, may be an on controlled rectifier unit and/or in other PWM controll block of VFD.
Seems that around 1000rpm speed, VFD do not keep same V/Hz ratio, but higer, so higher current than normal; vibration may be from high oscilation/variation of DCbus and/or PWM duty, particulary around that speed.
I'll do some tests:
- test VFD on other motor (better the same but not necesary / even a small one - need to modify motor VFD parameters); same beheavior confirm a VFD problem otherwise motor
- if can't test with other motor:
- increase accelerating time
- start and check motor voltage with a low impedance AC voltmeter or better a scope
- check in the same time DCbus voltage (can then be correlate then with the AC output voltage)
- check if V/Hz ratio is increasing around 1000rpm
- if VFD is confirmed, check DC bus voltage sensing and related circuits up to firing controll

 

[CelsoSC]

On a side note, I am an electronic engineer and am always curious.
And since I fixed the VFD, I have to give some warranty. Blowing up components because of overcurrent will get very expensive on this particular VFD. I prefer to spend some time trying to understand what is happening. I'm not an electric motor expert at all; I'm searching for help and trying to learn as much as possible. I'm so glad that this forum exists with nice people trying to help. A big thank you to all!

@electricpete
Thank you for the tips to measure the phase of the windings; I will do it next week. And I know Mr. Murphy very well, too.

@Pandit1
I still stick to my statement that the VFD and motor settings were matched at the factory and no modifications were made to the parameters of the VFD. I serviced this compressor some time ago, and when it was working normally, it operated in the range of 800–2500 rpm, and the motor worked very smoothly.

Could the exchange of the standard bearings with insulated bearings have changed something that is now showing resonance at 1000 rpm? If yes, I'm happy to set a skip frequency.

Is it the wrong phase of the winding? I'm happier, it's easy to fix, and the old conditions are met.

Something more? I have to ask to take the motor out and send it to the workshop, and it's a nightmare to take the motor out of the compressor. 1000 kg and no point to hook up anything.

 

[edison123]

Insulated bearings has got nothing to do with vibration, resonance etc.

If your other speeds are fine, just skip over the 1000 RPM coasting up and down as the drive gurus like iop are saying.



Muthu

http://www.edison.co.in

 

[electricpete]

Could the exchange of the standard bearings with insulated bearings have changed something that is now showing resonance at 1000 rpm? If yes, I'm happy to set a skip frequency.

Insulated bearings could reduce the bearing stiffness which alters the resonant frequency (it would generally shift the resonant frequencies down lower in frequency than they were before). To my thinking that wouldn't explain the increased current UNLESS it is such severe vibration that the rotor contacts the stator enough to increase the torque load significantly. I can recall we had a small motor (something 5hp) that kept tripping on overload and when we investigated we found the bearing fits severely degraded and rotor rubbing the stator (apparently enough to cause overload tripping). I've never seen that scenario on a larger motor like yours (I've seen rubs but never accompanied by a high current). It doesn't seem likely that rubbing scenario could continue for very long without revealing itself in some other way (a lot of noise, possible smoke from mechanical rubbing, eventually damage the windings and cause a ground fault trip). Nevertheless it's the best scenario I can some up with to explain what you're seeing.

Some thoughts if you wanted to investigate that particular scenario further (rather than just blocking). If you are able to listen to the motor when operating in the problem speed that might provide a clue if that scenario is occurring. If you have stator winding temperature indication, it might show up by comparing no-load stator running stator temperature with previous no-load stator running temperatures since the stator core damage from a previous rub could increase core losses and increase stator temperature (and it would be easier to notice small difference at low load than it is at high load). Certainly if you disassembled the motor you'd know for sure whether or not rub had occurred but I know you mentioned that's not an attractive option. If it's an open style motor and you might be able to look directly into an air intake at the end and see the ends of the airgap which might (if you're lucky) show some type of evidence of a rub (although not seeing anything from that limited view wouldn't disprove rub, and most open or open drip proof motors don't provide this direct access to view the airgap).

 

[edison123]

Insulated bearings have a very thin veneer of coating on the outer race. I don't see them altering the bearing stiffness or motor resonance in anyway.

OP
Did you sort out the issue? A feedback would be nice.

Muthu

http://www.edison.co.in

 

[electricpete]

> I don't see them altering the bearing stiffness or motor resonance in anyway.

I have no proof one way or the other. The coating is effectively a series spring, so to the extent that material is any softer then the other bearing components, the associated coating stiffness would have an an outsized effect on the overall bearing stiffness (series bearing stiffnesses combine in the same manner as parallel electrical resistances).

Either way, it doesn't change my recommendation. The only cause I can come up with to explain both increased vib and increased current at one particular speed is resonance severe enough to cause rub. Hence the recommendations targeted to investigating that possibility of resonance causing a rub.

 

[zlatkodo]

I don't believe that the cause could be a phase swapping, as the consequences would be significantly worse.
I think the cause should be sought in the settings of the VFD.
ACW

 

[CelsoSC]

Hello edison123,

Motor was taken out of the compressor and now is on his way to the workshop.
Hope to have some news soon.

 

[edison123]

Ok, good opportunity to check out the windings thoroughly. Ask them to do surge comparison testing of all circuits of all 3 phases with the rotor pulled out of the stator. This test will deduct phase reversal, turn shorts, phase shorts etc.

If the shop has a VFD, then you can test the motor for your speed range on no-load.

Muthu

http://www.edison.co.in

 

[CelsoSC]

Hello, today I can an update that the motor was tested fine.
It was tested under full load on a test stand and performed OK.

Well, VFD it is.

Wish me luck.

 

[CelsoSC]

Yesterday, before the old motor was delivered, I made a factory reset on the VFD.
Connected in star a 30 kW,4 pole motor to do some tests on the VFD, and the motor worked normally. I checked all firing pulses of the IGBTs, and gate charge and discharge times were fine too.

Today I hooked up the motor again and set the Motor as on the Motor plate:
400V delta, 310A rated current, 76 Hz, 0,82 CosPhi, 2255 RPM.

The motor started sluggish, and the current surged to the limit very fast.
Motor was connected in delta.

Just to test, I connected the motor in star (removed U2, V2, and W2 and connected them together), and to my surprise, the motor started normally, and the current was about 35A to 45A from 10 Hz to 50 Hz.

What I'm missing?

 

[wayne440]

(claiming no expertise) So for whatever reason the VFD "doesn't like" delta connected loads, or the motor has some fault that only shows up when delta connected. Can you try the 30kW motor in delta configuration? If so, that might point you to one or the other situation.

Edited to fix "brain fart".

 

[kflan]

Although the nameplate of the motor at the top states 400v @ 76Hz, for some reason, from your findings, that V/Hz ratio does not appear to be working. At the bottom of the nameplate, it states 400V @ 103Hz. Have you tried that V/Hz ratio? This motor should be connected delta and more than likely was ran at the repair shop on delta. I am curious as to what V/Hz ratio the repair shop ran the motor on and loaded the motor. I assume there was some sort of report? I think your hitting current limit on the drive too fast with 400v/76hz and it is making the motor struggle.

 

[Parchie]

@kflan,
Therefore, when the motor was run at 400/76Hz, the V/Hz =5.25 V/Hz, the ratio will be less at 400v/103Hz = 3.88 V/Hz!
Needless to say, operating at 400V at 76 Hz, the motor operates at a higher flux (flux is proportional to the V/Hz ratio). Any slight airgap differences between the rotor and the stator will be exacerbated hence the higher vibration.

 

[CelsoSC]

Hello Kflan,
Thank you for the detail, I overlooked it.
I tested with 400V delta at 103Hz setting and it worked!
Current with no load is higher at 80A to 95A.

Would like to know how to set up the correct Motor magnetization current?

 

[CelsoSC]

Hello,
Finally, I can close this topic.
It turned out that it was a firmware issue on the CUVC board.

The motor started and operated normally after replacing the board with a spare and uploading the parameter set.
I found that the firmware versions on the old and new boards were different when I compared them.

The old board also functioned as it should once I updated it with the most recent firmware and uploaded the parameter set once more.
No more vibrations!

I sincerely appreciate all of your help!