HV Motor damage - Root cause analysis 2

[jmmeunier]

Hi,

We discovered damages on a HV Motor having the following parameters:
Rated voltage: 6kV
Power: 5500 kW
Freq: 50 Hz
RPM: 1490
Operation: water pump (power station)
Airgap: 3 mm

Pictures from the stator, the rotor and the bearings are presented in attachement.

- This motor was rewound, the bearings were machined and the rotor pressure plates were replaced.
- The damages were discovered in 2011 during a minor overhaul, after 1000 hours of operation.
- Nothing was noticed during operation.
- The vibration analyses did not indicate any eccentricity, broken rotor bars or abnormal vibration.
- No problem with the temperature of the oil was registered during operation.
- The Ring test indicates hot spots on the damaged zone of the stator core.
- no element is missing inside the machine.
- the melted metal on the rotor is made up of same composition as the metal of the stator and the rotor core.

The measurement results (motor is not permantly monitered) do not indicate any problem during operation, but it is clear that the rotor has rubbed the stator.

Do you have an idea about the possible root cause?
- eccentricity not always present?
- bad lubrification of bearings?
- ...


Thank you for your help.

 

[krsh]

From the pictures you attached i say the bearings are the root cause and the fact that the rotor has rubbed the stator is only a side effect. This problem occurred because of the poor construction of the bearings or misalignement between the bearings and the shaft. Another cause could be incorrect load distribution on the shaft but that would have led to vibrations.
I hope this explanation helps you.

 

[electricpete]

Those are great photos.

Bearing failure certainly is something to consider. . We can see clearly there is something wrong with the bearing.

FWIW, my initial reaction (right or wrong) is that there is not enough bearing degradation to cause rotor/stator contact. Consider the bearing clearance may be something like 0.008” diametrical, while the airgap may be 0.060” or more on each side. (do you have exact numbers and particularly as-found airgaps checked in 4 locations?) . Also, if bearing clearance significantly enlarges or is significantly misaligned, we MIGHT expect to see a more regular pattern of contact all around the rotor. (not guaranteed… perhaps if rotor bow it might only contact one point).

My alternative suggested scenario for consideration is that the rotor has touched the stator intermittently during start due to flexible rotor design and high magnetic pull during start. That could explain the chaotic and seemingly unrelated locations of the various rubs.

1 - Is this DOL start motor?
2 - Is this 2-pole motor?
3 – Is damaged bearing on different end than the rub?
4 – Is there damage on the other side of the rotor too (now shown?)
To be consistent with my scenario, 1 should be yes. Perhaps 2 should be yes (less likely to have pullover on slower speed motor). 3 and 4 not as relevant but yeses may be more aligned with my scenario.

The other feature that is curious is heating of the stator coil area between the rubbed laminations in lower left corner of 1st page. Some possible explanations:
A – Rotor/stator rubbed for a long time causing heating of these coils… not consistent with my suggestion.
B – Rotor/stator rub damaged stator laminations, which overheated adjacent coils over time during subsequent operation. I vote for this explanation, although not positive.

By the way, you mentioned monitoring of bearing oil temperature. Was bearing metal temperature also measured (thermocouple inserted to bottom of recess in back side of bottom bearing). I would have expected temperature measurements of metal temperature to spike whenever the bearing damage occurred since melted babbit is evident in the distribution groove.


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

 

[electricpete]

Now I am looking at those stator marks again. The circumferential stripes in the stator damage pattern correspond to the rotor vent ducts, correct? Then these would be the areas where rotor/stator contact did NOT occur. Oddly, in the lower right photo on the 1st page, the striped areas show as brown from overheating. Why show up brown/overheated in an area that didn’t contact? Maybe these areas where just overheated from adjacent areas and the adjacent black represents much more severe overheating?...not sure.

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

 

[jmmeunier]

Hi Krsh and Electricpete,

Thank you both for you replies.

Krsh,
The babbit was machined in 2009. Do you think that particles from the babbit could have torn away the bearings and damaged the babbit surface?

Electricpete,
The bearing clearance is 0,0095 inch (0,24 mm). The depth of the marks is about 0,039 inch (1 mm).

The patterns observed on the stator and on the rotor cores perfectly fit. To me the rotor obliquely rugged the stator.

To answer your questions:
1 - Is this DOL start motor? to be checked but normally yes
2 - Is this 2-pole motor? 4 poles (squirrel cage)
3 – Is damaged bearing on different end than the rub? both bearings are damaged but the marks are more important on the bearing which is on the same end than the rub.
4 – Is there damage on the other side of the rotor too (now shown?) no there is no apparent damage on the other side.

To me, the rotor/stator rub damaged stator laminations, which overheated adjacent coils over time during subsequent operation. Indeed the ring flux test which was performed afterwards revealed that the laminations are shorted in the damaged zone. The brown corresponds to burnt resin.

I will check for information about the temperature of the bearings.

Kind Regards,

 

[electricpete]

As clarification, it is the “wedges” that we see overheated, not “coils” as I said. But it makes sense the smeared laminations would heat them over time after the rub occurred.

Now another question. In the stator orientation shown, the rubs occur on the side of stator (rather than top or bottom). Is the stator in the same orientation as during operation? This is a horizontal sleeve bearing machine. You may be familiar for sleeve bearings, if we look from the end where the rotor goes CCW, the shaft will move toward left during start (due to dry friction), and then will settle to the right during operation (due to oil wedge action). If the rub is on the stator side, I would try to figure out whether it is the side where the rotor tends to move by dry friction during start or by oil wedge during run. That is not at all definitive, just a data point which may provide a slight nudge toward one conclusion or antoher. (I realize that electromagnetic forces can be much bigger than anything the bearing friction/wedge creates…. but those forces become strongest when rotor is pushed offcenter within airgap which may be initiated during start.)

Also just for pure information, it would be interesting to map the TIR along the rotor, noting angle of the highest point as you travel axially.

You have 2 different areas of contact on the rotor (one on the end, and one a little bit closer to center) which appear separated by approx 90 degrees from each other. Are the associated contact areas on the stator also separated by 90 degrees?

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

 

[jmmeunier]

The rub is on the left side when looking from the end where the rotor goes CCW.

I also think that the problem was intermittent because the vibration measurements - which are never performed during the motor start phase - do not show any (permanent) eccentricity.

 

[electricpete]

The rub is on the left side when looking from the end where the rotor goes CCW.

That is the location that tends to be consistent with rub on starting. Because the dry friction between shaft and bearing will initially cause the shaft roll uphill to the left, then the mag force from the huge starting currents may pulling it further in that direction. In contrast, for simple bearing wear, you might expect to see the rub somewhere near the bottom or slightly to the right. At any rate by no means definitive, keep an open mind to both possibilities and others we haven’t thought of.

I mentioned before flexible rotor… actually that would generally tend to result in rubs more toward the axial center I think. Even without flexible rotors, some motors are more prone to rub based on electrical characteristics (for example single-circuit wye stator winding more susceptible than winding with multiple parallel circuits)

The fact that you had both bearings damaged is interesting… that argues against a bearing failure as an initiator of the event unless there was some common cause of both bearings failing. (is there is a common oil system feeding the bearings? … I’d guess instead the bearings have their own sump). More likely there was for some reason a high vibration for a brief period of time and the rotor/stator rub and bearing damage both occurred at that same time….. one or the other could have been contributor to the high vibration. Contributors may have been EM forces of start, or could be some other factor… resonant vibration as pass through a certain speed during startup or shutdown…. or some weird temporary mechanical disturbance from connected load or environment.

I am sort of brainstorming here. Others welcome to chime in their ideas.


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

 

[electricpete]

The rub is on the left side when looking from the end where the rotor goes CCW.

That is the location that tends to be consistent with rub on starting. Because the dry friction between shaft and bearing will initially cause the shaft roll uphill to the left, then the mag force from the huge starting currents may pulling it further in that direction. In contrast, for simple bearing wear, you might expect to see the rub somewhere near the bottom or slightly to the right. At any rate by no means definitive, keep an open mind to both possibilities and others we haven't thought of.

I mentioned before flexible rotor... actually that would generally tend to result in rubs more toward the axial center I think. Even without flexible rotors, some motors are more prone to rub based on electrical characteristics (for example single-circuit wye stator winding more susceptible than winding with multiple parallel circuits)

The fact that you had both bearings damaged is interesting... that argues against a bearing failure as an initiator of the event unless there was some common cause of both bearings failing. (is there is a common oil system feeding the bearings? ... I'd guess instead the bearings have their own sump). More likely there was for some reason a high vibration for a brief period of time and the rotor/stator rub and bearing damage both occurred at that same time..... one or the other could have been contributor to the high vibration. Contributors may have been EM forces of start, or could be some other factor... resonant vibration as pass through a certain speed during startup or shutdown.... or some weird temporary mechanical disturbance from connected load or environment.

I am sort of brainstorming here. Others welcome to chime in their ideas.


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

 

[electricpete]

Or if this machine is VFD driven, it may have operated at resonant speed for some period of time.

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

 

[ScottyUK]

Stator core movement? Certainly not the first time a core pack has moved within a frame, although I have never seen movement which caused a rub.



----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[jmmeunier]

Thank you for your replies!
rem: sorry for my english: that is not my mother language.

WHAT IS CERTAIN:
A) Stator/Rotor
a1 - the motor is DOL started
a2 - the rotor rubbed the stator on NDE on the left when looking from CCW side.
a3 - the recent ring test showed hot spots on the whole damaged zone of the stator core.
a4 - Melted metal is found on the rotor core. This metal is of the same composition as the stator and the rotor cores (as if the rotor stayed against the hot stator at the still stand)

B) Bearings
b1 - following the measurements performed after the reparation of 2009, the rotor clearance was correct at this moment (0.0095 inch). There is no information about the current clearance.
b2 - the recent vibration analyses did not show any vibration or eccentricity during operation (no measurement during start)
b3 - both bearings are damaged: the NDE bearing is maybe more rubbed
b4 - the oil system feeding is common to both bearings


ASSUMPTIONS
Considering the initial bearing clearance 0.0095" and an airgap of 0.118", the rotor should not reach the stator if the alignment is correct.

So it seems clear that, at a moment:
- either the bearing clearance has enlarged
- or the rotor was misaligned
Anyhow this phenomenon was not permanent but long enough to melt the stator core surface.

Scenario 1: bearing clearance enlargement
The addition of metal to the bearing (in 2009) was maybe poorly done. The shaft moved toward left during each DOL start and could have degrade the metal surface of bearing in such a manner that after a certain time the clearance was large enough to allow the rotor to rub the stator. The current clearance should be checked to know.

If the problem was due to dirty oil, shouldn't it be the case for other motors in the power plant?

Scenario 2: intermittent misaligment
Note that the bearings are external to the motor frame and the alignement is performed on site.
The bearing could maybe float inside their frame (?). During starts they could then not be correctly aligned anymore because of the fact the shaft moves toward left.

Other scenarios:
3 - a piece felt in the airgap and pushed the rotor against the stator ? not likely because no such piece was found in the machine and another type of degradation should have been found on the stator core
4 - The stator core deformed sufficiently to be rubbed by the rotor. However this deformation should only be intemittent and only on one side. Is it possible?
5 - The issue comes from the load (water pump)?
6 - ...?

Thank you for your opinion and you help

 

[zlatkodo]

Does this motor is driven by VFD?
Zlatkodo

 

[jmmeunier]

Hi Zlatkodo,

No it isn't...

 

[krsh]

After your updated info i am 90% sure that the bearings are the root cause. It is clear for me that the repairs done in 2009 led to this problem. Normally you should have replaced the damaged bearings with new ones not to repair them. The added alloy to the bearing separated and that led to cavitation on the bearing circumference and unbalanced the shaft and you know the rest... The melted metal on the rotor is only a result of friction.
Give this explanation to managers and maybe next time they will give you the money to buy new replacement parts instead to repair them.

 

[electricpete]

I will say that I have seen bearings with damage that looked almost exactly like this, and it occurred after operating at elevated vibration levels. In our case, the bearings were a symptom, not a cause (the cause of the elevated vibration was thermal rotor bow).

fwiw I am 90% certain that the machine must have seen elevated vibration levels for some period of time. The reason is: you don’t get wear up around the top of the bearing unless that rotor is orbitting at almost the full clearance of the bearing.

The cause of that high vibration could possibly have been bearings (which affect rotordynamics) or could have been something else discussed or not discussed above.

By the way, you mentioned you have surveyed vibration and didn’t see anything unusual. I’ll bet you were using housing accelerometer (rather than prox probes). That’s what we do also for most of our sleeve bearing machines. I have heard many people say that there can be a lot going on in a sleeve bearing machine that doesn’t show up on the housing accelerometer readings. In our case, when we had a rotor anomaly that resulted in bearing damage, we saw seen evidence on the housing, including change in 1x and harmonics but I’m sure there are a lot of variables that affect the response. Perhaps you can share some of your vibration data?

Also, I was curious if you have measured as-found airgap? If not perhaps the opportunity is gone. However, you can still at least map out the runout of the rotor (magnitude and orientation relative to keyway) along the rotor as part of piecing together the whole picture. Also if you have a coordinate measuring device you might get some info on the stator... although that’s trickier to do and less likely to have a problem there.


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

 

[jmmeunier]

Krsh, Electricpete,

Thank you for your answers!!!

I also tend to think that the added metal has been eroded on the side of the bearing (there are not only scoring). The current bearing clearance should be measured to get an answer.
On the top of the bearings, there are only scratches probably done by metal particles detached from the bearing.

Herewith attached are some vibration results on the motor before and after the reparation (accelerometers).
I add that, from time to time, current signature analyses are performed. They don't show any problem.
The airgap was not measured after the problem was observed.

Thank you!

 

[electricpete]

On the top of the bearings, there are only scratches probably done by metal particles detached from the bearing.

See 4 questions attached.


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

 

[electricpete]

Also by looking at the bearing, there are no oil ring groove in top of the bearing so it relies completely on pumped oil. Is it possible machine was ever started without oil... and then oil applied by shaft pump after started? (Normally there is electric oil pump that should be running before start). Not likely that such a scenario would occur but just figured I'd mention it in case you might have some computer recordings of oil pressures or oil alarms which can be overlaid against pump run/stop times to for double-checking.

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

 

[electricpete]

Correction: "oil ring groove" should've been "oil ring slot" (the slot that the oil ring passes through... different than a distribution groove).

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