choice among two theories for a random winding turn failure 1

[electricpete]

A random wound 460 vac motor failed by turn to turn short. Only one phase and one coil was involved.

The failed coil happened to be a line-end coil, but 25% of coils in the machine are line end coil and steep voltage surges are not expected in this power system. So we ruled out surge as a cause.

The motor had been in service about 30 years service. During inspection, it was reported some parts of winding were dry and brittle, similar to what is expected after 30 years service.

Slides 1 and 2 show overview of endwinding and failure area.

Based on inspection, two different people opined two different apparent causes of the failure.

Scenario 1 - as highlighted in slide 3, the slot fill appears low. General looseness of winding from shrinkage and/or low initial slot fill are proposed as cause of failure. Note that in slide 4 the particular coil that failed does not have same appearance of low slot fill but it is postulated that this coil moved.

Scenario 2 - as highlighted in slide 5, the coils immediately bend as they exit the slot (rather than bending a little ways outside the slot). It is enough to distort the phase paper (which makes the looseness look worse than it really is from the end, even though it is admittedly loose looking). The cause of this early bend may be result of coil forming (the slot sections are not long enough, the end span is not long enough) and it also may be a result of winding assembly process (no attempt was made to establish a bend further outside the slot prior to dip/bake). Either way it places a tremendous stress on the coil in that area, which can be aggravated by thermal changes and vibration.

Among scenario 1 or 2, which seems more likely to you?

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

 

[electricpete]

Scenario 2 - as highlighted in slide 5, the coils immediately bend as they exit the slot (rather than bending a little ways outside the slot). It is enough to distort the slot liner phase paper (which makes the looseness look worse than it really is from the end, even though it is admittedly loose looking).

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

 

[waross]

Hi Pete. Both perceived issues of construction remained in service for 30 years before failure. I doubt that either one was the cause of failure.

During inspection, it was reported some parts of winding were dry and brittle, similar to what is expected after 30 years service.

I suspect that this condition plus a little moisture was the cause of failure.
Why close to the slot ends?
Moisture and or condensation and tracking across the end of the stator iron.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[electricpete]

Thanks Bill. The motor is indoors in a controlled environment, has a space heater automatically switched in whenever the motor is shutdown, and the problem happened to show up after the motor had been running for weeks (not soon after start). So moisture does not jump out as a likely cause for this one.

I am particularly interested in discussing the relative merits of these two scenarios, neither of which is definitive of course… just the best guesses from available info.

The timing does play a role in how we view these as you say. Scenario A allows for the possibility that looseness may have gotten worse over time based on the thermal aging which was apparent during inspection. Both scenarios provide a mechanism whereby stresses from vibration are magnified so that it might take awhile for vibration to degrade the winding to the point of failure.

Regarding coil looseness, at first view it seems very loose from the photos, but then considering that the slot liners are bowed outwards abnormally, the view from the end might be deceiving in that respect. What do you think?


Attached I included photos of two other random wound motors I grabbed randomly from my files for comparison (not sister motors, not even same OEM). The first one definitely has the ideal prounounced bend well outside the slot. The second one does not (although has not failed). The looseness does not appear as severe fromo the end in either although again it may be misleading since the slot liners are not as far bent out as the failed motor


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

 

[dArsonval]

Pete;

It’s likely safe to rule out a surge as suggested because that type of failure tends to be more catastrophic in visual appearance.
The coil exiting the iron would look more severed, and have more copper bead debris like the one small one seen in your photo of the failure.
I would couple your Scenario 2 with an additional factor. “Workmanship”. It kind of looks like the motor was rewound at some point in its operating history.
A nicked wire nestled among that failed coil eventually vibrated with a neighboring conductor (helped by the loose slot fill) causing what happened.
The coil’s sharp bend exiting the iron slot can often not be avoided due to lack of physical space available within the end bracket of the motor.
The “cuffed” slot liners certainly helped provide additional protection in making the exiting formed bend.

If this type of failure occurred on a new or recently rewound motor, it could be promoted a case for warranty.

John

 

[zlatkodo]

Hi, Pete,
First, you should check the value of "volts per coil."
Such a damage occurs very often in case of random wound coils , if this value is excessive.
I suppose it is about 4 pole winding in 48 slots, but the number of parallel circuits is unknown . If this winding is done with 4 circuits for example, this could be a cause.

Winding Design and Repair

 

[Hoxton]

It is 30 years old!!

Just how long did you expect it to last!

In the real world nowadays we go for 15 to 20 years design life. Very often this is longer than the life of the plant.

 

[electricpete]

The “cuffed” slot liners certainly helped provide additional protection in making the exiting formed bend.

I'm not sure what you mean. Slot liner looks pretty flimsy and I'm thinking it provides nothing more than abrasion protection. On the other hand (now that you mention it) the top stick looks a little more rigid and may have played a role of how some of the uppermost strands bent outside the slot. Can you talk a little more about what you meant?

First, you should check the value of "volts per coil."
Such a damage occurs very often in case of random wound coils , if this value is excessive.
I suppose it is about 4 pole winding in 48 slots, but the number of parallel circuits is unknown . If this winding is done with 4 circuits for example, this could be a cause

Thanks, it two-circuit delta 4-pole machine. 48 slots. Phase to phase voltage dropped across 8 coils. I guess that makes it 460/8 ~ 58 volts/coil. Random wound, so of course full coil voltage could appear across physically adjacent turns. But I can't say I have ever heard of a motor failing from excess steady state turns per coil. And these motors are from a respected OEM. We have 6 of these motors and the first two failures among them occurred after approx 30 years (180 motor years) service last year and this year. We didn't take that close a look at last year's failure but looking closer at this one.

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

 

[electricpete]

But I can't say I have ever heard of a motor failing from excess steady state turns volts per coil

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

 

[electricpete]

Some other factors that hopefully explain where I'm coming from (I didn't think it would be necessary to explain the world to ask a simple question).

We have already completed our formal evaluation and put in place our corrective actions.
The actions include proactively rewinding the older motors in this family over the next several years to address age.

The reason for the question is primarily a matter of curiosity. In the back of my mind I'm also wondering whether I might be able to effectively priortize the rewinds by pulling back the endbells and looking at the endwindings to judge these conditions.


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

 

[waross]

So I am not looking for further input from others who don't have the complete information.

Inspect the other slot ends for any sign that a similar failure is starting. If either of those scenarios is correct it is likely but not a certainty that other slot ends are near failure.
with the information that we have been given I would have to say
"None of the above."
From the information that I have, which is mostly pictorial, I see several turns in the winding that are the color of annealed copper and appear to have the insulation completely baked off. I see several other turns with insulation that appears to have been burnt black from heat.
A single point failure tends to heat all winding conductors equally. The winding conductors in this instance seem to have not been heated by a point failure.
I see many conductors that appear to have good insulation.
I see several conductors that appear to have severe heat damage to the insulation. This damage appears to extend for several inches from the slot end.
I see several conductors that appear to have the insulation completely burnt off and the copper annealed by excess heat.
These may all be indications of a turn to turn short internal to the coil.
The heat developed by shorted turns eventually led to damage to the slot liner and flashover.
I understand that your plant runs ungrounded. That would explain the relatively little arcing damage to the stator.

"Among scenario 1 or 2, which seems more likely to you?" "I am particularly interested in discussing the relative merits of these two scenarios, neither of which is definitive of course… just the best guesses from available info."

Based on the information that you have given I have to say that the likely hood of either of the scenarios causing that damage is so ow that I have to reply;
"None of the above."
Based on the available information the motor has been in service for 30 years. Those conditions have not caused a failure for the first 40 years.
I have to reply;
"None of the above."

Other information that may not have been given enough consideration.

We have 6 of these motors and the first two failures among them occurred after approx 30 years (180 motor years) service last year and this year.

Surges have been ruled out.
BUT I understand that this may be an ungrounded system.
You may have a high impedance, discontinuous ground fault or an intermittent high impedance, discontinuous ground fault.
Discontinuous ground faults have been known to superimpose high voltage high frequency potentials on ungrounded system.
The fault may be steady but it may also be a transient related to switching and/or motor starting.
A common feature of this type of fault is random motor burnout throughout the plant.
A common feature of this type of failure is winding turn failure near the line end of the winding.
I do not have enough information to rule this possibility out, and this may be much more likely than either scenario #1 or #2.
This type of failure may often initiate turn to turn shorts buried in the coil bundle. From the pictures this may have been a possibility.
Pete, my friend, please take a walk around the plant to cool off before you flame we with your response.
Respectfully.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[electricpete]

I'm not sure what you read that possibly lead you to believe that I am looking for random brainstorming of possible causes of the failure. I thought I already make that clear. I'm not sure what else I could possibly say to make it any clearer. Thanks anyway for your response.

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

 

[waross]

I am trying to say that if either one of the scenarios was the cause of failure, the motor would not have lasted 30 years.

I am particularly interested in discussing the relative merits of these two scenarios, neither of which is definitive of course… just the best guesses from available info.

Neither one.
The heat damage to the windings may not be consistent with either guess.
The other slots don't seem to show any damage from low fill.
Other tight bends don't seem to show any damage.
My discussion of the relative merits of these two scenarios is that they don't have much merit.

A suggestion that may save some money. We had some generators severely damaged by being flooded by salt water during hurricane Mitch. We were able to wash, dip and re-bake the generator windings for a saving of thousands of dollars. depending on your evaluation of motors for future rewind, this may save some money particularly on larger sized motors.
Respectfully.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[electricpete]

Thanks for your comments. However I disagree.

I am trying to say that if either one of the scenarios was the cause of failure, the motor would not have lasted 30 years.

Both mechanisms can degrade the strand insulation over time. We can get many individual strands damaged before a turn fault occurs (it takes two strands damaged at the same location for turn to turn fault). There is no time limit for these conditions to result in failure. Furthermore as noted the looseness seen today may certainly be worse than it was initially. And if I'm trying to decide whether inspecting for looseness is effective it matters not whether the looseness was there initially or got worse over time.

The heat damage to the windings may not be consistent with either guess.

The heat damage is entirely consistent with a turn to turn short.

The other slots don't seem to show any damage from low fill.
Other tight bends don't seem to show any damage.

I don't expect to see any other damage... all the visible damage unique to that coil appears to have occurred after the turn to turn short. After the first turn to turn short motor tends not to last very long so I don't expect to see any other turn to turnh shorts in the same motor.

Failure analysis is certainly not an exact science but again I have gotten the best available input based on a lot more than what was presented here. I'm not expecting anyone to have the clarvoyance to know the cause of failure from those photos. It was a simple question. Please take it at face value.


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

 

[electricpete]

It was brand new to me when the one guy mentioned the sharp bend coming out of the slot. Personally I have heard of a lot of motor failures attributed to looseness but I haven't heard any attributed to this. Curious if anyone else has heard of this as a contributor to a failure of random wound machine. And also curious if rewinders would consider the bend shown in the photo bad (I did get some input from John who pointed out a possible challenge in establshing that bend).

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

 

[Strong]

I vote for Scenario 2.

Walt

 

[electricpete]

Other tight bends don't seem to show any damage.

I was thinking about that. The motor is not yet rewound. Before they put it in the bakeout oven, I'll ask them to remove a topstick from a healthy slot, cut the top coil at both ends and pull it out of the slot strand by strand (maybe with a little local heat to soften the resin). It's probably a longshot to be able to distinguish any mechanical rubbing damage from the damage that occurs from pulling the strands out, but you never know. I'll ask them to try to quantify the slot looseness in the process also.

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

 

[biggeorgetechnical]

Dear electricpete,

I was wondering: How sure are you that surges were essentially absent? Do you have some kind of scans on the voltages?

Thanks,

George




"When you go looking for problems, you will find them... But they may not be the problems you went looking for!"

 

[hvmtr]

Pete:

Of the two scenarios I have to choose #1. In my opinion both are weak and probably not directly related to the failure. I'm not concerned with voids in the slot however if strands are loose that could be a contributing factor. Have to checked/monitored the THD?

 

[electricpete]

I was wondering: How sure are you that surges were essentially absent? Do you have some kind of scans on the voltages?

Sure enough for my purposes. To get from the grid to the motor requires to travel through 345kv/13.8kv transformer, 13.8kv to 4.16kv transformer and 4.16kv to 480v transformer. The only overhead wiring is on the 345kv system (all else is bus duct or cable). We have lightning arresters in place on the high side of the 345kv/13.8kv transformer. The motor is not fed from vfd. No plant evolutions were in progress at the time the ground alarm occurred and no other alarms/anomalies were observed at the time of the ground alarm.

Of the two scenarios I have to choose #1. In my opinion both are weak and probably not directly related to the failure. I'm not concerned with voids in the slot however if strands are loose that could be a contributing factor. Have to checked/monitored the THD?

Why do you say these scenarios are weak?
I'm not sure I see the distinction between the empty space we see from end of the slot and potential looseness of the strands in the slot.... they seem roughly the same thing to my thinking. Perhaps good resin can overcome low slot fill to some extent.
Have not checked/monitored THD in voltage or current on this particular equipment. There are no power electronics feeding the motor nor directly fed from the same load center transformer. Have not seen any significant voltage distortion during the times we have had an opportunity to monitor voltage on similar 480 volt systems in our plant. Have seen some significant distortion on current for some motors running at low load. Voltage does run on the high side from time to time but never more than 10% above nameplate.

I appreciate the interest, but I'll repeat I did not ask the forum to help figure out the cause. I asked to discuss relative merits of two scenarios.

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