Does anyone have experience with premature winding failure on appropriately derated LV motors installed above 3000M on Inverter PWM? I am looking for data that would help to explain some premature motor winding failures (multiple manufacturers) on VFD (multiple manufacturers) driven motors in high elevations. Some motors are VPI, all are inverter rated. Voltage varies from 460/480V - 575 and 690V. PWM varies from 1.5 - 3kHz. The failures are not seen at elevations below 3000M and the actual analysis of the failed motors are unreliable. Motors are appropriately derated for elevation in all cases.

Are the failures on end turns or in the slot?

I will assume most if not all Motor/ASD combinations have reactance in the output and it looks to be a plant wide phenomenon based on the description. De-rating already having been addressed tells me that these failed units are not failed due to thermal degradation. Since ASD connected motors are the only ones mentioned I will assume motors started with ATL and RVSS are not having high rates of failures. I would start at the plants transformer and meter the mains voltage, check plant ground. The odd thing I am seeing said is that motors under the altitude de-rate do not have these problems. Any ASD tripping codes that have been noted?

Chris, as it seems as though the problem is universal across different manufacturers, I would guess the problem lies somewhere in your installation. This may due to the usual suspects of Voltage, Harmonics etc. As you have pointed out, the problems do not appear below 3000m. Now this is a guess only as have not had experience in high altitude installations and someone more knowlegable could confirm this, but when derating for these altitudes, do you need to apply a correction factor. In aviation, performance of a aircraft is 'normalised' by applying a temp factor over altitude. This accounts for the air density, less air, less lift and maybe less cooling in your case?. For example at 2000 feet and a temp of 28 Celcius, gives the aircraft a performance equivalent to say 3500' or so. When manufacturers state a derating factor, do we need to take this into account. As I said, only a guess! Good Luck

"Appropriately derated" - can you define what the derate was? Dean and Bill have both mentioned corona and end-turn winding failures. My only experience with extreme altitude (4500m+) VFD applications required voltage derating to prevent corona at reduced atmospheric pressure - essentially, 575V rated VFDs operated at 380V, and motors wound with insulation suitable for 480V inverter duty @ sea level. You say you are operating drives by multiple manufacturers at 460, 575 and 690V. I have to assume the motors connected to those drives match the VFD voltage. If this is the case, corona is suspected as the cause of most of the failures at least. On the VPI'd motors, it's still a possibility if the insulation class is just not up to the duty.

Chris,

If this were one location with continued failures then the various ideas of incoming voltage, harmonics, etc etc etc suggested might have value to help you. But that is not what you described.

Your problem & solution is probably 99.5% as described by Bill in his reply above.

If you do not have inductors in series with your motors, add them. If you do, make them higher L.

Bill is 100% correct that books have been written about this corona discharge eating motor insulation from the voltage spikes. If by LV you mean 230v, those spikes start out at 6-700v and go upto 1200-1600 quickly with long leads. Double these spike voltages for 480v systems. Study your various motors and buy from those who have the least failures or try other mfgrs.

Although VPI is the best way to get rid of the small air spaces inside the windings, where the ozone generated eats the varnish, it is not a guarantee that ALL air spaces are removed. So one mfgr may be less capable of removing all air than another even if both use vacuum impregnation.....

I would advise before supplying a solution and assumption that the problem is voltage punch through of the winding insulation that you get a few pictures of the motor windings in failed units. It will be fairly clear if the problem is one of overvoltage spikes to the windings or not having enough derate in the machines for the altitude and low air density.

I agree with Bill, and had an experience with a single drive powering 6 LV motors simultaneously. Motors #3 and 5 failed regularly, and were replaced three times about every 10 days. The drive and motor were manufactured by the same company. As this was the beginning of higher efficiency motors, it was originally thought that this was the culprit, and standard Eff. units were installed. Same thing happened.

I was in the MV group, but was in the general area on other business when called in to physically look at the installation. A hand-held meter showed the variance in voltage at each motor's terminal box. It took a great deal of pushing to get my employer, the distributor and the site to bring in an oscilloscope and even more to encourage them to measure the wave forms every couple of feet down the 150 foot line.

Amazing how plotting voltage vs. length validated university antenna theory and clearly illustrated standing waveforms. Especially at the locations #3 and #5 which were over 3x voltage.

This was before the use of drive filter packs at the motor terminals (reactors or a combination of tuned devices). The solution here was to replace the original cable (now with tiny holes in the insulation) with new cable and add seemingly useless loops of wire in between the motors. They were not all of the same extra length, but were specifically designed to move the motor's electrical distance to a zero-effect point on the voltage-length plot.
And the problems went away.

what is the distance beetween motors and VFD?
are been used shelded cable?

The length of cable is VFD specific and depends on the impedance of the cable, the motor and the VFD. The average of 7 different manufactures in the LV region is:
Screened cable: 25m-150m
Unscreened cable: 50m-300m

The pictures sent back from the locations with the failed motors show blachen patches across serveral end turn windings. The balance if the winding looks like new. These are 460 volt motors operating from an AFE. So instead of the normal 650VDC buss, the buss is 720 VDC. All other motors that we have operating at high alitude have had VPI treatment. Only the two locations were the motors do not have VPI treatment have failed.

The cable length from drive to motor is between 10 and 15 feet. There is no output reactor. The voltage rise time of this inverter is rather slow at 275usec to 300usec.
My suspicoin is that the motor insulation needs to have a voltage derating to prevent isulation break down due to partial discharge but with only two cases of this failure and lack of documentation I have little to back this up.

The evidence is the failure in the end turns which is indicative of HV breakdown of the insulation. 1st step I would take is to place 3% Reactor on the load side between motor and VSD.

Both motors were rewound by a motor shop in Lima, Peru. The only report we could get from the motor shop was to say the "insulation was not sufficient for the allitude". A load reactor has been added in the one case that occured last year but the older one that occurred about two years ago still does not have load reactor. Since being rewound there has not been a failure. I am sure that the motor shop did a VPI treatment.
No failures at high alltitude have occured with motors with VPI treatment.