Motor running above max temp for decades? 6

[Skogsgurra]

Now this:

A compressor motor was installed in 1972. A 1600 kW 3 kV asynchronous motor. It is fed from its own transformer, 20/3 kV 2 MVA and 7 percent uk.

The motors have always been equipped with Pt100 temperature transducers but these have been connected to a not-so-reliable temperature indicator and no-one really cared what they said. The temperatures seemed to be in the 90 - 95 centigrades region most of the time.

Now, the indicator system has been replaced with a new one and everything has been checked out. The temperature is now rising to 140 C within one hour and still increasing. The isolation class is B (yes, old machine) so the people closed the machine down and yelled for help.

Checked cooling, heat exchanger, motor voltage, symmetry, distortion, slip, current consumption - I think everything. All OK.

The conclusion is that the motor probably has been running quite hot before the change of temperature indicators - but no-one knew. My question: Is it at all possible to run a class B motor for years at temperatures close to or slightly above 150 C?

(Yes, I know about the halfing for each 8 - 10 C increase in temperature. It is not that part of the problem I am trying to grasp. The question is more like this: are there any "magic limits" that makes a vintage class B insulation break down very quickly or do we still follow the "2exp(-T/10)" law?

In other words: Is it probable that the motor has been running 20 - 30 C above rated temperature for many years?

 

[electricpete]

We don't have any really old motors but it seems to be a common war story told by old-timer electicians that where they used to work they abused their motors and never maintained them and never had any problems.

Early 1970's vintage would possibly be asphaltic based insulation. If heated I believe it will become puffy, so inspection may tell.

Temperature rating for class B hotspot (which should be RTD) is in fact 130C (140C for 1.15 SF).

So when you were at 140C that was the SF temperature. How much higher we don't know. Clearly continuous loading at this temperature is beyond the design or normal practice. But sometimes equipment does tolerate abuse pretty well.

Bottom line after all that rambling - who knows! It's certainly possible.

One question - did they double check the rtd measurment setup - proper for 3-wire RTD if applicable. If you don't compensate for lead length temperature can read much hotter than actual.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Skogsgurra]

Thanks Pete!

Yes, we also measured the Pt100s directly at the junction box at the motor and got similar results.

Appreciate your views on this. I tried to argue along your lines - but needed more support. A few more opinions and I will go to the next meeting better armed.

 

[Laplacian]

Are you 100% positive they are platinum RTD's; is there a slight posibility they could be 100ohm nickle? I have run into this with 1960's to 1970's vintage motors made in the US.

 

[Skogsgurra]

That's a new thought. What would it imply? Faster ageing? We measured cold motor. R was then 108 ohms, and motor temp was 20 - 25 C. So the transducers seemed to be OK at that temperature. But there is surely room for more precision in that check.

I haven't been able to find the tempco for Ni100. Will have a look in CRC handbook.

 

[Skogsgurra]

Found it elsewhere. The resistance for a Ni100 at 100 C is 161 ohms. That corresponds to more than 150 C if a Pt100 were used. So there is a definitive risk that the new indicators are wrong. And the old ones were OK all the time. I'll check with the manufacturer - if he still exists.

 

[buzzp]

Sounds like you found the problem. I just thought I would add that I had a case where a motor megged bad (less than 1k) and the dang thing kept on running. As far as I know, it is still running. Those old motors are the toughest you will find.
However, Pete makes a good point about the ashphalt insulation. That stuff tends to soften up significantly when heated. Seems at that temp it would be oozing out the casing.

 

[aolalde]

Great tip by Laplacian!! . I think he deserves at least a star.

 

[jraef]

I gave him one. Even if it turns out not to be the problem, he deserves it just because he was savy enough to question what we all took as a given!

"Venditori de oleum-vipera non vigere excordis populi"

 

[Skogsgurra]

Agree. Laplacian gave a completely new angle of view. We have finally got in touch with the manufacturer. Waiting for his answer.

 

[UKpete]

Hi skogsgurra, just a general point but it is common for manufacturers to put in insulation that is better than the specified temperature class (particularly that nearest the copper). For example on standard industrial LV motors in the UK they are class B but actually manufactured with class F insulation, but I don't know if the same can be said for MV motors.

If you are doubtful about the accuracy of the temperature monitoring, is it practical to do a resistance shut-down test?

Also 1972 doesn't sound that old to me, I don't think materials and practices have changed much in that time.

 

[vanstoja]

Motor insulation class temperature ratings generally are based on a 40,000 hour lifetime which amounts to about 4.57 years at 100% utilization. Conservative, reliable motor manufacturers may elect to use better insulation materials than required by the class standard and will run Arrenius time-temperature tests to confirm compliance with the class lifetime at rated temperature. This may result in an actual insulation system that meets next higher class ratings. Operating voltage may also result in some additional conservatism in insulation system design since voltage surge withstand requirements, when specified, usually dictate greater ground/turn insulation thicknesses for higher voltages such as your 3 KV motors. If insulation systems operate in an inert(say nitrogen-filled) environment, such as is usually provided in canned induction motors cooled by process fluids, then several multiples (up to 6X) of insulation class lifetime at rated temperature may be obtainable. One really needs the test-determined Arrenhius plots of the actual motor insulation system and a good description of the winding environmental conditions to make a reliable estimate of the real winding insulation life at class rated temperature.

 

[Skogsgurra]

Yes, UK. It is possible to take readings on a cold motor. The resistance was measured, but I was not there. The guys said that they measured about 108 ohms and the temperature was probably between 20 and 25 C. So it clearly indicates Pt100. But I do not know how reliable these data are. Not even who did the measurement.

Vanstoja, your data support the thought that the motors have been running quite hot for many years and that the problem was detected only when new temperature indicators were installed. The local repair shop has cleaned and "painted" the motors several times during the last twenty years. "Painting" means dipping and hardening. It is possible that important cooling channels have been narrowed over the years.

No answers from the manufacturer so far.

 

[UKpete]

skogs, what I meant was checking the motor winding temperature by the resistance cooling curve method, only practical if the motor can be stopped and quickly isolated from the transformer.

 

[Skogsgurra]

No, UK. That has not been done. Shall see to it ASAP.

 

[Skogsgurra]

Results from manufacturer: They are Ni100!

Result from winding resistance test: The windings were not room temperature but lower. The motors have forced ventilation and air/water heat exchanger. The cooling was active also at standstill. That's why we measured 108 ohms at room temperature. It should have been around 112 ohms if the windings were at room temperature.

All in all: You never know what your next mistake will be. Thanks to all - and especially Laplacian.

 

[aolalde]

skogsgurra:

Thanks for sharing your findings.

 

[edison123]

Good posts skogs and laplacian.

 

[Laplacian]

I'm glad you found the problem and I appreciate all the kind comments and stars.

A while back we added a Multilin 469 to a 2500HP motor starter and decided to use the RTD's to monitor temperature. I found some poor quality original documentation on the motor and saw that the RTD's were 100 ohm. I "assumed" platinum since I couldn't make out the type from the documentation.

About a week into service, operators were screaming that the machine was running hot and about to trip up operations. We monitored that air temp coming out of the enclosure while motor was running and it did not correlate with reported RTD temperature.

I went back into the Multilin and was able to easily change the RTD type to 100 ohm nickel. That was the first time I had seen 100 ohm nickel RTD's used in a motor.

I'm glad I was able to share my learning on this forum.

 

[Skogsgurra]

me too!