smoke puff out of 13.2kv motor, but no trip 1

[electricpete]

This is a strange one.

Outdoor WP2 Vertical 13.2kv 3500 hp, 324rpm S.C.I.M driving a circulating water pump. Upper bearing in Kingsbury style in 50 gallon oil bath with cooling coil immersed in the bath.

History:
2005 - Motor rewound. Core noted to be marginal but was not restacked. Cracking performed to bring core hotspots in spec.

2009 – B phase Motor T-lead open-circuited resulting in motor trip. Bizarre failure described here.
thread237-241996

February 2010 – Partial discharge increases from approx 350 millivolts to approx 550 millivolts on B phase

3/2/2010 – motor removed for routine preventive maintenance – Megger (sat), bridge (sat), filter change (tortuous path between filters and winding by desing). Oil level was normal (high end of band).

3/11/2010 – Motor restarted. Motor was not watched carefully upon restart.
· 20 minutes after start, there was a report of blue haze coming from the bottom of the motor (this motor takes in air at the top and discharges at the bottom). Motor was manually secured 5 minutes later (did not trip).

Repeat megger test sat – 1,000 megaohms (corrected), pi > 2.
Repeat winding resistance test sat – all phases match very closely.
Inspected space heaters, inspected lower bearing area, inspected packing area. Inspected rotor, air baffle and limited view of stator through air box ports. No anomalies noted.

3/13/2010 – Restart the motor.
No smoke noted (initially). Although there was a smell that seemed to get worse.
Vibration sat.
Partial discharge still around 550 millivolts on B phase.
Thermography showed no anomalies at pump packing, motor lower bearing housing, coupling. About 12F hotter than the other motors on the motor frame.
· After about 40 minutes of running a fairly large puff of white smoke came out of the bottom of the motor lasting about 3-5 seconds and then stopped. Motor was manually secured 3-5 minutes later.

Electrical test has not yet repeated but I anticipate it is still sat (if not, then the course of action is obvious).

Discussion of temperature trend: This motor ran about 20F hotter than others before rewind in 2005, has continued 20F hotter than others. Was cleaned in April 2009 and remained hotter. Has actually increased to about 30F hotter than others over the past year. But hottest temperature last summer was 230F (hottest of 6... we check them all). And the insulation is Class F.

What say you?
1 - Practical question: Is there any logical further troubleshooting before we pull the motor and send it to the shop for disassembly / core test etc?

2 - Just for fun question. What is your prediction of likely/possible causes of puffs of smoke from a motor that does not trip and looks good on electrical testing? How likely are the following scenarios to cause this:
- Core overheating?
- Water leak into upper bearing oil reservoir causing overflow of standpipe (we haven't seen any evidence of oil but do plan to draw a bottom sample)
- Foreign material in the motor?
- Others?

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

 

[edison123]

Did you do a surge test ?

What does sat mean ?

Muthu

http://www.edison.co.in

 

[oldfieldguy]

What's the smoke smell like?

In the automotive world, the old rule was blue smoke = oil, white smoke = water, and black or gray smoke = burning carboniferous material.

I've seen strange-colored puffs from foreign materials dislodged from cooling passages, too.

Insulation-related burning usually has a distinctive odor. If the motor is stopped, a careful wipe with the fingertips may collect enough residue for s "sniff test".

old field guy

 

[amptramp]

pete:

Forgive my ignorance, but what procedure is indicated by "Cracking performed to bring core hotspots in spec."

Thanks, amp.

 

[electricpete]

Thanks for those replies.

Muthu

No, we have not done any surge test. That is beyond our capability to do on-site at the plant. I was thinking the possibility of a turn-to-turn fault in the end-turns which could burn for awhile before causing a fault. But this motor ran for 40 minutes between 1st puff of smoke and (after shutting down and restarting) 2nd puff of smoke. Do you think it is credible? At any rate if turn fault is the case, then it requires motor removal so doesn't suggest any mandatory test on-site (question 1), but interesting idea for question 2... will keep surge test in mind for when/if the motor gets to the shop.

"sat" means "satisfactory". I think it has origins in the US military and pretty common term in US nuke which is populated by a lot of military. Anyway, sorry for using obscure terminology.

oldfieldguy - It was described as an electrical smell by some that were there. I wasn't there to smell it (was watching a motor test 1000 miles away... but got to participate in problem solving by phone... will try your wipe/smell test on Monday). I know I associate a certain sharp smell with low voltage windings. Personally I don't ever remember smelling any distinctive smell like that when investigating a medium voltage vpi winding failure... does anyone else? We had "blue haze" the first time and "white smoke" the 2nd time....maybe oil and water!

AmpTramp - Regarding cracking - see 1st column of 2nd page here:

http://www.lexseco.com/files/Lexseco_WhitePaper.pdf

Cracking and spreading and a few other techniques are used to try to improve core hot-spots without restacking the core. Cracking is mechanical agitation to get relative movement between lamination plates. Exactly why it helps I'm not sure (breaks burrs off? moves around insulating/oxidation layers betwen lams?... not sure).

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

 

[waross]

White smoke may also indicate vaporized light oil. a diesel engine that is cranking but not firing will put the unburned fuel out the exhaust as white smoke.

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

 

[waross]

Is there a noise associated with the smoke. I remember a motor with a line connection flashing to ground. Loud noise and black smoke. It would then heal itself for about a minute and blow again. We were some distance away when we first heard the noise. We wondered what it was until we spotted smoke rising in the area of the motor.
A wild thought. Could you have lube oil dripping on an overheated space heater? Very hot may produce a blue haze. Lower temp (more oil leaking?) may produce white smoke.

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

 

[edison123]

pete - You don't to need to dismantle the motor to do a surge test. You can do it in situ. Just turn the rotor by hand to rule out the air-gap variations showing up as faults in the waveforms.

Muthu

http://www.edison.co.in

 

[amptramp]

pete:

Thanks for the info.

 

[electricpete]

Thanks Muthu. I think you misunderstood my response (maybe it was misleading). I never said it is required to dismantle a motor to do a surge test (although as a side note, that does give a more sensitive test since influence of rotor can be removed, wouldn't you agree?). What I said was we cannot do the surge test ourselves on-site. The practical question (question 1) is what do we need to do on-site before going through the expense and time to ship the motor to the shop. It is not logical in our situation to waste the time/money to bring someone on-site to do a surge test... if it passes we found nothing, if it fails we still have to send the motor off-site. So we are focusing on question 1 .... relevant checks we should do now (before pulling the trigger to ship the motor off-site) to avoid realizing later (when the motor is in the shop) that it was something we could have easily fixed on-site. Of course perhaps it requires you to know what we are capable of on-site, so perhaps not a fair question if you are not a mind reader.

Thanks Bill. We did inspect the space heaters and saw nothing unusual there. It is remotely possible there may be oil dripping on the stator someplace we cannot see, I guess a boroscope is a possibility there.

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

 

[electricpete]

By the way, if I thought turn-to-turn failure was one a really likely scenario, it would make sense to bring someone in to do a surge test. That would help us make the correct decision quickly (since it is desired to get the motor back in service as soon as possible). But my subjective take is that it is not a likely scenario, so more likely than not doing a surge test just delays us making the right decision. So I did ask your opinion... how likely?

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

 

[waross]

Hi Pete. You are the guy on site and your gut may be more accurate than our guesses.
My suggestion, If there is noise, look for electrical problems. No noise look for oil on hot surfaces. But I could be wrong and you are the guy on site.

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

 

[Younsi]

Hi Pete,
I am new on this Eng-Tips as you know. While doing accelerated aging test on two motors last October we saw the same smoke event followed by normal operation on one of the motors. My engineer said it "burped" and gave white smoke. After shut down and routine insulation testing IR, C, DF, etc. The motor behaved normally and we put it back under testing. After 10 days of accelerated life testing with severe 0 to 250% load cycling the motor failed. It did have the same "burping" event a second time a few days before failure.
After dissection we did discover that the motor had several turn to turn failures that seemed to have self cleared ! This could be the case for your motor but honestly I never heard or read about a scenario like what we saw before.
Your motor is a 15kV rating one. I suggest the following:
-Analyze the PD data better and go beyond looking at magnitudes. If you send phase resolved plots and patterns I can look at them and comment for you. If you have a IRIS monitor get us the 2D and 3D plots.
-Turn to turn faults are detectable by both vibrations and MCSA. Did you higher vibs ?
-MCSA requires CT sensors and a data logger at high sampling freq. If this motor is critical I recommend you consider monitoring the stator currents and their harmonics.
-Do you see unusual behavior on your RTDs ? Are they all functional ?

Other indicators of trouble are ozone, load dependance of PD levels, positive PD predominance etc.

I hope this can help you or at least create more comments from good maintenance folks.

Best regards

Karim

 

[edison123]

In my opinion, any smoke from the electrical machines is more due to electrical reasons than mechanical ones. You answered your own question

"relevant checks we should do now (before pulling the trigger to ship the motor off-site) to avoid realizing later (when the motor is in the shop) that it was something we could have easily fixed on-site".

A negative (surge result) is a positive.

So, go for it.


Hi Karim. Welcome to ET.

What exactly do you mean by "doing accelerated aging test on two motors" ? Accelerated aging test is a prototype test for a few sample stator coils/bars done outside the machine to assess the insulation system used and those tested coils/bars are never used in the motor windings themselves.


Muthu

http://www.edison.co.in

 

[electricpete]

Thanks Bill and Muthu – will keep that in mind.

Karim

I intend to check partial discharge phase pattern when I get a chance. I get emailed overall results but don’t have access to the detailed data without extra effort to go to the acquisition laptop where the data/software lives.... haven’t been near that since the smoke occurred until today.

Interesting case to have turn to turn fault which caused smoke but did not immediately fail. Was that a random wound machine (ours is of course form wound)? The traditional wisdom is that an autotransformer effect is created which creates much higher current in the shorted-turn loop. I guess maybe perhaps the shorted turn could burn open the short between turns without opening the circuit (?). Did you ever formulate an explanation?

-Turn to turn faults are detectable by both vibrations and MCSA. Did you higher vibs ?

We did collect vibration – what would you look for.. increase in 2*LF?


ALL – We have made decision to pull motor – the motor and me will be in the shop Wednesday for testing/inspection.

I have some more info pulled together in the attachment.

Slide 1 attached is vibration overalls... you can see that the 3/13 reading after seeing blue haze the first time, the overalls are mostly unchanged.

Slides 2-6 are “waterfall” spectral plots of vibration vs frequency (horizontal axis) and time (diagonal axis) where the most recent spectrum is on top (pink). The horizontal axis is frequency in “orders” (multiple of running speed), starting above 1x to the smaller stuff is more visible. You can see many of the points had a distinct increase in 9x and 10x during the most recent run and on slide 3 1H2 position all the harmonics 2x-10x increased during most recent test. (Note this data was taken by a very experienced individual so we are confident this is not a result of poor magnet attachment).

Slides 7-13 show the similar waterfalls over a wider frequecy range. You can see slide 9 a possible increase in broadband noise floor.

The vibration charts might tell us harmonics suggest looseness. I tend to there is a possibility that a rub may have occurred and caused the smoke, although that’s just a swag (why would it come and go).

Slides 14 and 15 give an overview of winding temperautre performance with discussion on the top of slide 14.


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

 

[Younsi]

Pete
Sorry I did not get a chance to comment sooner. Based on the data I agree there is a clear vibration signature. I doubt it is bearingrelated otherwise it would stay. I will get someone to help with identifying what it can be. Did your expert mention the potential for "intermittent turn fault" ?
The RTD data shows seasonal changes which is OK but also shows transients and quite a noisy signal in general. Is this normal to you if you compare it to your other motors ?

Another possibility for the smoke. If you have a broken bar or end ring, which very often becomes intermnittent, that could also fit your descrition of Sunday.
Any PD data yet ?


Muthu,
I agree with you accelerated aging tests are done in general on coils or bars for lower cost. In my case I am developing a new technology that the business feels should be validated with data of full size motors even if they cost more ! That is what I can says at this stage. You might see it as a new product soon.

Regards

karim

 

[edison123]

Thanks Karim. Must feel good to have a rich client.

Muthu

http://www.edison.co.in

 

[electricpete]

Karim - Thanks for your comments. There is no doubt some mechanical problem existing based on change in vibration (and update below). I have heard of in intermittent rotor sparking in case of loose rotor bar, and I guess it is not too much of a stretch to extend that to rotor bars. We last did current signature analysis in Fall 2008, and bars appeared very healthy (pole pass sidebands something like 55 db below 60hz magnitude under load).

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

 

[electricpete]

UPDATE - It was just reported the motor is difficult to rotate while removing it for sending to the shop. (Note we couldn't check this until machine was uncoupled because the associated pump has Cutlass bearings which are notoriously difficult to rotate).

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

 

[electricpete]

UPDATE

1 – the above report of inability to rotate freely by hand was ERRONEOUS. An 18" strap wrench was used. Typically a larger wrench is required to break free the Kingsbury thrust bearing. When motor was received in the shop it rotated with expected amount of forc.e

2 – As found airgap readings CCW From Term box:
Bottom: 85 40 85 85
Top: 85 48 94 94
Since a part had been dropped into the machine, the rotor was not rotated to check the airgap. Note the 180 degree opposite positions don't sum to equal numbers as expected suggesting mild ovality in rotor or stator.

3 – Rotor was removed and inspection showed evidence of past rotor to stator rub. Multiple contact points on stator concentrated in a very small area where the airgap was smallest. Multiple contact points around the rotor circumference. Rotor TIR fairly typical - <= 1 mil on shaft. 10 mils on top and bottom of rotor iron. 15 mil at middle of rotor iron. The high spot of the 15 mil middle reading was 180 opposite of the worst of the rotor contact points. Stator bore was also mapped using a high-tech Coordinate Mapping machine and showed typical amounts of deviation from concentricity between airgaps and flange/rabbet surfaces. There was no signs of unusual movement at upper or lower bearings.

In April 2009 when machine was last disassembled/reassembled for T-lead repair, there was no evidence of rub at that time (based on my old photos). At that time, the lower bearing was not replaced, upper bearing radial pads not adjusted, and airgap was not checked. There was perhaps a suspicion that rabbet fit of endbrackets would put the airgap back where it started although in retrospect it certainly should have been checked. My theory at this time is that there was enough slop in the rabbet fits that when we reassembled the machine it ended up in a different position as a result of this slop. Alhtough it is a little coincidental that the offset from centered is similar on top and bottom. The part I struggle to explain is why would this rub come and go with the machine at steady state (based on appearance of smoke after 40 minutes of running).


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