PM meggering of motors 2

[iceworm]

Equipment is 480V TEFC induction motors, NEMA, SF 1.15, operated 24/7, however, generally possible to take out of service for PM.
I'll split into two different size ranges:
Less than 50Hp;
200Hp to 50Hp

Clean industrial environment
Not over-greased or under-greased.
Question concerns recommended frequency for meggering during PM.

So far I have looked at 2016 NFPA 70B, 2015 NETA-MTS, IEEE 112 -2004, IEEE 43-2013, 2014 NEMA MG-1.

And yes, I also looked at the MFG recommendations:
Reliance (Baldor)
GE
US Motors
Marathon

I'm not asking for an opinion on a frequency. Rather does anyone have any suggested references?

ice

Harmless flakes working together can unleash an avalanche of destruction

 

[dArsonval]

Recommended Tests/Inspections and Associated Performance found at 5.2 of this document may provide what you're seeking.

http://www.epri.com/abstracts/Pages/ProductAbstract.aspx?ProductId=NP-7502

You may have to refine your question. It reads to me as if you're asking how many times is an apparatus meggered
during a preventive maintenance service procedure. Which would be similar to asking how many times does one operate
a wrench, or a screwdriver. Perhaps clarifying your question will provide more assistance from the astute crowd here.

John

 

[iceworm]

dA -
Thanks for the reference.

You may have to refine your question. It reads to me as if you're asking how many times is an apparatus meggered
during a preventive maintenance service procedure.

Thank you for the suggestion. That would have never occurred to me.

Re-wording question:

Looking for references on, "How often one should perform a preventative maintenance megger of a motor?"
Please read the OP for the context.

ice

Harmless flakes working together can unleash an avalanche of destruction

 

[controlsdude]

I have seen this PM done from 1 - 5 year interval. Also, since its done so infrequent, its usually outsourced to a more expertise type electrical house that does this type of work frequently.

 

[iceworm]

cd -
Thanks for the response.
Yes, that is the frequency I have seen (am seeing) as well. And, yes, the customer has the expertise and test equipment to both do the testing and interpret the results.

I'll give some context:
The question revolves around, "Why are we doing this? We never see any motor going bad. We just find them after they have gone bad - and that is rare. And 90% of the time it is the bearings that went out, not the winding insulation - unless the bearings went out, rotor hit stator, and that took out the windings."

Okay - reasonable request. If it were safety related, no brainer, we're doing it and reviewing.

But if it is not safety related - what does it benefit? Even if they were to catch one in the process of insulation failure, was it caught in time where cleaning, baking out, tying, glyptol, bearings made it good for another ten years? Or is in in for a rewind anyway?

It costs money and time to do the testing. There are no limits on stuff to do. There is always a limit on money and time. That's why I'm looking for papers/research on frequency, and in particular, what were the parameters used in selecting the frequency.

Opinions, while interesting, aren't too much help - unless they are followed with, "And here is why we selected that particular frequency." That would be helpful.

The link from dA was interesting (Thank you dA). It did discuss documented failure rates, and that was helpful. However when it got to recommended frequencies, it stuck to safety related items. Bummer, we already know the answer for that.

Definitely interested if you know of any references.

ice

Harmless flakes working together can unleash an avalanche of destruction

 

[jraef]

I think it also depends on how organized and prevention oriented your facility is. One place I worked at did annual PM inspections of almost all motors, including megger testing. One thing I liked about it was that by looking at records on a motor, you could see the slow decline of insulation values when relatively new, and the sharper change rate as they aged. Basically once the insulation resistance began to drop more than a percent or two in a single year, the rate of decline steadily accelerated toward failure. High moisture areas where the motors cycled on and off a lot fared the worst, motors that ran 24/7/365 (364) tended to last the longest. Bottom line though, our down time from motor failures was nearly nil, because we had a good idea WHEN a motor would fail so we could do a preemptive replacement when something else was being worked on.

Most other places I worked where I was involved in motor work were not that organized or disciplined, so consequently down time from unexpected motor failures was more of a problem.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[waross]

In a long and varied career I can't remember ever seeing motors meggered for the sake of meggering.
I like the procedure that jraef shared with us. If the records are accurately maintained and available when needed it is a good idea.
I have been known to record megger readings inside the terminal box. After hurricane Mitch, I found that the megger readings of the generators that I had marked in the back of the switchboard a few years before made an excellent benchmark as to when the generators were dry enough to put back in service. Hurricane Mitch hit our little diesel plant three times in three days but that's a story for another day. When you open a terminal box and see inside a white salty high water mark at the level of the opening into the windings you know you're in trouble.


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

 

[electricpete]

EPRI NP7502 = Electric Motor Predictive and Preventive Maintenance Guide (linked above) - Table 5.1 (random wound < 200hp) - Insulation Resistance: 24-36 months.

EPRI TR-106857-V8 (1997) - Low Voltage Motor PM Basis - Table 3.2 - Electrical Off-line testing: 3-4 Years.

EPRI "PM Database" (living document) - "low voltage motors' - "motor performance testing" - 3 years. "Motor Performance Testing is performed to ensure that the electrical characteristics of the motor, including performance, are within specifications, and to diagnose the source of many potential problems.". It is open to interpretation what test this particular author had in mind. However since the same database also recommends bridge testing at the same interval (3 years), I interpret that they mean insulation resistance testing.

EPRI 1003095 (Electric Motor Tiered Maintenance Approach) outlines an approach for you to develop your own intereval based on equipment criticality, spares, environment, history, etc.

IEEE Std 1415-2006 - IEEE GUIDE TO INDUCTION MACHINERY MAINTENANCE TESTING AND FAILURE ANALYSIS - Table 1 - "insulation resistance" - "typical frequency" 1 to 2 years. (This may not be the latest version… there has been some activity on this standard recently).

NETA MTS-1 2001 (there are later versions but this is the one I have access to) - lists insulation resistance as a maintenance test although not interval is mentioned that I can find.

NFPA 70B - Recommended Practice for Electrical Equipment Maintenance (1998) - (not the latest, but the latest available to me) - lists I.R. as maintenance test although no interval mentioned that I could find (same as MTS1).

One question to face is whether to determinate the motor from the cable for the test. Most people (including me) do not generally require to determinate there unless there is indication of a problem. Determinating the motor from cable creates more work and also creates potential to introduce problems during determination. Unfortunately this approach often means that the insulation resistance you are seeing is more influenced by the cable than the motor (especially long cable runs). That makes trending not as effective.

There are some other challenges for trending. We already talked about cable (if left terminated during test). Also insulation resistance is highly sensitive to temperature of the test (which affects the insulation itself) as well as surface conditions on the leads and endwindings related to humidty. There are temperature corrections which can be applied (IEEE 43 suggests to develop your own for individual equipment based on history but realistically no-one is going to do that… just stick with the standard ones). Since the corrections are not perfect it's preferable to test near the same temperatrue every time and/or near the reference temperature (104F). Some people like to test soon after equipment is taken out of service to minimize effects of humidity (warm winding won't have humidity effects) although implementing that can be a coordination challenge and space heaters make it less critical. I do some trending of insulation resistance under varying ideal and non-ideal test conditions and I certainly don't look for a few percent change between test. To get my attention would require a consistent trend of decrease over several tests or else something like a factor of 2 change between tests unless the value is low enough to concern me in absence of trend.

I can appreciate some people view it as a low-value-added test. On the other hand if you have a failure of a critical motor to answer for, and you have never bothered to test it since installation some decades ago, it can in some circumstances create some uncomfortable questions.

Some plants have ungrounded 480 volt distribution systems with ground indication/alarms but no ground trip. Motors can and do develop grounds without tripping. One might argue that insulation resistance testing is less critical under that scenario. (no I'm not asking for a critique of ungrounded electrical systems).


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

 

[waross]

Nice research Pete. lps
Re: Ungrounded systems. My feeling is that given the possible issues with an arcing ground fault on an ungrounded system, regular insulation testing may be more important with ungrounded systems.
Not contradicting you, just responding to your argument with a different opinion.
At the end of it all, whatever procedure that the owner wishes to implement is the correct procedure for that plant.

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

 

[djs]

Back when I was in the motor business, I found some information on predicting motor insulation failure. The short of it was that a simple megger test would not reliably predict insulation failure. However a related test called dialect absorption test would. Different and more expensive equipment similar test setup. As always good record keeping and regular test intervals are important. For small motor probably not a cost effective procedure. However with process lines with high associated high costs of un-expected downtime, it might pay off.

 

[electricpete]

Thanks Bill I appreciate your comments and opinion but we definitely view it differently.

I had a motor a month ago that gave a ground alarm but continued to run. Took it out of service, pulled off the TEFC shroud/fan and pulled back the outboard endbell and saw clear evidence of turn to turn short in one coil (photo attached). But interestingly it had continued to run just fine until we manually took it out of service (although I don't necessarily expect to be the norm on turn to turn short... which is why I think it's worthy of mentioning).

Had this been a grounded system this same fault (or any fault that results in very lower ground resistance) would have gone undetected or more likely tripped. Neither outcome inspires me to think the grounded system should be tested less. I've never seen any evidence of this mentioned arcing ground fault in my 16 years working with somewhere more than a thousand 460 volt motors at our plant.

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

 

[iceworm]

electricpete -
Thanks for your response.
EPRI TR-106857-V8 (1997) Couldn't find this one on the EPRI site
EPRI PM Database Couldn't find this one
EPRI 1003095 Found this one interesting

IEEE Std 1415-2006 Yes, had this one.

I'm off to work for several weeks. So, I won't be checking in for a while. I'm not ignoring anyone - just out of touch.

Appreciate everyone's effort.

ice

Harmless flakes working together can unleash an avalanche of destruction

 

[electricpete]

EPRI TR-106857-V8 (1997) Couldn't find this one on the EPRI site

I have a copy I downloaded from the website in the past but I didn't realize it is no longer available. The entire PM Basis series 106857 was issued in 1997 as a snapshot, frozen in time. It makes sense that it has been pulled because it has been replaced by the next one below - " EPRI PM Database" - which is supposed to be a "living" continually updated database (although the particular item you're interested in for low voltage motors hasn't been updated since 2001).

EPRI PM Database Couldn't find this one

You need to be logged into the website as an EPRI member for this one.
From homepage select program cockpits. Then select NMAC.
Then look at lower right hand corner for: Links / Web Applications / EPRI Preventive Maintenance Basis Database (PMBD) Version 3.1
Once you're into the web app, enter "motor" in the search box and then select "low voltage motor" from the search results.

edit - by the way the database includes not only intervals but a variety of calculational tools. One of them is a what-if analysis which attempts to provide estimates of reliability impact of the various pm strategies. I think maybe if you spent time figuring out how to use it (the terminology and approach is convoluted), you could get it to spit out a number representing a prediction of impact of insulation resistance testing (or lack thereof) upon reliability. I don't think most people (myself included) would really trust any number like that without knowing something about the underlying model, assumptions and data used to generate the prediction, but maybe it will be of some help to you in your discussions.

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

 

[waross]

Hi Pete. I seldom disagree with you but but even when I disagree I always respect your knowledge and experience.
May I share my point of view? Thanks I will. (grin)
I have been at it for a lot longer than 16 years but in that time I have worked in only a few ungrounded plants.
I haven't personally seen an arcing ground fault.
However several of my instructors spent years working in ungrounded plants before going into teaching.
Several of my instructors had seen arcing ground fault issues more than once and shared their experience with their classes.
If a motor or transformer develops an arcing ground fault the induction of the winding combined with the capacitance to ground of the system may form a spark gap transmitter. If the fault is near the wye point or neutral point, the winding may form an auto-transformer and result in a high frequency, high voltage being superimposed on the system. This stresses the insulation of all running motors and can lead to widespread motor failure throughout the plant.
The faulted motor may continue to run and may not result in a trip or ground fault indication.
I hope you never see this first hand.

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

 

[electricpete]

I hope you never see this first hand.

I hope for the best in all cases, but I can tell you this doesn't keep me up at night. I don't want to get into specifics about my industry, but I can share some info about my plant. We aggressively investigate ground alarms. We have redundant loads powered from independent (not paralleled) load center transformers, so that any failure on one ungrounded system (be it transient overvoltage or otherwise) does not affect the redundant motors performing the same function.

We have resistive ground-indicating lights (all three should be same intensity) and ground relays connected from each phase to ground. Certainly that will bleed off any dc with an RC time constant. I have a strong suspicion that it does help with the transient overvoltage scenarios although I can't say for certain. I feel somewhat at a disadvantage to analyse this since the transient overvoltage scenarios don't seem to be accompanied by well documented case studies (which is not to say they didn't happen but only to say electric system description and event details are not well communicated) and have different explanations depending on who is doing the explaining (maybe different explanations apply to different events).


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