We have had a rash of gear coupling failures lately. Mostly due to lubrication issues. I was told that infrared and ultrasonics would be the best method of checking couplings on the run.

My question is how much warning would infrared give you before the coupling failed? Would UE be an earlier warning system and use IR to confirm or give a better indicator of time before failure?

Aubrey

It would seem that a baseline temperature (thermography image) would have to be established. Ambient would have to be taken at the same time. Temperature difference could lead to a reasonable conclusion. One concern would be how hot would you have to be to decide that a lubrication issue is there? Over lubrication could also give hot temperatures. The trending might help but trial and error may take place to feel confortable with the decision to check the lube?
I have found vibration is a good indicator that a coupling may be locking up as your normal vibration will increase axially although the amplitude might not be very great.
Not to sure about ultrasonics. If the coupling locks up you may not have any movement to make noise.

We have had gear coupling issues before and found a good PM program really helped.

Just some thoughts!

We tried using thermography to monitor couplings. The attached report shows a gear coupling that was found damaged in 1/2004 during the normal coupling PM. The coupling had been last greased in 6/1999. The last thermogram taken before the failure was found was in 7/2003. I would not have predicted a failure based on the report.

ACCW_B_coup.pdf (374 Kb, 99 downloads)

I prefer ultrasound for evaluating coupling condition. Temperature can be hard to measure with coupling guard in place and then shaft heat conduction from the coupled machine rotors has to be accounted for.
Temperature, vibrations, and visual inspection with strobelight are complementary technologies that can be usefull.
Walt

quote:

Originally posted by Dean Riley:
I would not have predicted a failure based on the report.

Dean,

I would not have predicted either. But I had once a successful prediction ( no grease) even without trending. May be this was an easy case since what caught my eye was uneven temperature distribution (although max delta T did not exceed 13F).

I need to ask you couple of specific questions on IR technigues used in your example. I can see Emissivity is set to 0.96. Is it a valid setting in this case? For couplings made out of shiny stainless steel ( if this is your case)I doubt it is correct. Since temperature differences as indication of a problem could be subtle it therefore should be taken with high precision and could have indicated a problem either in trend or in the IR image pattern.

Also, what are those "cold" spots in the front of the coupling? Some kind of reflection?

David

Coupling_IR.doc (303 Kb, 56 downloads)

I would say the conventional rule for rolling element bearings is that temperature primarily responds to lubrication problems, and doesn't respond to mechanical degradation (pits in the race) until the failure is very advanced. It might respond a little to loading factors such as misalignment that tend to accelerate bearing wear.

My gut tells me the same logic should apply for gear couplings. Tooth wear itself isn't going to cause much change in temperature until it is advanced. Misalignment that leads to wear might cause temperature increase. Lubrication condition will be the biggest driver of coupling temperature (perhaps expect a spike in temperature after lube a coupling just like after lube a bearing?)

For the bearing, we have a very powerful and effective alternative to detect mechanical defects/degradation within the bearing... vibration. I'm not sure vibration is as powerful for monitoring coupling degradation. I have seen a gear coupling progress to the point that the machine train completely failed... I.E. the motor stopped driving the pump. We didn't see that one coming by vibration (and we don't monitor coupling temperature)

All good comments! Because we are in the predictive maintenance business we do keep an open mind to the possibilities. The Infraspection Institute several years ago with support from another company (Eli Lilly maybe?) did a paper on misalignment and infrared thermography. Depending on the severity of misalignment the infrared thermography did see significant temperature difference. I can't prove it but it would stand to reason that minor lubrication differences would not bring much change but no lube or way over lube would bring some differences if you have zeroed in on what is normal. I don't think that it is full proof by any means.

This message has been edited. Last edited by: Don Jones,

Suppose you are in the office and on the table is an infrared camera, vibration analyzer, strobelight, and an ultrasound meter. The boss walks in a asks you to evaluate couplings on 100 machines because of a rash of recent failures. Which one would you choose?

I'd be off and running (now its a fast walk) with the ultrasound meter!

Walt

Walt - It seems to makes sense that ultrasound might be a good tool for this. Have you ever been able to find gear tooth wear or other gear coupling conditions using ultrasound?

I had very good success at monitoring lubrication problems and resulting gear tooth wear and cracking in Zurn Amerigear couplings in a powerplant feedwater pump application in the early 1990's. A copy of my published paper is available at

IR monitoring of lubricated couplings

Note that the framing rate of the camera being used (AGEMA 870) happened to closely correspond to the rotational frequency of the shaft, allowing the shaft to effectively be "thermally strobed", and thus being able to see the individual hotspots around the circumference that corresponded to damaged teeth.

Rich Wurzbach
Maintenance Reliability Group, LLC

Sorry, just realized the paper isn't up there yet. I'll get it so that link is working. Here is a link to the abstract, and where a copy can be purchased direct from STLE, but give me a day or two and I'll post my copy.

SPIE Thermosense Abstract

Rich Wurzbach
Maintenance Reliability Group, LLC

David_G

I happened across the report the same day I read Aubrey's original post and thought it would be useful since I knew that the coupling had failed and there were yearly thermograms on it.

You are right about the emissivity settings. I went back to the report and changed the emissivity settings to .85 (the setting in the 1st thermogram). The temperatures went up about 5 to 7 degrees. I think the cool spots are the reflection of the concrete wall directly behind the thermographer. There is a basin of water behind the concrete that the pump draws its suction from.

Even with the adjusted temperatures, I don't think I would have predicted the failure.

ACCW_B_coup1.pdf (389 Kb, 41 downloads)

Dean,

The powers that be have given me the orders to report every coupling that measures above 130 degrees. I questioned this as one end of the paper machine is about 95 degrees and the other end of the dryer section is about 125 degrees. I was told the coupling rep. said 130 degrees should be the cut off point and to go with it.

Just wait a week when everything on the hot end of the machine is 135 degrees. You can bet I will be asked about all those hot couplings.

Aubrey

Aubrey,
I'm with you. I don't think that a rule like that is realistic. Multiple technologies may be necessary for best results. Whatever will work the best for the situation.

Don

El'Pete,

I have found damaged/worn teeth in gear couplings. There are several lubricated types of couplings that ultrasound can work well: gear (single and with spool piece), metal grid, quill shaft, U-joint, and constant velocity joint. Poor maintenance practice often comes in two or more fault situations for couplings:
* Misalignment -- force generator
* Unbalance -- force generator
* Dry/contaminated/deteriorated lubricant -- friction instigator
* Worn/damaged teeth or internal components -- looseness/impacting instigator

Find one problem and one or more of the other problems may be present!

Walt

Aubrey
The reason we started trending the temperatures was to extend the time between PMs. We were regreasing the coupling every year. We extended the interval to 3 years. This particular PM was missed for some reason and went 4 1/2 years. If we would have followed our 3 year plan, this failure would not have occured. We also changed from purging the coupling to complete disassembly, cleaning, inspection, and hand regreasing. We use Mobil EP111 exclusively on all greased couplings. I've been here since '87 and can think of only 3 lube related failures, all caused by overdue PMs.

I'm sure ultrasonic monitoring will help you out. I wish I had the time and resources to learn more about it.

Good luck.

Hey Dean,

That is about what happened to us. We have an annual outage. A contractor is supposed to disassemble the coupling, inspect it, repack the grease, then reassemble it. Somewhere along the line, greasing and proper assembly got overlooked.

We discussed the situation and with some advice of one of our forum members we decided to go with a two prong attack. We will do ultrasound inspections backed up with a infrared. Both of the methods have a learning curve. On the UE route I have already detected a coupling that could be dry. We will pull it apart on the next field day and inspect it.

Aubrey

You may want to consider a "three-prong" approach. I believe that there is significant value in performing strobe inspections of operating couplings, if you can get line-of-sight on them. This posting elsewhere in reliabilityweb.com shows a few failed couplings and a strobed image of the coupling inservice before failure: Coupling forum discussion pictures

While there is much more to see in flexible element couplings with the strobe than with gear couplings, you can have some success at inspecting the fasteners, and strobing at 2x and slightly offsetting the speed to see some wobble or misalignment.

We may actually be at a disadvantage with the more modern infrared cameras (FPAs) than the older raster scanners, because they seemed to have a better chance at stopping the coupling motion and showing hotspots as they rotated. I think perhaps some of the current IR cameras, depending on their framing rates, might blur the image and make that difficult.

Regarding having an absolute temperature go-nogo for the couplings, that doesn't make a lot of sense. If you have heat being contributed from the adjacent machine that is propagating down the shaft to the coupling on one side, that is not any indication of the condition of that coupling. Rather, we need to look for temperature rises above normal for indications of problems. But I do understand that you might have a hard time convincing supervision of that if the coupling manufacturer provided a temperature guideline. Unfortunately, manufacturers don't always understand just what we are measuring, and in some cases have never seen IR images of their equipment before. I had that case recently with a synchronous belt installation that had a hotspot along the shoulder of the drive sprocket. When we showed that to the manufacturer's engineering department, they said they had never seen such an image, they would have to look for internal funding to study it, and it would take about 18 months to complete the study.

Rich Wurzbach
ASNT Level III Certificate #117658

Rich is correct about using a strobe for visual inspection, if you have one. I once looked at a pump coupling (while operating and no strobe) and found a loose shaft key. The plant person wanted to know how I knew. I said the lighting (60 Hz flicker) was acting as a strobe, and I could "see" its position change.

Walt

Don,

You have never mentioned what was the failure mode in your coupling.

Rich, Walt,

It is OK to use a strobe, but for detecting lack of lubrication in a gear/grid coupling this method obviously won't help. I agree, heat generated by the coupled equipment and propogating through the shaft is masking a problem in the coupling, but assuming that this is a stable process and trending ambient temperature as well may help.

One factor impeding IR method may be insignificant heat amount generated by the faulty coupling on this background, so that signal is lost in the noise.

David