Alignment Targets for Large TEFC Motors 6

[JJPellin]

I posted a question back in August about motor shaft axial shuttling in some newer pump trains with relatively large TEFC motors. Update: We have largely resolved our shuttling problem. Our motor supplier and motor repair shop establish mag center on these large TEFC motors by running them solo with the cooling fan removed. If we have to establish axial running position in the field, we place the rotor in the center of mechanical float. We have less axial shuttling problems, but have another more serious problem.

On at least four occasions, we have had coupling disk pack failures on these machines. The failures are characteristic of severe misalignment. For the rest of my comments on this post, I will use one machine train as an example. Energy control policy does not allow us to perform a true hot alignment. We align these machines cold using alignment targets based on predicted thermal growth. For a large pump train, the pump OEM predicted that the pump would grow 0.016” vertically and the 800 HP TEFC motor would grow 0.006” vertically. Based on this, we set the motor 0.010” high for our cold alignment. Over the last six months, we have failed disks in the coupling twice. We recognized a need to validate the alignment targets.

We placed dial indicators on the coupling hubs with the pump in hot condition and monitored as it cooled off. The pump hub dropped 0.003” and the motor dropped 0.018”. Next we performed a very detailed temperature survey of all pump and motor supports. This showed that the motor supports averaged 95 F on the drive end and 60 F on the non-drive end. The cooling fan arrangement on these TEFC motors is not cooling the motor housing uniformly. Similarly, the pump supports were averaging 95 F on the NDE and 70 F on the DE. The air from the motor fan is blowing hard against the inboard supports. Using these temperatures, we would predict that the motor needs to be set about 0.010” low for a cold alignment. Recall, the OEM had us setting the motor high by 0.010”. We were running with 0.020” parallel offset misalignment. No surprise we were having coupling problems.

We will adjust our alignment targets and realign the motors next week. I am still curious if anyone else has particular experience with uneven vertical growth on large TEFC motors.


Johnny Pellin

 

[edison123]

I have seen a few 1700 KW ID fan motors having a vertical thermal growth and such motors were cold aligned with the motor low. One theory was since the motor bottom is bolted, the only way for it to grow was upwards.

Muthu

http://www.edison.co.in

 

[JJPellin]

I should have mentioned that the pump I was referring to runs at about 650 F. In theory, the pump should grow much more than the motor, even though it is centerline mounted (API pump). The fact that we need to install the motor low of the pump is very unusual and is based on the uneven growth of the motor and pump. I was taught that motors grow straight up with approximately equal growth on the inboard and outboard ends. And, that seems to be true for WPII motors. But, for TEFC, it does not appear to be true.

Johnny Pellin

 

[Tmoose]

Do you have some pictures of the failed coupling disks?

What is the distance between the disk packs?

I'd rig up something to allow viewing the coupling in operation, or take high speed pictures.

The intent is to view the coupling in operation to reveal if the disks are puckering or bending?

Do you loosen all the disk pack clamp bolts in the aligned state, and then tighten them, to allow the disks to start life fairly "neutral?

3 down 17 to go.

 

[JJPellin]

The disks were broken at the outermost disks at the edge of the washers. The coupling manufacturer's manual indicates this is suggestive of misalignment. The coupling spacer is built as an assembly with a hub-to-hub gap of 8 inches. The distance between disk packs is probably about 6.5 inches. The center members with disk packs assembled are built up on the bench. I have not heard of any benefit from loosening bolts in the field.

Johnny Pellin

 

[MikeHalloran]

Next time of course, you want much longer coupling spacers.
... and someone's selection criteria need adjusting.

I assume that moving motor or pump farther apart and using a longer coupling assembly is now a huge deal.

I wonder if the coupling people might offer disc packs with a greater number of thinner discs for improved fatigue life in your application. Are they working on that problem, or are you left on your own?


The problem is a little extra messy because of the motor's asymmetrical expansion.
If it were a little closer to symmetrical, i.e., just a changing parallel offset, I'd look at Schmidt couplings.






Mike Halloran
Pembroke Pines, FL, USA

 

[JJPellin]

The coupling spacing meets API and our internal design criteria. The coupling is a standard design with a service factor greater than 1.5. We have approximately 1500 centrifugal pumps that use this same coupling design and have coupling spacing shorter than this. Failures of these couplings are extremely rare. Moving the pump or the motor is impractical at this point.

Johnny Pellin

 

[electricpete]

Thanks for the feedback on the axial shuttling.

We don't have any TEFC motors above 200hp. Where we do have challenging thermal growth situations (turbine driven pumps with turbine and pump on separate/independent platforms), we attempt to measure off-line-to-running change in alignment using Permalign. I see you attempted to determine something similar with dial indicators but your efforts were limited because safety procedures prevented you from performing a true/full alignment check hot… instead just monitored change in vertical position at each shaft. With the Permalign, we install it with equipment tagged out (typically cold) and then get the useful data during the next transition to hot conditions.

fwiw I could easily imagine that TEFC motor grows more on the inboard end (DE) than the outboard end (NDE) based on the external cooling pattern where the outboard endbell is heavily cooled, inboard endbell has no cooling airflow, and the cooling air applied to the frame fins warms as it flows from outboard toward inboard. Infrared thermal image while loaded might give a reasonable basis for developing a thermal correction estimate. But again we've never measured thermal growth on a TEFC.

I have not heard of any benefit from bolt loosening...

I agree with Tmoose on this (I think I learned it from him on the other board). For Thomas Shim pack couplings, if the machine is aligned to a non-zero target cold then you should if possible loosen/retighten coupling bolts hot to relieve stresses you induced into the coupling through the growth. It is recommended in the Thomas / Rexnord literature. Attached poweropint is my attempt to illustrate the reasoning.


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

 

[CouplingGuru]

JJPellin,

You are correct. Base on the numbers you mentioned of 8" distance between shaft ends that should allow you a 0.046" parallel offset with out any problems. But that is assuming perfect angular alignment and zero axial movement. Based on your failure description, the disc coupling manufacture is correct, you have a strain problem. The manufacturer has pointed to mis-alignment but that generally is a summation of many alignments. Angular-Parallel-Axial alignment. Now all those alignments resolve back to angular deflection in the disc pack. I would be willing to bet that you actually have some twisting going on that is actually providing some angular misalignment as well as parallel misalignment. Since you have an API coupling, I would take a measurement across the flex element from adapter to center member flange at 90 degree intervals around the coupling. This measurement will give you the exact angular misalignment that the packs are seeing. It would be best if you did this measurement in the hot condition and measure again when the pump cools down. The manufacturer should have these values defined in their manual, if not I can walk you through how to calculate them base on the coupling manufacturers misalignment rating.

When it comes to couplings we are always here to help.

http://WWW.PSCCOUPLINGS.COM

 

[electricpete]

Typically these shim pack couplings give warning before failure. If you are inspecting them, you can see the cracks early. If you are not inspecting them, then eventually (at a later stage), you will see pieces of individual shim packs ejected somewhere below coupling guard (but machine continues to run fine at this point). We have seen both the above, but never experienced a complete failure of the coupling that actually stopped the machine. I'm curious what you saw on your four failures.

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

 

[JJPellin]

Thank you for several good points and suggestions.

I am surprised at the comments about loosening shim pack (disk pack) bolts to relax stresses. I think this may be a result of me providing inadequate information. I referring to Thomas Series 71 couplings with removable spacers. We rebuild these spacers in the shop and tighten the bolts with no stress and no imposed misalignment and then install them as a complete assembly. There are other types of couplings (Series 42, Series 52, AMR) where you assemble the disk packs and associated bolting in the field. In these, you have no choice but to tighten the bolts with the coupling under some amount of misalignment. For those types, it makes sense to consider loosening and tightening bolting in a hot condition once the misalignment is at a minimum. Our energy control practices would make this impractical in most situations, but I can still see the value. If I needed to address this with a field-assembled coupling, I would probably do it differently. I would ask the mechanics to align the machine to zero-zero, assemble the coupling, and then make the moves needed to achieve the cold-offset targets. This should have the coupling disk packs in a relaxed condition with a nearly perfect alignment.

I agree that I should embrace the technology and get the laser tools to be able to measure alignment changes on a running machine. We have tried this with our laser systems and have hired outside contractors to come in and do this for us. In all cases, I was disappointed with the results. I think the lasers are very capable to take these measurements. But the bracketry needed to mount them is tricky. Unless the bracketry is perfect, it can distort the results. I need to revisit this technology and get current on its capability.

The coupling failures we have had were not complete failures. We have caught all of them when the outer 1 or 2 disks cracked. But, we still have to go in and rebuild the coupling. We have viewing windows in our coupling guards and inspect the couplings with a strobe if we see evidence in the vibration that we have coupling issues.

I am generally skeptical of coupling manufacturers’ values for allowable misalignment. They have a strong incentive to rate their product capabilities higher. I would never willingly run a coupling this large with an offset more than about 0.015”, no matter what the coupling manufacturer claimed. Even if the coupling could take this for a period of time, I would be concerned about the loads imposed on the driver and driven machine and the resulting vibration.

We have a thermal imaging camera in our group. I should use that to get a better picture (literally) of the temperature distributions on the running equipment. I don’t know why I didn’t think of that. Thanks for the suggestion.


Johnny Pellin

 

[electricpete]

I'm not familiar with that 71 type. I agree with you that my comments about loosening probably don't apply. They wouldn't apply if the capscrews through the hub thread directly into the adapter (nomenclature per figure 2 here ) rather than a loose fit in a non-threaded hole with nut on the other side where the bolt can end up held off-center within the hole.

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

 

[CouplingGuru]

JJPellin,

I agree with almost everything you have stated. I am very familiar with the Thomas product, although you have seen first hand the detriments of running with misalignment, I have seen first hand what it takes for a coupling of that nature to fail. While I don't contend you should run it at rated misalignment I do know that somehow you are seeing excess misalignment in order to get that failure. In terms of coupling design really the only way to achieve better misalignment rating is to have a higher tensile material as the disc pack(i.e. more springy) or a longer chord length (4 bolt vs 6 bolt vs 8 bolt) These high performance materials are generally way too costly to incorporate since most OEMs are drawn to the lowest cost product. Anyway, getting back to your problem, if you see cracking of the discs next to the washer it is 100% a strain induced failure, this is only cause by angular misalignment across the flex element. So I am fairly positive you have some additional misalignment sources aside from parallel misalignment. Or you may have compensated so greatly in the cold condition, that at start up you are running above rated misalignment. Either way, best way to verify is to measure the distance from the adapter flange to the center spool flange at 90 degree intervals around the coupling. This will give you the minimum and maximum spacing that the flex element is seeing. Then you can use the outer diameter of the coupling in the following formula INV-SIN[((Max-Min)/2)/(Spool Dia/2)]= Angluar element flex. This angular element flex has to be less that 0.25 degrees, if that is the case the disc will last a very very very long time.

Also Technically for API compliance those center-member assemblies are suppose to be factory tightened and balanced, maybe you are doing all that but switching disc packs on a factory balanced and match marked center member assembly could create some serious vibration issues, I am sure you are on top of that, but just giving you a heads up on that.

When it comes to couplings we are always here to help.

http://WWW.PSCCOUPLINGS.COM

 

[electricpete]

So I am fairly positive you have some additional misalignment sources aside from parallel misalignment.

I'm not sure why you say that. Parallel misalignment of the shafts translate to an equal/opposite angular misalignment at each of the two flex planes as you know. So couldn't pure parallel misalignment of shafts cause this? (I'm not saying it was pure parallel misalignment in this particular case, I'm just checking the logic).

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

 

[CouplingGuru]

electricpete,

You are correct, pure parallel misalignment can cause this, but base on the numbers JJPellin described he is well with in the prescribed parallel misalignment rating. However, how do you know the shafts stay perfectly parallel to one another through-out the thermal growth range? Plus I know for a fact that it impossible to have "perfectly" parallel shafts, just like it is virtually impossible to achieve "perfect" alignment. That is why I always recommend measuring the flex element gap, it is pure angular misalignment, it takes into account face run-outs, bolt location deviation, angular shaft misalignment, parallel shaft misalignment and even axial misalignment if the nominal flex gap value is known (or measured prior to install). Either way the disc pack failure is from over flexing the element. Now we just have to figure out where it is coming from.

When it comes to couplings we are always here to help.

http://WWW.PSCCOUPLINGS.COM

 

[JJPellin]

I think what electricpete is trying to say is that any angular or offset misalignment will result in pure angular misalignment across an individual disk pack. In that regard, you are correct, if we could directly measure the angular misalignment across each disk pack, we would know for certain if it was within OEM specifications or not. But, it would not matter if we achieved that amount of angular misalignment across that disk pack by having a purely offset, purely angular or combination misalignment between the driver and driven shaft. You mentioned a good point about the chord length affecting the allowable misalignment. This coupling is a Series 71-8 which is a smaller OD with 8 bolts through the disk packs. This has a shorter chord length than the old Series-71 original design. According to the OEM manual for this coupling, the maximum misalignment is 0.015" TIR angular and 0.012” TIR offset. I have not found out how to interpret these numbers. If these are truly TIR readings from indicator sweeps, then we are only allowed 0.006” parallel offset and an angular misalignment of 0.11 degrees. Based on our readings, we had almost zero angularity, but 0.020” to 0.030” parallel offset. Am I interpreting this incorrectly? Our coupling is a size 375.

Johnny Pellin

 

[CouplingGuru]

Using the right side of the chart is a combination measurement, it defines the mid-point of the angular and parallel misalignment relationship.

The better way is by getting that "N" dimension. Measure that, that automatically combines both of them into angular deflection of the pack.
So your flex element gap has to be between .70" and .68", again that it highly conservative, because their rating of 1/3 degree per pack would allow.
SIN[.333]*[7.62/2]+Nominal gap & SIN[.333]*[7.62/2]-Nominal gap (Nominal gap = (.70+.68)/2 = .69)
Or a gap tolerance +\- 0.022" so with their published 1/3 degree rating that "N" gap needs to fall with in 0.712" to 0.668"
The "N" is the measurable distance between the adapter and center member flange.

Also you mentioned chord length on the 71 vs 71-8, that is why the angular misalignment rating of the 71 (6 bolt) is 1/2 degree and and 71-8 (8 bolt) is 1/3 degree.

When it comes to couplings we are always here to help.

http://WWW.PSCCOUPLINGS.COM

 

[MikeHalloran]

Did you all mean to say "chord length"?


Mike Halloran
Pembroke Pines, FL, USA

 

[CouplingGuru]

Yes on the chord, auto correct cell phone

When it comes to couplings we are always here to help.

http://WWW.PSCCOUPLINGS.COM

 

[JJPellin]

As requested, I have attached a picture from the last failure. Certainly not a dramatic or extreme failure. But, a very serious concern, none the less.

Johnny Pellin