Bearing load increase due to misaligned flexible coupling 1

[geesamand]

I work with flex couplings (Lovejoy L- series or Falk Wrapflex) mounted between a gear reducer and C-face 1800rpm electric motor. The motor is mounted to the reducer using a piloted pedestal without adjustment. In my experience the couplings are rated for much higher misalignment levels than is acceptable for machine vibration limits and bearing loads. So we want to estimate or experiment to understand the practical limits for misalignment at this coupling without damaging the bearings or creating excess vibration. I have some ideas for how to build an adjustable fixture, and measuring vibration levels is easy. But determining the effect on the bearings is tougher.

Is there a way to measure or calculate the bearing load increase caused by a misaligned flexible coupling?

David

 

[BrianE22]

The coupling data sheet should include stiffnesses for offsets and angles.

 

[electricpete]

Some people have attempted to measure:

1A -

http://www.mt-online.com/component/...nment-reduces-bearing-life-.html?directory=90

1B -

http://www.mhm.assetweb.com/DRKNOW/APLPAPR.NSF/apweb/A46D0489F7CC8D41852567ED005AB6C4?OpenDocument

2 -

http://www.ulb.ac.be/polytech/laborulb/couplings/couplings.pdf

I have built a spreadsheet for attempting to calculate. It uses a model where shaft sections are modeled as beams, bearings are modeled as linear springs, and coupling is modeled as two angular springs. You can vary the parameters of models to see the results. Attached is the spreadsheet.

I can only attach one file per message, so will continue in next post...

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

 

[electricpete]

Attached powerpoint is the results of using the previously-attached spreadsheet to try to model a wide range of shaft to shaft misalignment conditions. There are four variables: Shaft, Coupling, Bearing, Misalignment. Each of these variables is varied in two configurations: Shaft is thin or thick. Coupling is hard (stiff) or soft (flexible). Bearing is hard (stiff) or soft (flexible). Misalignment is offset or angle. This creates 16 possibilities which correspond to 16 slides in attachment.

In each slide we have one between-bearings shaft on left, one between-bearings shaft on right, and the shaft extensions and coupling between the inboard bearings of those machines (4 bearings total). In case it is difficult to see how x-coordinate of the graph relates to the machine: the four bearings are at the locations of the 4 dark-purple-colored spikes (dark purple is "force" and the spikes are the reaction forces at the bearings). You can also see that shear is integral of these discrete forces, moment is integral of shear, slope is integral of moment (with 1/EI factor), and displacement is integral of slope.

You can see from the results that the reaction force at the bearings depends on all those variables (stiffness of bearings, stifness of shaft, stiffness of coupling, nature of misalignment). Also will depend on the distances between bearings etc.

The geometric inputs should be easy to get (representative shaft sizes and distance between bearings). The stiffnesses will be the tricky part. As mentioned above coupling stiffnesses might be available from manufacturers. (I have seen some links discussing these values somewhere). As far as "bearing stiffness", it would actually represent series combination of bearing stiffness and support stiffness. Knowledge of this variable is also a prerequisite for rotordynamic analysis, and there is plenty of literature on how to estimate it. You could look for representative values in textooks. You could do a bump test with instrumented hammer and look at low-frequency asymptote to estimate stiffness. You could also push with a pressure-instrumented jack and measure movement with dial indicator. I doubt anyone here will be that energetic...



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

 

[electricpete]

Whoops. The attached powerpoint should be used instead of the previous powerpoint. The first 16 slides correspond to the 16 cases discussed. After that some general discussion.

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

 

[geesamand]

Thank you for all of the detailed information. I will contact our coupling suppliers for stiffness data to perform an analysis. It is not in their literature.

By chance has anyone extended this type of data to a practical limit for parallel and angular misalignment suitable for elastomeric jaw couplings? The couplings permit large misalignments but in my experience these limits cause the machine to run like a washing machine with a brick in it. Machine size 1-100hp, 1800rpm max.

 

[electricpete]

Machine alignment tolerances are generally defined by machine speed (and possibly size and criticality) without reference to coupling type (other than spacer couplings where limits depend on spacing).

There are plenty of recommendations out there. Here is one extracted from an EASA document

http://www.bradleysmotors.com/pdf/Technical Data/Alignment Tolerances.pdf

For 1800rpm machine, it gives a limit of 1 mil offset, 0.3 mils/inch angle.

When you get down toward the 1hp end of your range, you may have a hard time getting dial indicator bracket onto your machine so it may not be practical to achieve these tolerances...

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

 

[electricpete]

Other alignment recommendations from google

http://www.peerlesspump.com/shared_docs/Update Your Shaft Alignment Knowledge.pdf

(last page)

http://www.impactengineering.com/pdf/Alignment_Tolerances.pdf

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

 

[diskullman]

You may want to look at the Lovejoy Deltaflex coupling

http://www.lovejoy-inc.com/products/specialty-products/deltaflex.aspx

Russell Giuliano

 

[diskullman]

If this is a c face motor mating with the gearbox, how much misalignment can there be? Should be neglible. Can you get a gearbox that accepts C face motor without thepedestal and coupling? What kind of gearbox are you using?

Russell Giuliano

 

[geesamand]

We manufacture these gear reducers and adapters. The couplings mentioned are chosen for low cost, performance, and availability. The pedestals are cast/machined and were sized to fit Lovejoy L- and C- series. There are many other coupling options but they don't fit in the available space and cost.

The misalignment is never negligible in production. Due to the form of the pedestal, the pilots must be machined in separate operations, so getting perfect alignment isn't automatic. We are working to improve the quality of the pedestal machining, but we also need to develop a rational understanding of the tolerances.

David

 

[geesamand]

I wanted to share a link that would have been extremely useful and concise for my needs, but they didn't test any elastomeric jaw couplings.

http://www.mt-online.com/component/...nment-reduces-bearing-life-.html?directory=90

In any case, this has given me some ideas for how to build a test fixture that could provide the direct data to support my study.

 

[electricpete]

Hmmm, that's familiar. It is in fact the same address of item 1A of my post dated 15 Jun 11 23:05

If you keep reading to item 2 of same post, then look at pages 29 and 30 you will see some results for one type of elastomeric jaw coupling (looks like 6 jaws on each half, unlike typical small 3-jaw)

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

 

[geesamand]

I'm way too young to be having senior moments like that.

So far my major coupling manufacturers have provided zero useful or relevant information. I am a little surprised, given that jaw couplings are so immensely popular.

 

[Tmoose]

I respectfully submit that after you establish loads vs misalignment:
1 - Resulting bearing life should be based on considerations of your actual bearings' fatigue or endurance limit and effectiveness of lubrication and sealing, not just the L10 paintbrush
2 - Seal "life" should consider deflections compared to seal manufacturer's specs for allowable eccentricity, and a tolerance analysis of the concentricity of the seal bore and bearing bores.

no harm, no foul.

Dan T

 

[geesamand]

Thanks Dan, that is essentially the intent here.

With point 1, I am limited to publishing L10 values as our customers strictly understand only that. However in our design process we also consider L10a calculations. That's another can of worms. For this exercise L10 will show the effects of increased load just as well as an L10a.

David