Fretting Wear/Fatigue in Splined Connections 1

[chez311]

Does anyone know a method for determining a design safety factor for fretting wear and fretting fatigue for a involute splined connection? I am thinking there must be some sort of required safety factor of tooth contact stresses to material yield. Also would these requirements be different for clearance/line to line fits vs. press fit connections?

 

[tbuelna]

All metal-metal spline joints experience some amount of fretting. It's just the nature of the device. Fretting in a spline tooth contact is a very complex process, and it is not easy to analyze with accuracy.

There are several threads where this subject has been discussed, with many excellent technical references provided. A good place to start would be a search of posts using the term "spline fretting".

Good luck to you.

 

[MikeHalloran]

If the spline has to slide, I suppose some fretting is inevitable.

If the spline is intended only for torque transmission without relative motion, defenses are available, e.g. rolling the external spline with a very slight twist, and mating it to a straight broached internal spline, such that a hammer or a press or a drawbar are required to accomplish axial assembly.



Mike Halloran
Pembroke Pines, FL, USA

 

[tbuelna]

If you don't require a connection that permits some axial motion, then there are better methods for coupling a shaft than using a spline. The only way to prevent fretting is to ensure there is absolutely zero possibility of relative motion at the coupling interface. Shrink fits, curvics, or clamped friction from bolt flanges are all better choices than splines when fretting is a concern.

 

[chez311]

Thanks for the replies:

1) tbuelna - I read through quite a few of the posts, as well as technical articles and specifications, however I was re-reading through the posts when I noticed an attachment that you mentioned in thread406-309426. It references an AGMA spec with a section 10.4.4 having a calculation for fretting/wear rating - as far as I can tell this would be ANSI/AGMA 6123-B06, correct? Its the first reference to such a calculation I've seen, I think this could be very useful in determining if under operating loads (not peak/spike loads) fretting is a concern for a particular interface.

2) This is in a joint which does not require axial motion, its only purpose is to transmit torque. In regards to the elimination of relative motion, there seems to be some discrepancy between sources. However, from what I've read it seems that the mechanism of fretting wear is extremely small displacements that can occur through either gross slip (ie: sliding on the micro-scale) or stick/stick-slip conditions. This means that even in an application which has no allowable axial or radial movement by design (zero clearance press fit) there is still relative movement due to elastic deformation at the contact surface which leads to fretting. I guess it would be possible to design a joint with enough preload that any additional elastic deformation under design loading would be negligible, however that would require very high press forces and is not an option for this particular application due to assembly requirements.

One thing, however, I have noticed is not discussed in the sources I have read - whereas it is stated by many experts that fretting can be hindered (or almost completely eliminated) by a constant flow of lubricant across the contact surface to carry away particulates generated from fretting (which is not an option in many instances, and thinking about it - would this not require some clearance between mating parts?) wouldn't some inherent clearance/lash between components (which under no load and/or torque reversals allow the mating components to come out of contact) accomplish something relatively similar - maybe not as effective, but still similar?

 

[mfgenggear]

Chez311

see the link to this article.
interesting

http://www.machinerylubrication.com/Read/693/fretting-wear

Mfgenggear

 

[tbuelna]

chez311-

You did not mention what application this spline is used for, but attached are some pages from helicopter design handbooks on oil lubricated spline design. Sheets 1-5 are from the US Army helicopter design handbook and sheet 6 is from the Sikorsky helicopter transmission design handbook. Sheet 7 is a crude sketch showing how I would typically design an oil lubricated spline connection on a rotating shaft for an aircraft drivetrain application. The spline teeth operate submerged in oil and there is a constant flow of oil across the tooth faces to flush out any loose debris that is generated by contact sliding. I think this is similar to what you describe in your last post. This approach is standard in the aircraft industry and has been shown to be very effective at minimizing fretting wear of splines, but it does not completely eliminate the problem.

If you look at table 7-4 on sheet 5 you'll see some recommended limits for contact stress for various types of splines/materials. The handbook this data came from is probably 40 years old, but it is still commonly used in the aircraft industry for sizing splines. You asked specifically about a design factor of safety for fretting wear in spline teeth, but I don't recall ever seeing any. The aircraft industry is extremely conservative when it comes to something like the design of a spline connection in a flight critical system, especially when the spline presents a single point of failure that can disable the entire system. So a typical fillet root side fit spline, with an L/D below 1.0, using the oil lubrication approach shown in my sketch, made from a thru hardened alloy steel, might be sized for a simple P/A contact stress of around 5ksi. This would be considered acceptable for unlimited service life, but obviously it is also extremely conservative.

Good luck to you.
Terry

 

[gearcutter]

If anyone is looking for better quality print of the excellent info that tbuelna has posted; here's the link to the pages in JPEG -

https://flic.kr/s/aHsk8YcywF

On the bottom RH side of the images; click on the down-pointing arrow for download options.

 

[tbuelna]

gearcutter-

Thanks for providing better quality images of that spline reference.

Just one last comment regarding the point chez311 brought up about fretting wear in oil lubricated splines. It is hard to overstate just how beneficial providing a controlled flow of oil across the spline joint, like that shown in my sketch, can be in terms of fretting wear. A similar fretting wear situation exists with sprag type clutches used to allow a helicopter main rotor transmission to overrun the engine. These sprag clutches are designed so that the inner (output) race overruns when required, but they spend most of their operating life locked up. Torque is transferred across the clutch by friction at the contacts between the smooth, hard metal sprag and race surfaces. The hertzian contact stress levels at the sprag and race surfaces can be extremely high, and every effort is made when designing these clutches to prevent any relative movement at the contacts when the clutch is locked since that will produce fretting wear. But even with good designs some small amount of fretting wear always occurs and this generates metallic debris. To ensure this metallic debris is flushed from the clutch, the same oil control approach shown in my sketch is used. Where "dams" are located at each end of the clutch and the clutch elements operate submerged in oil, with a continuous flow of oil thru the clutch.

An example of this approach is shown in the attachment below. It is from Sikorsky's Helicopter Freewheel Unit Design Guide USAAMRDL-TR-77-18. It's 240 pages and is an incredible technical reference. If you have an interest in the subject I highly recommend giving it a read.