Fatigue failure of BS 970 : 817M40T 3

[mog69]

Does anyone have any data on allowable stress for 817M40T ?
I have an application which requires a 6" Circular Pitch spur pinion to be overhung from it's last supporting bearing to engage with a large Ring Gear (approx 28 ft pitch diameter). The pinion and it's shaft are machined from a single shape specific forging 1732 mm long. The pinion has a face width of 500 mm. The shaft is supported vertically (pinion at the bottom) on two spherical roller bearings nominally 400 mm bore and on 600 mm centres. I have experienced failure of the shaft between the two bearings from fatigue, propagating from an undercut used to locate a split collar which supports the weight of the shaft & pinion on the inner race of the lower bearing. Fatigue is from stress reversal due to separating forces of Pinion and Ring Gear.

 

[Goahead]

It seems doubtful if a stronger steel is the right answer to the fatigue problem. A design review should try to reduce the stress concentration factor of the undercut. Furthermore a local application of shot peening could possibly modify the surface stress conditions that in the present configuration bring about the crack initiation.

http://www.welding-advisers.com/

 

[mog69]

I agree that changing the material (say to 826M40T) would have little effect. I'm looking at increasing the radii which blend the sides into the flat bottom of the undercut to reduce stress concentration. Shot peening is also a good idea. The problem is that a number of these forged shaft and pinions are already in the field overseas, so although these actions can be incorporated for future, I have a potential liability I need to try and evaluate for existing product. Is the driving torque / acceleration / deceleration too great? If I can get fatigue stress data for this material, I may be able to specify a reduction in the working stress by increasing the ramp acceleration times of the driving DC motor for example, which would bring the expected service life up to an acceptable level.

 

[richardmorris]

Don't know if you have seen this thread on eng-tips forum 'Metal and Mettallurgy Egineering', dated July 2003, I found it by typing 'ESDU' and 'BS970' on Google. He's basically asking the same question as you and a number of reference sources are given in responce.

If AISI 4340 is nearly identical to your material as mentioned then ESDU definitely include data for this material in data item 'ENDURANCE OF HIGH STRENGTH STEELS (IN BENDING)'.
This might be a bit too expensive for a 'one off', although the other references might be of some use.

 

[Goahead]

If you can reduce the working stress it would probably pay in extended part life even if you do not know exactly the fatigue data.
Such data may probably be available from your steel supplier.
If you can determine from past failures that they are not Low Cycle Fatigue driven, that is if it takes quite some time from crack initiation to failure, it may be advantageous to schedule non destructive testing to catch the initial crack before it is too late. Advance alert would allow to plan for spare part purchase and substitution at convenient time.
Suitable local NDT could be Flurorescent Magnetic Particles Inspection or Ultrasonic inspection, possibly without dismantling the part from the assembly if accessible.

http://www.welding-advisers.com/