Failure analysis on friction welding shaft 4

[Best99]

This failure occurred at friction welding area on induction hardened hollow shaft after rotation bending fatigue test about 10,000 cycles. So, please specify the failure mode and crack initiation site for this case.

Thank you for every suggestions.

 

[micalbrch]

Hmmm. Do you have a side view photo or a drawing of the shaft that you can post?

 

[metengr]

Here is a tip, send the failed sample to a comptent metallurgical lab for a proper failure analysis. Otherwise, all we can do is speculate.

 

[CoryPad]

As requested previously, a side view of the fractured part and a part drawing would be necessary to begin our speculation. Did the failure occur through the center of the weld? It appears there are two parts in your axial view image - what is the darker central one? Is it part of the welded joint?

 

[tbuelna]

Bodyslam,

I'm less formal than metengr. I think it's lots of fun to speculate! However, it would make it easier to speculate if you provided at least a little more information. Such as material, heat treat, loading conditions, etc.

I would also inquire as to whether your friction welding process was qualified and controlled, and whether your weldments underwent any type of post weld inspection. At first glance, it appears as though your friction weld process may not have produced a sound joint. A condition like that would have shown up in a post weld mag particle or ultrasound inspection.

Tell us more.

 

[dbooker630]

Also please let us know if there was any proof load testing such as a bend test. Poor friction welds typically don't survive if there was lack of fusion or misalignment present.

 

[Best99]

Dear alls,
Thank you for all good suggestions. Please consider additional information as following items.
1.Side view photo
2. Yes, this failure occurred through the center of the weld
3. What is the darker central one?
It was hole in seamless tube
4. Is it part of the welded joint?
Yes, this failure occurred on friction weld line position
5.Material, heat treat, loading conditions
Material S45C, Induction heat treatment, Bending moment 280 kgf.m. Total life is 10,000 cycles

Any more requirements, please let me know.

 

[redpicker]

I'll take a stab at it.

Understand that an analysis done over the internet based only on two un-scaled photographs and a guess at the base material chemistry and heat treat condition is likely to be worth less than the electrons used to transmit it.

That said, the dark ring in the center is the weld flash. They form on the OD and ID during the friction welding process. The flash on the OD was removed during processing. On the ID, it was obviously left alone.

It appears a fatigue cracked formed on the ID due to rotational bending (the flat, lighter colored region around the ID). The fact the crack propagated considerably more rapidly at the lower right quadrant (in the first picture) than in the upper left indicates bending (either due to an out-of-balance condition or buckling) was involved in addition to the rotational loads. Once the fatigue crack propagated close enough to the surface, rapid failure occurred by torsional overload (the flat region around the OD). The darkest, mountainous crescent-shaped region between the two flat regions was the last to fracture.

I'd take a look at the microstructure and hardness near at the ID. It could be that the post weld heat treatment did not penetrate through thickness. Because of the friction welding process, the weld zone cools very rapidly and assuming this material is similar to AISI 1040, could reach 55 HRC or more in the weld zone. It does not appear to have been a "bad weld" since the weld line appears to be slightly below the plane of the fracture, so complete bonding was achieved. Fracture occurred in the heat affected zone.

There is my 10-minute analysis. Facts not mentioned probably have more to bear than what has been said, so don't put much validity in this analysis.

rp

 

[CoryPad]

rp, thanks for identifying the darker region (flash) - I don't know why I couldn't identify that before. I think your analysis was excellent. I am not convinced about the ID initiation, but oblique lighting of the fracture surface would be needed to confirm. I do agree that the weld appears to be below this fracture surface and should be intact. Time for a sectional view with proper polishing and etching to reveal microstructures.

 

[redpicker]

In re-reading the OP, I see this was a rotational bending fatigue test sample. So, my guess that failure was due to rotational bending fatigue has a bit more support.

The ratchet marks on the ID suggests the ID initation. The chevron marks on the OD torsinal crack point to the locaton where the fatigue crack was closest to the surface. The ID fatigue crack initiated the OD crack.

CoryPad, you are correct, some oblique lighting would improve the appearnce of these fracture characteristics. I have also found that sometimes wetting the surface with alcohol can also highlight some fracture markings. A decent close-up camera and lighting and this could be a textbook example (maybe it is?).

rp

 

[cloa]

I'd say the point of posting on an online forum is for speculation. Anonymous engineering analysis can only lead you to analysis that you can do yourself but have missed yourself or you need the teaching materials or you have a large number of possible directions- which one to try first?

 

[Best99]

Thank you for all valuable posts. I will take this sample for deep metallographic examination as your suggestions