Reference for allowable Credit on Weld Repair

[edmeister]

Recently performed a weld repair to a solid 4130 1/2" dia shaft. My initial direction was to Weld / fill the chafe damage on the side of the shaft & perform a post-weld heat treatment to restore to original condition. It was brought to my attention that this was not an acceptable process. Say 10% of material diameter was compromised - best acceptance was 90% of shaft properties useful post weld repair. To regain the missing 10% - 4130 Sleeve is to be installed over top the filled in weld/blend & the sleeve welded at both ends. Is there any document / reference that does not allow original strength of a weld/repair? Any input would be interesting.

 

[Sparweb]

All major aircraft OEM's publish a repair manual for their products. Does the aircraft SRM (or component CMM) allow you to do this?

Does this 1/2" shaft rotate once per day or 12,000 times per minute?

Did anybody inspect the shaft for damage after the chafing was blended out? Liquid penetrant?

Also, consider an example like this: Two steel flat bars of equal thickness and width are glued together. Then a bending moment is applied that causes equal amounts of tension on the upper bar and compression on the lower bar. The upper bar is annealed with a yield strength of 35 ksi, and the lower bar is heat treated to yield at 125 ksi. Since both are steel E=30 msi.

When the bending moment applies 20 ksi of stress, the neutral axis of bending is probably exactly at the glue joint.
When the bending moment applies 100 ksi of stress, the neutral axis of bending is nowhere near the glue joint.

This classroom example isn't meant to serve as a realistic structural design example, but as a thought experiment to illustrate what could go wrong with a build-up + grind down of weld blob on the side of your shaft. You said PWHT in your post, but didn't qualify the statement with anything to indicate that it accomplished what you think it might have done.

 

[verymadmac]

MMPDS-01 Table 8.2.2.1.1(a). Strength of Fusion Welded Joints of Steel Alloys

Gives 72 ksi for no stress relief, and 85 ksi for stress relived. 4130N is 90 ksi.


That's before one gets to the interesting questions, relating to failure modes and proving they are still acceptable. A sleeve is cheap and easily analyzed.

Normally I have found this followed by "oh by the way we ran out of spec rods but had this one the local supplier assured us would produce the same strength weld" Que the rest of the day spent playing hunt the weld spec.