Stainless steel fitting - embrittlement

[kclim]

Folks,

Have a stainless steel fitting, small valve welded to a pipe section, maybe about 3/4" in diameter.

The operating conditions it sees internally are 260C, 600 kPa.

Was told by operating personnel that the line suffered from an impact, which caused failure.

Material analysis has shown the pipe to be 321, and 316 on the valve side and weld.

Given that the failure has minimal deformation, despite the materials being typically ductile I suspected fatigue (although can't find evidence of this on the fracture surface) or embrittlement (sigma phase or otherwise). Although the fitting doesn't operate in the range which would promote embrittlement, perhaps uncontrolled welding may have contributed?

Was wondering if anyone had thoughts on what the mechanism could be - have attached a word doc with some images.

Thanks

 

[metengr]

I don't see impact as being a probable cause because if you evaluate the location of failure, I see no evidence of impact damage/deformation away from the failure.

At this operating temperature, and how the line is configured I would believe more in low cycle fatigue crack propagation along the toe of the weld just based on the appearance of the fracture surface.

 

[EdStainless]

Given the deep notch along the inside of the joint any minor weld defect would make crack initiation/propagation easy.
What filler was used for the weld?


= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[mrfailure]

Looks like brittle overload to me and I can see how a single event impact would cause fracture at the toe of the weld. You may want to consider more in-depth analysis by a lab specializing in failure who can perform SEM fractography performed along with characterization of the base metal and weld.

 

[Metaljon]

Agree with Metengr it looks like a fatigue failure propagating from the toe of the weld. If it was a ductile failure from impact then I would expect a shear failure which it is not. Not sure what the operating environment but would not rule out some SCC contribution. You need to do a failure analysis to determine cause.

http://nz.linkedin.com/pub/jonathan-smith/39/25/589