update**

after gaining some more information I have a feeling it's not thermal cyclic stress but *drum roll*

Stress corrosion cracking.

The original part was in service for about 20 years, and failed. a repair section was welded in place with 316 material and matching filler (not sure if low carbon procedures/ material was used). Part remained in service for about 1 month then failed again. Then a monster 3-4" fillet weld was then placed at the joints to "strengthen" them.  At which point the part deformed from the load.

My guess changed to SCC because the baskets are being sent between an acidic wash ( 1%  sulfuric acid) to a caustic wash, then a sulfate coating (soap) All the tanks are held at 160-180 F(71-85 C) which is well above the temperature threshold for SCC in austenitic stainless steels.  To top it off, the original material was unknown, the maintainence supervisor assumed it was made out of 316 as "that's the only thing that can hold up to sulfiric acid" (His words not mine). I agree this precludes them being made from a lesser grade of austenitic SS like 304, but not from Martensitic SS or Inconel, or Haynes alloy.

I'm guessing the original material was made out of another material altogether that has better corrosion properties than 316. It finally reached it's failure limit at which point the 316 was welded to it. The welds cracked directly down the middle which points to low melt point inclusions or a bad welding procedure as opposed to reaching the yield strength of the base metal (welds were properly sized). If the original material was 316 perhaps it was annealed as to reduce the residual stress, something which was not done on the repair procedure...

BUT then the part failed from plastic deformation away from the weld joint after the monster fillet weld was placed on it. This leads me to think that the part is being overloaded or not designed to carry the weight we are now putting in it. The design is rather poor using open ended C channel material..

hmm more to think about now....better put on the thinking cap

I hope I'm not too late for an answer.

Hooke's law states that in the metals elastic zone the elongation is directly proportional to the stress (tension) required to produce it (the elongation).

Mathematically:  S = E per (L - Lo) / Lo , being  S the applied stress; E the Young modulus, a physical property of every metal that you'll find on engineering handbooks, Lo the original length of the metal specimen and L the actual length produced by the stress S.

Let's suppose now that you have a piece of steel bar on which you've made some sort of jig so as to preclude it to expand freely. The length of the steel bar at room temperature is Lo. The bar material is steel, and in an engineering handbook you've found out the thermal expansion coefficient of steel, i.e., how much of its original length the bar expands for every degree C (or F) of temperature increase. You've also found out E, the Young modulus of steel.

Now you heat the bar to a given temperature, but it can not expand because your jig doesn't allow the bar to expand. Knowing the increase of temperature the bar is supporting and the thermal expansion coefficient, it's easy to calculate the LENGTH THE BAR WOULD HAVE IF IT WAS FREE TO EXPAND, i.e., length L.

So, now that you know E, Lo and L, it's easy to calculate S, i.e., the stress the bar is supporting. This apply when the bar is submitted to compression only; if apart from compression the bar supports also buckling, the calculation is another one.

Giovanni S. Crisi
Sao Paulo - Brazil