Failure analysis help

[kingnero]

I've got some positive stops that have failed recently. see file attached for dimensions and pictures (10 MB pdf !).
Loads are ~0 to 50 kN, applied according the longitudinal axis of the legs. There is a minimal slope on the edge, so there is a small lateral component.

Parts have been in service for 20 years, meaning about 10^7 cycles.

No fatigue marks to be seen. This is also a failure % of 0.01% (lots of similar components, but not allot of them are 20 years old).

I see a number of welding errors (lack of fusion, not enough penetration).
Is there a way to check if dirt (coolant fluid from bandsaw, mill scale, grease, ...) could have been the original cause of this,
or, which is my opinion, this is just a way too small weld seeing the plates are much larger/thicker.

Your opinions please?

 

[CoryPad]

I see evidence of fatigue in Figure 7. If the fracture surfaces are not cleaned, you won't be able to see much of anything. What is the desired lifetime? 10⁷ cycles is quite a bit.

 

[TVP]

king,

I agree, with Corypad, evidence of fatigue in both Fig 7 and Fig 2. I also concur with your conclusion that the welds are much too small for the application, and that there is insufficient depth of fusion (penetration) into those thick members.

 

[metengr]

Figure 7 shows evidence of light beach markings along the top weld fracture surface. These were simply fillet welds, which offer less than desirable resistance to fatigue loads in service. Also, lack of fusion was apparent for the weld fracture surface in Figure 6.

 

[Guest102023]

It is quite possible to have fatigue failure of steel after > 10^7 cycles. Stress cycles are rarely constant amplitude over long periods of time. In some components there is a wide spectrum of stresses, complicating the design process.

"If you don't have time to do the job right the first time, when are you going to find time to repair it?"