Creep and Stress Concentration Factors

[mikepagnotta]

I am working on a project with some fairly exotic materials at high temperatures (Molybdenum, Hastelloy, Inconel 718, all at about 800 degrees C). We have been able to get some creep data for these materials, but the data does not seem to distinguish between notched and unnotched specimens. The basic question that I have is: When calculating a stress to compare to creep allowables, do you need to calculate the notched stress (at a hole in a plate for example), or does the creep allowable pertain to the far field (gross cross-section) stress? Any insight would be greatly appreciated.

 

[plasgears]

You can take the creep data to be independent of SCF. SCF is derived from a set of variables, and it needs to be separate from the creep data.

 

[mikepagnotta]

My follow up question/comment would be that if using Finite Element Analysis to determine component stress levels, all things considered (accurate representation of hardware, proper boundary conditions, etc.), the model stresses are generally a conservative value to compare to the creep data. This is because depending on the model refinement, FEA will almost always have some stress concentration effects in the results. Again using the example of a plate with a hole, even if the hole is coarsely modeled, you will see an increased stress at the elements in the vicinity of the hole. These stress values would be conservative when comparing to creep data. Do you concur? An exception would be if a finite width plate has a hole but it is not modeled at all. In this case the FEA results would be unconservative even from a net section stress standpoint.