Failure criterion for crack development of metals?

[bongirs]

Hello,

I have seen that in a lot of industries, the failure criterion is unsigned von Mises stress. Sometimes it may refer to yield stress, sometimes ultimate stress depending upon the scenario simulated. My research peer suggested another approach that I never heard before. Use signed von Mises stress as a failure criterion, why? Because you always get reliable tensile stress strain curve for the material, but not so much reliable compression stress strain curve. Now, we neglect the compression stress in the object as long as visually there is no appreciable plastic flow. The reason is that the material never cracks in compression first, always in tension. So we compare the signed von Mises tension max value with the tensile stress strain curve's ultimate/yield stress to define failure. I would like to ask the experienced guys here for their comments on this method.

 

[realtime2]

Should be ok. A couple of limitations:
1. Not for fatigue. Single loading only.
2. Check the principal stresses too. -that they don't change
direction much during loading, if loading is multiple vectors
that are possible out of phase.
3. If the strains are huge in model you might want to use the
true stress-strain curve for values up to the engineering
ultimate stress. i.e.: apply the correction factors to
convert the ultimate stress and strain to true.
4. If you are using an elastic analysis and you expect hot spot
yield you should apply a plasticity correction to the elastic
model values.

 

[rb1957]

1) research project posts are not allowed

2) signed or unsigned von Mises has little to do with crack development (ie fracture mechanics, stress intensity, fracture toughness, ...)

another day in paradise, or is paradise one day closer ?