High compressive stress in steel

[gurmeet2003]

Greetings every body,

I have a question about interpretation of non linear results. I have made a non linear run on a compressor cylinder head. The loading is assembly bolt load. With a linear analysis high stresses are seen at the fillet (1/32 inch radius). These stresses are higher than yield and are the cause for doing nonlinear analysis with non linear material properties. Material is AISI 4340 and is modelled as Bilinear isotropic material (BISO; isotropic hardening with mises criterion).
Expected failure mode is static overload and not fatigue. Nonlinear FEA run is made at twice the design load. Von Mises stresses are just within limits (almost equal to UTS). However Minimum principal stresses (Compressive stresses)are 25% higher than UTS. My question is as follows:

A. Since the material is ductile (11 % elongation) should I consider the high (low?) compressive stresses or ignore them?
Or can I assume that since Von MIses stresses are ok I have Factor of Safety equal to 2?

Thanks,
Gurmeet

 

[sdm919]

My first thought is that you are OK assuming you accept the Mises criterion for both yield and failure, and that your model is good in all other respects. That is, yielding occurs when Mises stress = FTY and failure occurs when Mises stress = FTU. The failure evaluation assumes that you have done a post-yield material nonlinear analysis, as you have said.

As for the principal stresses, consider the 2D failure envelopes for the Mises criterion (an ellipse) and the maximum principal stress criterion (a square). It is possible for some states of stress to be outside the square but still inside the ellipse. That might be the case you have. As a check, look at the stress components in the critical element(s) and compute the Mises and principal stresses yourself and see where they fall relative to the failure envelopes. The 3D case might be a little harder to visualize, but the same idea applies.

Also, I don't think compression failure due to exceeding a material allowable stress has much meaning, assuming you don't have any buckling issues. It's common to set FCU = FTU, but that is mainly for convenience. A part breaks into 2 pieces when FTU is exceeded, but just keeps "squishing" when "FCU" is exceeded.

To be a little more conservative than Mises, you might want to ensure that the MAXIMUM principal stress is less than FTU, but I would say that if the maximum Mises stress from an accurate nonlinear analysis is less than FTU, you are OK.