Crompto29
Mechanical
- Sep 25, 2008
- 45
Example: If I am modeling a complex object like a weld neck flange, there can be areas where the three principle stresses are positive, similar and well above yield, however according to Von Mises Criterion the Equivalent Von Mises Stress is close to zero.
i.e. if s1=s2=s3~800MPa then Svm~0. Material yield is 260MPa. (the ultimate strength is 515MPa).
If the area of high principle stresses is large enough the flange will fail, wouldn't it? Yet the Von Mises Stress is well below its allowable value. How is this dealt with or even predicted in an FEA model? There must be an additional test.
This issue seems to arise as the Von Mises stress surface is an infinitely long cylinder.
ASME VIII Div 2 requires the additional test: S1+S2+S3 < 4Sy to which my example will fail. This effectively puts a lid on the Von Mises cylinder. Is this the most common method in the wider FEA design industry?
Am I right in saying that Von Mises Criterion is for yielding only and that even if stresses are large and similar (i.e. Svm=0) the object will theorically never yield. (But will at some point catastrophically fail which is a material issue and has nothing to do with Von Mises.)
My main question is how do you deal with the Von Mises infinitely long stress cylinder when Analysing FEA results?
(Von Mises seems to work fine by its self with thin shell analysis as s3 is about equal to 0.)
i.e. if s1=s2=s3~800MPa then Svm~0. Material yield is 260MPa. (the ultimate strength is 515MPa).
If the area of high principle stresses is large enough the flange will fail, wouldn't it? Yet the Von Mises Stress is well below its allowable value. How is this dealt with or even predicted in an FEA model? There must be an additional test.
This issue seems to arise as the Von Mises stress surface is an infinitely long cylinder.
ASME VIII Div 2 requires the additional test: S1+S2+S3 < 4Sy to which my example will fail. This effectively puts a lid on the Von Mises cylinder. Is this the most common method in the wider FEA design industry?
Am I right in saying that Von Mises Criterion is for yielding only and that even if stresses are large and similar (i.e. Svm=0) the object will theorically never yield. (But will at some point catastrophically fail which is a material issue and has nothing to do with Von Mises.)
My main question is how do you deal with the Von Mises infinitely long stress cylinder when Analysing FEA results?
(Von Mises seems to work fine by its self with thin shell analysis as s3 is about equal to 0.)