contact stress

[gio1]

Hi

I am looking at the sensitivity of contact stress between a piston and its gudgeon pin to different pin diameters. The piston is loaded by inertia force and bending+shear develops on the pin. The analysis is in Abaqus/standard and the contacting elements are C3D10M on one side and linear hexahedral on the other (pin).

In the area of highest contact stress, on the edge of the engagement, VM contours (both nodal and at integration points) are smooth but when looking for a "maximum contact stress value" they are misleading: despite a fine mesh density there are some hot spots (always on midside nodes) where peak stress increases with pin diameter (which is not logical!) whilst if I take an AVERAGE of VM stresses along the edge then this correctly decreases as the diameter increases (however the average value does not reflect the real magnitude of contact stress).

Do you have a possible explanation for this? And is it sensible to consider VM values for comparing contact stress? Or is it better to consider other values (P1, pressure?)

Thanks

 

[johnhors]

C3D10M elements have a problem recovering stress values, which Abaqus are well aware of and have been working to correct for some time ! The average stress is probably the most accurate value to go for in this case.

 

[gio1]

Thanks johnhors, I was not aware of this and was confident that C3D10M were well suited to contact analyses, as this is what Abaqus reccomend