Quote:
Originally Posted by Flieger
What I was trying to say was that this is not a case of uniform, triaxial normal stress, at least the far field stress. The compression is in one direction only and so Mohr's circle tells us that there will be orientations that will see shear stress. Combine that with the pressure in the cylinder which sets up a tensile hoop stress and there will be tension in the material. Aluminum's many slip planes combined with the polycrystalline nature should mean that there will always be a stress to cause fatigue. And since Aluminum alloys do not show an endurance limit there is nothing that can be done to the stud torque that will cause them not to fatigue, though their lives can be extended by using minimal pre-load.
|
I agree that materials will fail in uniaxial compression, and there will be shear stresses developed in this case. The point I was making was that grain orientation in ductile materials does not influence this failure as there are plenty of slip systems available. If it were a brittle material the presence of cleavage planes would make this a different situation.
I don't think stress has anything to do with polycrystaline materials. The crystal structure will clearly influence failure stress and failure modes but unless yielding occurs I am not sure there is much influence.
It is possible that elastic modulus could be slightly anisotropic but this is a bit too deep for the macroscopic behaviour we are considering and I would model for elastic behaviour that is entierly isotropic.
If you know the principal stresses than you could contruct a Mohr's Circle and predict failure stress. By drawing the Circles for Uniaxial Compression and Uniaxial Tension on the same axes you can derive an Circle for intermediate conditions and when stresses exceed this envelope failure will occur.
I think that the Mohr's Circle approach is, however, a little too conservative and think a simple maximum Normal Stress Approach is bit more appropriate.
I am not sure that the stud location of the case sees much of a hoop stress as the peak cylinder pressure only occurs within a few degrees of TDC and there will be a steep gradient along the liner. I am failry sure that the tensile loads produced will be the most significant.
I think that the most appropriate model is to consider the threads in the case and look at the stress distribution in this area.
It would be interesting to carefully model this aspect of the design with a good FEA package and try to evaluate the influence pf preload, temperature and peak cylinder pressure.