mesh quality and independency..!??

michelle

hi..i am trying out some mesh independency tests for a a cylindrical can with a steel rod inserted from the top.. the height of steel rod is half the height of the can and the axis of the can and the rod coincide...

while simulating for the different meshes..i 'carelessly' did not check for the mesh quality.i noticed that the contour plots of temperature show a great variation (from mesh to mesh) while the volume avgd temperatures are more or less same.Once I noticed the discrepancy i chked the mesh quality and found that values of mesh parameters do not adhere to the'mesh visualization advice' (in The HELp).

I am wondering why my results converged at all in the first case ..if the mesh was not of good quality..?

secondly could this be the reason for my not achieving mesh independency?

Thridly I am working with workbench(for geometry creation) and CFX mesh--(so that means unstructured mesh..).. I would like to know the significance of the 'default body and face spacing' and if there is any way of guessing reasonable values of these parameters so that mesh quality is good..

thanks in advance.. michelle