You ca use skewness, but the best method of measuring mesh quality is to look and resulting error -- and determine whether the magnitude is acceptable for your given problem and requried accuracy. Most error calculations in ANSYS are derived by comparing different algorithms of result distribution within the elements, the most common being between element and nodal stresses -- thus often analysts plot these two results and intuition and experience also helps judge how acceptable the discretization error is.

I suggest you look into some of the publications on mesh convergence. If you are interested, i can send some more documents. 

Hi,

• you should compare the difference between results you have obtained via different mesh quality. you must check that the results must be independent of mesh size. you have to refine element size until your result does not change.