gearcutter- Thanks for the link.
Spline couplings do present a very difficult analysis case, primarily due to the large number/combination of geometric variables involved, and how even tiny variations in surface geometry can produce large variations in tooth contact and bending stresses due to the high relative stiffness of the tooth structures.
The basic approach described in the article of using statistical analysis to characterize the fit of a large sample of mating components is often used to optimize the manufacturing tolerances of less complicated situations like fastener patterns. The authors greatly simplify their problem by ignoring the effects of lead variation and torsional stiffness in the spline contact axial direction, so that they are only left with a 2D analysis.
I know very little about statistical analysis, but the article was still interesting.
Terry
