How to properly constrain a spherical pressure vessel?

I'm doing a simple proof of concept analysis between a spherical pressure vessel and a cylindrical one, however, I'm getting results roughly 5% higher than my hand calculations.

My main question is with regards to how to properly constrain my sphere/cylinder. In actual operation, the vessel will simply be resting in a rack. It will not be bolted or fixed to anything rigid. I'm currently using an h-Adaptive test with the Inertial Relief option set, and I am starting to wonder if this is good practice?

In general, is there any way to restrain a model via datum planes or other reference geometry to prevent unwanted stresses?