An internal combustion engine has a piston that moves back and forth along the cylinder, a crank shaft that spins and a connecting rod (conrod) between them. If you build a rigid dynamics model of those three parts, you can solve for the forces acting on each end of the conrod at every angle of the crank for any speed (rpm) of the crank.
A classic use for inertia relief is to import from the rigid dynamics solution the force on each end of the conrod at a particular crank angle into static structural. Only the mesh of the conrod is imported into static structural, not the crank or piston. This is an example of a model with no supports. It is a mesh with only forces applied. That is what I mean. Solve that static structural model with inertia relief to look at the stress and deformation in the conrod at that particular crank angle.
If you have enough supports to avoid a singular matrix in static structural, why are you turning on inertia relief?
If you are curious, use inertia relief on a fully supported structure and let us know what you find.