Don't bother. Having carried out a very basic FEA on a spinning disc I came to the realisation that it's pointless and very difficult to put figures to things.

You can do the basic sums from kinetic energy calcs when it comes to robots hitting each other. For the discs take a figure of around 5kJ of kinetic energy and assume that everything is transferred (never happens) in a certain angle of rotation. You can take the sums further from there.

I was once told by an engineer Stop measurebating and get building. Take from that what you will.

Regarding the base, 4mm will be fine, you could get away with 3 or even 2 depending on the rest of your structure.