I'd have to look up the Dilivar thermal expansion properties, but then it would be a simple calculation. [thermal expansion rate * temperature change * normal length] / [elastic modulus * area of the stud (pi*diameter^2/4)]. That will give you a change in force due to temperature, which you add to the static stud stretch force to get the total force.
You would just compare the change in force between steel and Dilivar. Anyone have the elastic modulus and thermal expansion coefficient for Dilivar?
That would be a first approximation, assuming there is no change in length allowed. Reallly, there is a significant change in length since the Aluminum and Magnesium heat up and expand, but those materials are not changing so each change in force should be adjusted by the same multiplier constant.
I believe steel expands about 1/2 as much as Aluminum and Dilivar is about halfway between steel and Aluminum.
I do agree that up to a point tighter is better than looser. Aluminum is ductile enough that it could yield the first time a stud is too tight, relieving the stress at the high temperature but then meaning too loose cold. Magnesium is not quite as ductile so could cause cracks. This is in the extreme situation.
Porsche would have done the calculations back in the day and determined it was within the material's strength. But then they did subsequently invent Dilivar.
The uneven cooling should not mean significantly uneven stress since both the Aluminum and the steel are exposed to the unequal cooling equally. (
) It does still mean the cylinders go bannana shaped.
They did some interesting things on the 917 like the fiberglass jackets for the studs so that they would stay warm on the intake side.