Theoretically, from the post "Young And On Heat", for a sphere of 1m,
D.∂(ds)∂T=4πEαρΔL
For a given material such that Tmax<Tmelting,
∂(ds)∂T|To∂(ds)∂T|Tmax=ΔL|ToΔL|Tmax
and so,
FTmax−FTo=FTmax{1−L2eL2i.∂(ds)∂T|To∂(ds)∂T|Tmax}
from the post "KaBoom", becomes,
FTmax−FTo=FTmax{1−L2eL2i.ΔL|ToΔL|Tmax}
this implies
L2eL2i=ΔL|TmaxΔL|To
The containment shell would have an external to internal radius ratio given by,
LeLi=√ΔL|TmaxΔL|To
for a containment without internal stress differential,
FTmax−FTo=0
Big kaBoom!