Take a look at Advisory Circular 20-128A it shows methods of compliance for rotor burst (admittedly for Part 23 / 25 airframes). Basically the burst zones are defined by fragment size and cone angle from the disk centroid +/- 3, 5 and 15 degrees with fragment size going from a third of a hub down to blade fragments. Each Part 33 engine is certified to contain blades but the necessary shielding to contain a tri-hub failure would normally preclude the airframe from flight. I believe Allied Signal / Honeywell looked at ballistic shielding for the ill fated Ayres Loadmaster beastie, where two T800's were in close proximity to each other.
The AC shows a method of compliance based around system redundancy, sytem placement and total enery of the fragments, not sure how it would work in a Part 27 / 29 installation with so many critical systems in such a small space.
"Tangling turbines" (to shed blades during a shaft shear event and stop overspeed) are not a new idea and designed in to at least two engines I can think of but not sure if they have ever been proven "in battle". It is definately an energy "thing" since smaller APU's per TSO C77 are commonly certified to contain a tri-hub failure and actually do a burst test to prove it - dynamic computational modelling is getting very close in it's predictions though.
To Nick's comment regarding an L1011 burst, I thought it was a DC10 at Sioux City (?) that pushed changes into new design but the tri-jet is alive and well at Dassault!!
Vive la differance ??