Do you really need to do physical tests anymore?
Yes, absolutely. They have to rediscover what Oppenheimer et al figured out during the early '40s, not on the theoretical front, but on the engineering front. It isn't good enough to know the rough outline. To be sure that one doesn't get a misfire, or a disastrous mistake, something like 10 tests have to be run.
Can you get good enough certainty that a bomb design will work with software simulations?
You can't get anything with simulation because the dynamic regimes reached exceed those of other explosive materials so you don't know boundary values. It is possible to incrementally develop so that ever higher heat regimes are reached without going chain reaction. Controlled chain reaction is key. One must have rich enough fusile material plus "charcoal", neutron absorbers, to slow the reaction. If the heat regimes can be managed by configuration, then, in theory, a bomb reaction rate would be controllable by the density of the absorber. It may be possible to scale the explosion down to a level undetectable by the IAEC where one could fine tune the quantities and disposition of critical ingredients and then formulate a bomb. This would require that the formulate was scale invariant so that a powerful bomb would be possible by merely increasing the amounts of ingredients. Nonetheless, without testing in scale, you're only guessing. In the world of such bombs, you have to get it right the first time. |
