I think its just cautious testing to see whether the heat will go the way it should, before they place in the transmutatin' fuel.

Actually the lunar surface is lightly sprinkled with water in the form of hydrogen doping of lunar regolith from the solar wind. The He3 mining folks look to bake the volitiles out of the lunar regolith using solar ovens--the volatiles include hydrogen, helium, some carbon, nitrogen at the tens to hundreds of parts per million level. So it would take mining and processing few million tons of lunar regolith to extract enough hydrogen to fill a swimming pool with water.

The beauty is if you're gonna mine anyway--why not extract the volatiles first? I think this makes good sense...

As far as the reactor goes, how are they performing the 'nonnuclear' test? If they insert electric heat elements in place nuclear fuel rods in an engineering mockup, this will only give them an idea of the rate of diffusion of heat through conventional thermal conductivity of the core--but won't give them any data about non-thermal cunductivity heat diffusion, i.e., heat from diffusion of fast neutrons, gamma-rays, and decay products. Those are additional--and very important factors--in nuclear reactor heat balance and design. All of these things have to be accounted for, and this is only one of the --many-- reasons why nuclear core design is so bloody difficult. One needs a comprehensive design tool to simulate all aspects of an operating: I think the industry calls it a thermohydraulic model. Of course, eventually they're just going to have to build one and plop it on the moon. And that will be the first real engineering test!


-- Edited by GoogleNaut at 13:54, 2007-11-20

Yep, that would be a full up engineering test. And then once they had gained enough information of operating modes, it is traditional to go to destructive testing to find the failure modes. Of course with nuclear reactors this is a very messy and expensive affair--cause your test facility is usually destroyed in the process. So, logically, it seems to me that it woulds be best to develop solid designs for a limited class of reactors--debug them--and then more or less freeze the mechanical designs. Then produce them like crazy...

As engineering and operating experience grows, the basic designs may be slightly modified so that later generations will have fewer bugs and increased reliability--but this is the standard, logical evolution process of any mature technology.