At 12:08 PM 6/18/2011, Jed Rothwell wrote:
Nothing took so long. They have been doing tests without input for a couple of years. Levi described one in December. However, Rossi claims this mode of running is dangerous because it cannot be controlled.
I've seen some rather silly skeptical comment protesting that you can't control heat with heat. Of course you can. Unstated is that cooling is also a control mechanism. The cooling is by generation of steam or hot water.
Assume that the reaction rate increases with temperature. At a certain temperature, it runs away, and there is risk of destruction of the device and other damage. With a certain rate of cooling and a certain input heat, the reaction can be kept below the temperature at which self-heating is adequate to run away (under the cooling conditions). The heating would be started at a high input to bring the cell up to operating temperature, then lowered to just maintain that temperature. If it's lowered too much, the operating temperature drops and the generated heat drops with it, further lowering the temperature until the whole thing cools down to a (much) lower temperature.
It's being operated, apparently, at a balance point. Other designs might limit the heat by limiting the fuel input, but that might be difficult to control as well, that is, there might be some OOP that is very sensitive. I can see why they'd want to control with input heat, it's pretty simple to manage, electrically, with few failure modes, and it's fail-safe, as long as one doesn't take the temperature up too high. Power failure, the thing shuts down.
This would be the worry, that some uncontrolled condition cause an unexpected increase in reaction rate, taking the cell over the runaway temperature, in which case, obviously, lowering the input heat to zero would be ineffective, one would have to actually cool. Or quench with nitrogen, as was apparently done in one case. Having a cooling port where water would gain more direct thermal contact with the reaction chamber would be a shutdown mechanism that could be controlled.
It is possible that the system could be engineered so that the presence of water in the cooling channels guarantees that the temperature doesn't go to runaway. All it would take is a *lot* of data.... and hard work.
