I would not worry too much about the level of water in the boiler. See this classical example of dobule retroactive feedback for managing water level ;-) .
http://en.wikipedia.org/wiki/Flush_toilet mic 2011/11/2 Stephen A. Lawrence <sa...@pobox.com>: > > > On 11-11-01 09:36 PM, Jed Rothwell wrote: > > Stephen A. Lawrence <sa...@pobox.com> wrote: > >> >> Since the pump rate was constant, that means the power level was constant >> with a precision of +/- 0.09 percent. (That's 9/100 of 1 percent.) This, >> in a process which is said to be hard to start and hard to control. > > Either that, or the water level fluctuated. That seems more likely to me. > When it starts to rise, you increase the reaction. When it falls too far, > you throttle it. > > This is, of course, all old stuff being reiterated here. In the test from > last spring, the electrical power level was (supposedly) fixed; if it wasn't > then the calorimetry was nonsense. Consequently it's not at all clear how > the reaction rate was being controlled; the system, as described, was > apparently running open-loop. (Some people have imagined interesting > feedback controls in the blue box but no such thing has ever been claimed > by anyone who actually knew.) > > In the 1MW test it's less clear cut, but one thing stands out: There's no > obvious indicator that Rossi could have used to tell him when it was time to > turn it up or down. Output temp would lag too much to be used as the > control variable, and the result would have been a "hunting" temperature > which wandered all over the place, certainly not an essentially constant > temperature which was indicative of a power level which was nailed to better > than 1/2 %. It would be nice to imagine a sight glass, and Rossi's hand on > the throttle with his eye glued to the glass, but it's not clear such exists > anywhere except in our imaginations. > > Whatever, all such concerns have been dismissed in earlier posts, so there's > not a lot of point in arguing it further. > > >