David Roberson proposed a theory where a water clog forms because of condensation.
Because of this clog, pressure and temperature rises until the clog is cleared. If you look at the temperature profile for the 1 MW demonstration, you will see that from 12:45 to 12:55 the output temperature rises sharply from 100 to 112, and stays there until 13:25 where it drops sharply to around 104.5 at 13:27. Interestingly, this is when the input temperature starts to rise from 16 to 20.7 degrees, before slowly cooling. Now if we assume that the warm clog water went back into the pool, causing its temperature to rise, what size and temperature of clog would be reasonable? The main water reservoir has a capacity of 1000 l, but is about 2/3 full. Let's say it contains 600 l. If Tc is the clog water temperature, T1 = 16 C the initial reservoir temperature, T2 = 20.7 C the reservoir temperature after the clog has been added, m = 600 kg the mass of water in the reservoir, and if we take the clog temperature to be, say, Tc = 70 C, we can compute the clog mass by m_c = m(T2 - T1)/(Tc - T2) which in this case gives 48 kg. This is a fairly high but not implausible for a clog, and we don't know how fast the water mixed, nor where the input temperature sensor was. Here is a diagram explaining this scenario: http://i.imgur.com/LLtrR.png Also, note that for water to boil at 112 degrees you need a pressure of about 150 kPa, which is about five meters of water. -- Berke Durak
