Lugano http://www.elforsk.se/Global/O…er/LuganoReportSubmit.pdf has numerous problems (particularly emissivity, which I was one of the first to address in detail, not properly callibrated etc etc).
Some say that Lugano proves that the hotcat doesn't work. In fact, all it does is FAIL to prove that it DOES work.
So ... forget it!
Instead, go back to Levi http://arxiv.org/ftp/arxiv/papers/1305/1305.3913.pdf (March 2013 run).
HT2 was a STEEL cylinder coated with Macota Enamel (though not uniformly sprayed).
(Slightly complicated by the "flange" and a "breech" at one end).
Dummy run : this was done with known-emissivity dots, thermocouple (agreeing with the dots to 2%), same temperature range (around 300C), single-phase RMS power measurement of 810W to the resistors, correlating with the 910W 3-phase measurement into the control box.
(Only complaint : input power wasn't pulsed like the live run). Emissivity was set to to 0.80 for the cylinder, 0.88 for the breech.
They analyzed two areas directly from the IR cameras (Cylinder: 735W, breech 17W ... but calculated the flange indirectly as 58W -- eqn 22)
Calculated COP ignoring the flange 752 Out/810 In = 0.92 (by me .. not in the paper).
Less than 1, as expected, but close enough to validate the calculation.
Live run: Emissivity was found to vary (with position and time) between 0.76 to 0.80
They calculate the output power as 816W (including an estimate of the flange from the dummy run) and input power as 322 W : COP = 2.6 (equation 36)
Output power : ignoring the flange : 758 / 322 W = COP 2.35
I can't see any major error which gets it anywhere close to COP=1
(Edit : typo)