Appears that there was an inner reactor vessel wrapped with helical resistance wires (hence shadows) from size of wires and necessary wall thicknesses this vessel is likely around 12mm diameter.
Inner wall area of outer finned tube about Ø18mm, 0.2m long .0113m² Inconel metal resistance wires can only survive a maximum of about 1350°C without melting (actually probably lower than that over a month long period) >From photo 12a/12b the wires appear to be covering less than half of the core reactor vessel, giving them an area of (estimate) .005m² (this is only a guess) We know that they dissipate 900W of electricity, and inconel has emissivity of around 0.7. In order for finned tube inner wall to absorb 900W from the wires at 1350°C they would need to be around 1000°C. At that temperature they would also transfer approximately 900W to the external environment via radiation and convection. If the inner reactor was any hotter or adding any heat to the system then it would necessarily increase the finned tube wall temperature to increase dissipation to environment, that would in turn increase the wire temperature greatly, including a further bump from the radiative heat transfer from reactor to resistance wires, increasing their temperature to far above the point of failure. These numbers are only approximate (this is a crude calculation only), but I think that quantitatively at least it appears that there is a strong possibility that this demo was producing little if any power, based on pretty simple physical constraints. And most certainly not the 3.8 COP claimed. As to the explanation for the high temp readings - I suspect the IR camera was picking up the colour of the resistance wires and inner reactor vessel body through the partially transparent alumina to give an artificially high temperature reading.
