Another possibility is to make a small modification to each component:
Measure the flow rate a little bit wrong, measure temperatures a little bit wrong, calculate a little bit wrong, introduce so much errors and inaccuracies that a single one -if discovered- would prove nothing, but all together make an energy gain.
If he is a real talented chaos experimenter who doesnt doublecheck and who doesnt make plausibility tests, as this seems to be the case, then he might have done just this with a long series of dilletantic experiments and he could really believe in the energy production.

Reply via email to