Similarly, models for LENR/Cold-fusion are not possible with the low resolution data we have from current calorimeters. We have to push the metrology part in order to get sufficient data with low enough uncertainty before we sincerely get into modeling. Bruce_H may be ahead of the curve in his thought experiments on modeling, and who knows? Maybe he may find something that turns out to be useful in the future.
With the very low thermal mass of the systems Russ and I were using meaningful calorimetry wasn't practical, so we depended on 2 factors, relative thermometry between matched systems (working fuel/non-working fuel) and particle emissions- this last being the one we considered to be the main thing. I know that the latter is not really possible using your systems because of the metal containment, but we were using sub 1mm thick alumina.
On the topic of using 'non-working fuel' in control systems, a particular piece of Russ George wisdom was "there is no better control than a failed experiment."