Industrial Heat Amends E-Cat Patent (May 5, 2016)

This patent application, if it refers to the Lugano test(s), has some startling differences from the report by Levi et.al. Specifically, in para. 00063 of the Description section, the calculation of power makes several conservative assumptions, starting with setting the emissivity to 1. It further explains that the Optris thermal image data was taken from the bottom of the reactor, its coolest side. Other adjustments are also made, such as the exclusion of any conducted heat through the end caps where obscured by the support frame.

With these worst-case adjustments, the calculated COP is 5.6 ! Is this calculation in line with the original patent app. (pre-revision)? Please correct me if I'm wrong, but this seems to be new data, answering the objections MFMP and others raised regarding the Lugano report.

I am not sure Alan, I haven't found the original one yet.

However this document seems to contain the amendments:

http://www.cobraf.com/forum/immagini/R_123624373_3.pdf

None of the amendments seem to be related to what you are mentioning.

These applications refer to the E-Cat with the black painted stainless steel housing. Levi et al 2013

Quote

the calculation of power makes several conservative assumptions, starting with setting the emissivity to 1

That is not a conservative assumption: it leads to much higher COP than is real.

More to the point, it is scientifically just wrong since "the emissivity" does not exist in this case.

The revised patent shows data from three new experiments conducted. Unfortunately, and typically for Rossi devices, the conditions in each case are varied.

The first two experiments suffer the standard errors of wrongly using emissivity (which is assumed to be a single value for power estimation and temperature estimation) and not directly measuring temperature, also not having a control.

The third experiment which "particularly well deals with emissivity issues" is now completely different. This new test also has something like a control measurement. But it uses a completely new reactor.

In the (third experiment) control test an input power of ~ 900W gives an estimated temperature of 300C (or so) and similar output power. Black paint is now used, and all seems good.

In the active test the input power cycles between 900W and zero with 6 minute period. The surface temperature is claimed to be similar - 300C - which given the duty cycle is roughly 0.35 would naturally give the claimed COP (from this test) of around 2.8. This is what they observe.

The (first - maybe there are others) elephant in the room? Well, the temperature of the device will clearly change drastically during this procedure since we have 900W added or not to the power budget and there is 4 minutes for the device to cool down. There is no quantitative analysis of this change. The report talks about "average" temperature with reference to the spatial average previously calculated. We have no information about an additional temporal average. If the temperatures are taken at times just after the heating period when the surface temperature is at close to its equilibrium (900W input) value this is what you would expect.

How can I be confident in suggesting this?

Well of course I'm not confident. There are too many unknowns. Those unknowns make the experiment unsafe. But the key thing to note here is that the "SSM mode" behaviour is new, not previously used, and it makes the experimental analysis much more complex. It is therefore a good candidate for having hidden in it some miscalculation that delivers an error. In this case the most obvious error - taking the "900W input power on" temperature and not doing a temporal average over temperature at all times would fit the bill by delivering the observed COP (which is roughly the inverse of the stated duty cycle).

The very slightly revised patent application refers to devices similar to those tested for the first "third party report" by Levi et al., not Lugano.

http://arxiv.org/abs/1305.3913

I was about to defend the emissivity for these tests as being fairly reasonable, (and actually had to re-do this post before posting after catching something).
I did possibly catch an error in equation 3 in the Levi et al 2013 report this application is derived from, in equation 3, page 9 which causes a power of 1505 W instead of 1609 W, and then leads to a COP of 5.36 just using the other values following that (I haven't checked them all yet.)

Specifically, I have 2.56E10^11 rather than 2.74 E10^11 for 711.5^4
Interestingly, it takes about 724K to get the 2.74 E10^11 (724^4), which is only a 12.5 K difference that leads to a ~100 W drop in power!

....Leads to 1464 W after the 41 W from room temperature is subtracted....

Someone might want to see if I mixed something up along the way.

Edit: This certainly cannot account for the majority of the COP, however. Even at 650K, there is is still a significant COP, roughly >3.

Edit 2: Using an online convection-radiation calculator, the 1464 W figure I calculated above is corroborated with a value of 1465 W.
However, the calculator disagrees with the convective power from the report, and reports 361.4 W convection, rather than 466 W (report calculations 9-17), which is another 105 W less (in addition to the 100 W I discounted earlier). This ends up making the COP "5.07", rather than a COP of 5.6 . I am tracking down the differences between methods, and so far the Average Heat Transfer Coefficient (h) seems to be significantly different... but I'll dig in further and see what else crops up.

Edit3: The average heat transfer coefficient (h) seems to be the culprit for the disagreement between calculator and report. The calculator claims it is 8.25083, while the report claims it is 10.64 (calculations 16, 17, page 12). This almost exactly makes the difference.

The calculator used is this one supplied by Slad, and I have used it for emissivity previously and it does seems to give correct results for that.

http://www.thermal-wizard.com/…linder/horiz-cylinder.htm

Edit4: Out of curiosity, I re-ran the works through an ε of 0.8 (both for the camera and total hemispherical emissivity), correcting for the increased temperature (now 506.7 °C or 780 K), and arrived at a total emitted and convected heat energy of 2130 W, and a COP of 5.9 . So whatever the "trick" anyone thinks there may be, it is not from the emissivity of this particular experiment. There are some discrepancies which suggest that the report probably should be showing something like a COP of 5 rather than 5.6 when using an ε of 1.0 as was described in the report. Unless there is some good reason to seriously downgrade the total emissivity of the painted surface to a rather low value (which I cannot recommend at present), the heat output values in the calculated in the report are not a significant cause of the high COP values reported.

Tom Clarke wrote: "That is not a conservative assumption: it leads to much higher COP than is real.
More to the point, it is scientifically just wrong since "the emissivity" does not exist in this case."

Nonsense. Show how calculating radiated power based on ɛ=1 is more than what would be calculated from a lower emissivity.

Scientific analysis of data is based on known and tested behavior of physical phenomena. It is always an estimate, with accuracy characterized by the sum of possible errors of measurement. By setting the emissivity parameter to 1 (as an ideal black-body) the error contribution of the actual value in the experiment is removed from the calculation.

Unless of course you are proposing that the effective emissivity in that experiment is greater than one, or "does not exist" as you claimed. That would seem to be equivalent to saying the Stefan-Boltzmann Law is not valid.

I'm aware of your contention that the bolometer sensing bandwidth is an issue with regard to the Optris data, but that is a separate issue entirely. Here we are just discussing a conservative way to use that instrument in order to get useful data about the experiment. The Optris camera emissivity can be set as high as 1.10, which would be even more conservative. Would you also object to that?