Wikiversity/Cold fusion/Nickel-hydrogen system/Parkhomov
December 25, 2014, Alexander Parkhomov, a Russian scientist, released a preliminary report on an experiment he performed. In Russian. (Note that the first part of this is about the Lugano report, see Rossi Tests). In English.
This experiment used water phase-change calorimetry to estimate heat evolution, and in the final stage of the experiment, before the heater burned out, Parkhomov's calculations showed excess energy of 158% of input (i.e., COP 2.58].
Abd Lomax has done an analysis of the temperature data from the Parkhomov experiment.
The Parkhomov report explicitly specifies input powers of 0, 25, 300, 394, and 498 W.
To understand the temperature profile given in Parkhomov Figure 5, it was hypothesized that the additional input power steps (obvious from the temperature data) were at 50, 100, 150, 200, and 250 W. Then, for each period, a 10 minute period generally ending with the last reading before temperature increased again was translated to an average of minimum and peak temperature for that period. For the last period, the 498 W period, the 10 minute period was moved back to just before the erratic oscillations in temperature began.
In the Parkhomov experiment, there is no specific indication of when water began to boil. However, the first analysis of XP is for the 300 W. input period, and only 0.2 Kg of water was reported as evaporating. The three data points calculated COP for inputs of 300 W, 394 W, and 498 W., if extrapolated to below 300 W., indicate COP of 1.0 at about 275 W.
Later, Parkhomov revised his calculations, correcting two errors or inaccuracies, resulting in a COP at 300 W input of 0.99. Those corrections increased calculated COP at 500 W input to 2.74. -- note added by Abd (discuss • contribs) 16:40, 21 March 2015 (UTC)
The data, with its linear region followed by a parabolic curve, appears to be what would be expected from normal heating, it shows no room for major excess power; if the water bath is being heated, the reactor itself, from which the heat would be derived (if not from the heating coils around the reactor) is not increasing significantly in temperature.
In the region from 300 - 498 W input power, temperature is increasing roughly 1.4 degrees per watt. If the XP in the 498 W period were 158% of input power, as claimed (based on the evaporation of water), that would be 786 W additional power. It would be expected to raise the temperature by over 1000 C. Because the temperature will not be a linear function of power in that regime, it could be less, but even 100 C increase is more than the data can support.
The last 8 minutes of the 498 W input period, temperature began to oscillate above and below the prior mean temperature; the heater coil then broke, and temperature fell about 100 C, and then a "thermal arrest was observed." This has also been called "heat after death," i.e., apparent power production after the removal of heating power.
The apparent power level for the thermal arrest period, about 7 minutes, based on reactor temperature, can be seen to roughly match the 400 W input period. If there is XP in the data, this would be it. However, experts reviewing the Parkhomov data have suggested that the erratic temperature seen in the last 8 minutes of the 498 W period indicates thermocouple malfunction, notice that the power goes above and below the established temperature; XP appearing at that point would not, if relaxed, return the temperature to lower than the 498 W level.
The heater coil was about to fail, and the reactor might have been cracking. An analysis was done of the total hydrogen available (about 10 mg), perhaps it began to escape and burn. An estimate of the available power from this indicated that it could maintain the 400 W power level for about 7 seconds, not seven minutes. However, the apparent thermal arrest could also be due to thermocouple malfunction.
There is much information necessary in order to understand the Parkhomov experiment that is missing from the report, and there are anomalies or errors in the report. As an example of an anomaly, the temperature record shows, after the thermal arrest ended, a fall of temperature to about 20 C. However, at that point, the reactor would have still been in the metal box, immersed in water at almost 100 C. Perhaps Parkhomov removed the metal box from the water bath?
There is a presentation scheduled by Parkhomov in Russia, on January 27. Most reports of the Parkhomov experiment do not examine the temperature data, but only cover the phase change calorimetry claim. What is ironic here is that Parkhomov did his experiment using phase change calorimetry, pointing to problems with the complex calorimetry of the Lugano report.
Parkhomov has written people studying his work that he calibrated his calorimetry using a reactor without fuel, and up to 1000 W input., and that the calorimetry was accurate within 10%. However, the heater design would likely not be able to handle 1000 W input -- it burned out at 500 W. That conclusion, though, depends on the analysis above. If Parkhomov did calibrate using a reactor, then he would have had, presumably, temperature data, which he did not report. Nor did he report his actual calibration data.
Parkhomov did no-fuel runs on January 2, 2015, reported later. There is no detailed thermometry available from these runs, only his final conclusions and a temperature reached. It appears that his heater burned out during the third input power period, as expected, so that crucial period, which only had 928 W input, not 1000 watts, was short, only 16 minutes. There would be substantial thermal inertia, so his result of COP 1.12 for that period, besides being over 10% error, would likely be low. His method of measuring water was crude and imprecise, in later reports, he's periodically adding water, in multiples of 100 g., to a level mark inside the heat bath, not measuring actual water loss.
He reported no-fuel calibration at 433 W, reaching a temperature of 470 C. At first glance this would seem conclusive, because 400 W in the first experiment reached a temperature of 1150 C. However, he shows an input power of 211 W reaching only 210 C. In his first run, 200 W input reached a temperature of almost 800 C. That is a period when the phase change calorimetry shows no XP. Something is very different about the setup. It is clear: Parkhomov had insulation in the metal box in the December 20 run. Probably none in the January 2 calibrations. So the heat flow was radically different, and this is why calibrations should be done with the same conditions, as closely as possible, as the experimental run. --note added by Abd (discuss • contribs) 16:40, 21 March 2015 (UTC)
The largest problem with the Lugano report is that they did no calibration with input power. They did a "dummy" run, i.e., with the Rossi device without fuel, but input power was held below 500 W, stating a reason as being to avoid damage to the heater coils. However, in the actual experiment, input power reached 910 W.
So the Parkhomov experiment ends up with the same problem as the Lugano experiment: no clear calibration. Calibration of the Parkhomov evaporative calorimetry, if done with other than an exact copy of the device and the insulation, could result in different boiling behavior, and thus a difference in entrained water, the classic problem with water evaporation calorimetry.
Experts in the field, commenting privately, expressed doubt that anyone is able to seal an alumina cylinder against hydrogen leakage using high-temperature cement. The alumina itself is said to be porous and to leak hydrogen. However, discussions on the vortex-l mailing list have suggested that the lithium-aluminum alloy that forms when the LiAlH4 decomposes may seal the alumina.
There are now at least two other groups attempting to replicate Parkhomov. One of them, the Martin Fleischmann Memorial Project is also working to calibrate the Lugano experiment using an imitation "dogbone" reactor, matching the Lugano reactor in outward construction, planning to use the same Optris IR thermal imager that was used in Lugano.
This paper provides Parkhomov's corrected data.
There is discussion of this on the attached Talk page.