Rossi Lugano/early demo's revisited. (technical)

  • Lugano dummy run recalculation compare


    This post gives in a table the difference between the published Lugano dummy run data and the recalculation I did.

    (I had planned to make this comparison sooner, but fell and now have a concussion.)

    The difference can be seen in the following table


    --------------------Lugano---------------------------------Recalulation


    -----------Radiation-----Convection-----------Radiation------Convection

    Rods--------56.16----------74.71-------------------54.70------------63.39

    Ribs-------128.56---------101.79-----------------172.35-----------158.50

    Caps--------54.76----------31.29-------------------54.79----- -------34.43


    Totals -----238.57---------207.79------------------281.84-----------256.32


    Total Lugano-------------446.36

    Total recalculation-----538.17


    I have these comments concerning the differences :


    1. Rods


    Radiation for the rods is almost the same, however the convection is lower due to the lower value of the stacking correction factor for the convection.


    2. Ribs


    The radiation of the ribs is much higher in the recalculation then in the Lugano report.

    This is due to the increased emissivity because of the reflection berween the fins (infinite refelction method) an dthe larger area (real area times the view factor to the background)

    The convection is also significant higher an this is largely due to the difference in area's used in the recalculation and the Lugano report.

    In the Lugano report a fin area of 2 x Pi x (Rf^2-Rt^2) is used, Rf being the radius of the top of the fin (12.3 mm) and Rt the tube radius (10 mm) giving a fin area of 3.22E-4 squared meter.

    In the recalculation a fin area of 3.81E-4 squared meter was used, based on the real fin area calculated by substracting the area of two cones.

    This gives an increase of 19% in area but thsi difference is not enough to explain the increase of almost 50 % in convection for the rib area.


    5. Caps


    For the end caps the differences are minor.


    Conclusion :


    Most of the differences between the original Lugano dummy run data and the recalculation can be explained.

    However the reason for the large difference in convective heat transfer of the rib area is not known. This may need some further investigation.

  • I have to point out that inlet water temperature, measured close to the device, starts decreasing roughly one minute earlier.


    The horizontal distance (time delay) between the edge of the inlet water curve and the dropping yellow line of the outlet temperature is indeed almost one minute long, about 50 seconds, but the output temperature take a while to drop from 100 to 15 °C, the line is not vertical due to the thermal inertia of the surrounding metal, so the delay between the 2 edges is much less than 1 minute. If we assume that the flow rate was 100 g/s, considering that the internal volume of the Ecat was about 1 liter, it took less about 10 seconds for the cold water to go from the inlet to the outlet probe.


    Quote

    Also would a 100 g/s be reasonable as a large water flux given that in the same testing environment during the February 18-hour test a rate of almost 1000 g/s was reported?


    On the basis of the available images (1), the reported flow rate of the February 2011 test was an order of magnitude greater than the real one. Therefore the flow rate of 100 g/s you chose is quite reasonable.


    Quote

    I will post the digitized and interpolated data after this will have been clarified.


    I hope to have clarified the issue of the delay between the inlet and the outlet temperature.


    One more hint for the ambient and inlet water curves. Those shown in your last diagram have too many horizontal levels. The pixel lines on the PC screens in their range was only a few, less than a dozen. The jumps between the pixel lines are indicated by the upwards and downwards arrows on second image of my comment above. It would be useful to exactly locate in time these jumps.


    (1) https://www.lenr-forum.com/for…D/?postID=25977#post25977

  • On the basis of the available images (1), the reported flow rate of the February 2011 test was an order of magnitude greater than the real one. Therefore the flow rate of 100 g/s you chose is quite reasonable.


    I have to admit that you had me searching the web extensively (after checking out the sources you indicated) and I couldn't find any visual evidence that could prove or disprove your suggestion that the flow meter used in October 2011 was the exact same specimen used in February 2011 for the 18-hour test.


    One more hint for the ambient and inlet water curves. Those shown in your last diagram have too many horizontal levels. The pixel lines on the PC screens in their range was only a few, less than a dozen. The jumps between the pixel lines are indicated by the upwards and downwards arrows on second image of my comment above. It would be useful to exactly locate in time these jumps.


    This is a semi-artifact caused by the digitizing method, which wouldn't have been visible had I used a linear or higher order interpolation. The original photo does not have a high enough level of detail on that part of the graph to discern the horizontal position of the individual pixels, so I tried to match the overall shape of the curve instead.


    I hope to have clarified the issue of the delay between the inlet and the outlet temperature.


    Not really but it's late and here's the data as used in the last screenshot.


    EDIT: see post #266 for updated data.

  • I have to admit that you had me searching the web extensively (after checking out the sources you indicated) and I couldn't find any visual evidence that could prove or disprove your suggestion that the flow meter used in October 2011 was the exact same specimen used in February 2011 for the 18-hour test.


    Oh, I know. I too did researches on the web, but find no evidence that the two flow meters were exactly the same specimen. Otherwise I wouldn't have put also a question mark at the end of the title of my jpeg.


    It is however 100% sure that the flow meter(s) used on February 10 and on October 6 was(were) of the same make and model, but we cannot be 100% sure that the same specimen was used in both the occasion. Anyway, it is well known that in all the 2011 tests there was the habit to reuse the same stuff, from the yellow pump to the blue bucket. So, unless a contrary evidence is provided, I believe that the same specimen was used in all the Ecat tests carried out in 2011, where a photo shows a flow meter of the same make and model with a white ring and the cover removed.


    May I ask you if you think it is reasonable that a new brand flow meter was used every time it was decided to measure the water flow?


    Quote

    This is a semi-artifact caused by the digitizing method, which wouldn't have been visible had I used a linear or higher order interpolation. The original photo does not have a high enough level of detail on that part of the graph to discern the horizontal position of the individual pixels, so I tried to match the overall shape of the curve instead.


    That's fine. Thanks.


    Quote

    Not really but it's late and here's the data as used in the last screenshot.


    Would you please post the entire set of 4 graphs you posted yesterday, updated by using these last data?

  • [...] May I ask you if you think it is reasonable that a new brand flow meter was used every time it was decided to measure the water flow?


    How many times was it decided to measure the water flow with a flow meter? If this meter was the same one used during the February test, then clearly not very often (probably just twice, in fact), so this feels like a loaded question. We don't know if Rossi had more than one; unlike the pump these meters are cheap and easily replaceable.


    I find it's plausible that it was used in just these two occasions, but then you also have to concede that the likely rounded up value of 1 liter/s was probably a peak value that Levi did actually see at some point during the test, just not continuously. And it's easy to imagine that this was when he/they likely rushed to cool down the reported starting 100-130 kW peak before eventually stabilizing to an observed output of 15-20 kW. How to cool it down if not with a larger water flow than normal?


    To me this feels more like a tale of exaggeration and extrapolation rather than complete invention. Already in a previous occasion (e.g. January 2011) the peak output power observed at some point was thought or assumed to be what the device actually did (or could do) for the entire test.


    But this is just my 2c. I don't subscribe to the mass deception-collusion theory you often try to bring up but there seems to have been a pattern of overemphasizing partial results.


    Would you please post the entire set of 4 graphs you posted yesterday, updated by using these last data?


    Now 5 graphs.



    EDIT: by the way, the flooding flow will have to be revised upwards, since it was capable of at least 178 g/s as measured during the October test:

    https://animpossibleinvention.…st-of-e-cat-october-6.pdf


    And a rounded value was curiously reported on an associated Focus.it article as the flow capacity of the water supply:


    Quote

    Per circa tre ore e mezza l'E-Cat ha dunque funzionato producendo da sé l'energia che gli serviva. In questo periodo di tempo la temperatura del vapore ha oscillato di poco attorno ai 120 °C, mentre la differenza media di temperatura tra l'acqua in ingresso e quella in uscita dallo scambiatore di calore è stata di 5 °C (tenuto conto di un errore misurato nella taratura del termostato) per una portata di circa 600 litri/ora (la portata dell'acquedotto).




    So, despite everything, the energy output is still at least equal to the energy input.

  • LDM ,

    If we, for sake of discussion, suppose that the recursive method did exaggerate the dummy reported rib area temperature, and so reduce it to the temperature it might have been without the exaggeration, might then re-calculated output and reported input become much closer together? I propose about 380 C, rather than about 450 C, simply by looking at table 2, and comparing to the value demonstrated there when using an emissivity of 1.0 . (The temperatures of the other parts can be left the same for now).

  • I find it's plausible that it was used in just these two occasions, ...


    OK. On my side, I find it more than plausible, but it makes little difference. In any case, it is legitimate and reasonable to assume that the same flowmeter (the same specimen) was used on those two occasion. And if it was used ONLY in those two occasion, it means that the lowest value of the total flow - among those shown on the photos taken on October 6 - provides the maximum possible total outflow of the February 18 hours test. This lowest value was 7.6 m3, which corresponds to 422 liters for each of the 18 hours of the test, that is 117 g/s. This value could be even substantially lower if the experimental setup used on October 6 was tested for a while before the public demo. Consider also that the flow rate depends on the pressure at the tap, and that this pressure is higher at night, when most of the 18 hour test was carried out.


    So, in my opinion, the right value for the water flow during the final flooding phase of the December 2010 test should be the rounded up value of 100 g/s.


    Quote

    EDIT: by the way, the flooding flow will have to be revised upwards, since it was capable of at least 178 g/s as measured during the October test:


    [a new set of graphs in substitution of the set posted earlier this morning]


    So, despite everything, the energy output is still at least equal to the energy input.


    The value of 178 g/s is not suitable, because it was measured for a different setup. In the October 6 test, the tap water went through a heat exchanger, not through the Ecat. If you deem plausible that the same flowmeter was used in both the February and October tests, you should maintain the 100 g/s, as you initially assumed, and repost the first set of 5 diagrams which was on this page this morning.


    Using a higher value of 178 g/s - just to shift upward the last part of the EnergyOut curve in order to finally equalize the EnergyIn curve - does make no sense. If you really believe what has been told on the web about the water flow measured during the 18 hour test, you could have used 833 g/s (or even 1000 g/s), and you would have obtained at the end a wonderful COP. Nothing new under the Ecat sky, though.


    Anyway, the last part of the December transient (after the starting of the flooding) is so much affected by possible errors on the assumptions (for instance the delay between the temperature dropping of the inlet and outlet temperatures) that any speculation on its trend has little or no value.


    The important facts happened before the flooding, and your final set of graphs (even in the last badly updated version) leave very little doubt that the water flow have been stopped for about 20 minutes, as revealed by the contemporary increase of the temperature of the still water in the inlet pipe. Furthermore, the added 5th graph shows that, before the water stopping, the outlet power reached an asymptotic level very close to the input power, the balance being the heat dispersed to the ambient. This confirms that the flow rate of 3.73 g/s reported for the December 2010 test (Test 1) was correct, but it also means that with that flow and an input power of about 1220 W the Ecat is able to raise the water temperature from 12 to 80 °C only. This last outcome is very useful to understand the rationale of the successive January 2011 demo (Test 2).


    For the moment, the long and the short of this first test is that the Ecat behavior observed after the stopping of water flow was presented in a report issued with the logo of UniBo as the effect of a nuclear reaction working in self-sustaining mode.


    PS – As for the other observations raised by you about the February test, I would propose to examine them after the conclusion of the present discussion on the December 2010 test, and, if necessary, on the January 2011 test. I think that if we agree on what happened in these two first tests, it will be much easier to find an agreement on the third one.

  • Let's then revert it back to 100 g/s, ignore the floodings and focus more on what happened earlier. A few more questions on the December 2010 test.

    - In the end what happens exactly to the water between minute 37 and 58? In other words, what mechanism allows the assumed stationary water (amount unknown) in an unpressurized system to remain at exactly boiling temperature without it boiling away? Or did some water actually keep boiling in this phase?


    - Roughly how much water do you think was still contained inside the device after the flow stopped?

  • Let's then revert it back to 100 g/s, ignore the floodings and focus more on what happened earlier.


    OK. Thanks for having reposted the earlier graphs and data, and agreed on the next road map.


    Quote

    A few more questions on the December 2010 test.


    - In the end what happens exactly to the water between minute 37 and 58? In other words, what mechanism allowed the assumed stationary water (amount unknown) in an unpressurized system to remain at exactly boiling temperature without it boiling away? Or did some water actually keep boiling in this phase?


    - Roughly how much water do you think was still contained inside the device after the flow stopped?


    Very appropriate questions at this point of our analysis. It is certain that some water boiled away. I guess that the maintaining of constant boiling temperature along the whole period you mentioned depends by the extra heat accumulated in the metal in the period between the stopping of the water flow and the complete switching off of the electric heaters.


    In order to have a numeric confirmation of this hypothesis, we should try to compute all the main components of the energy balance:


    Ein = Emet + Ewat + Eout + Evap + Edisp


    Where:

    - Ein is the energy provided by the electric heaters;

    - Emet is the energy stored in the Ecat metal (and other solid parts) with respect to initial Tamb;

    - Ewat is the energy stored in the water inside the Ecat;

    - Eout is the energy of water exiting the Ecat as liquid;

    - Evap is the energy of water exiting the Ecat as steam;

    - Edisp is the energy dispersed to the ambient through the wall of the Ecat.


    Not an easy task, but not impossible.


    Ein and Eout are the same quantity that you already have calculated.


    Ewat could be roughly calculated assuming that the internal volume of the Ecat is 1 liter, and that all the water is at the temperature measured by the outlet probe.


    Edisp is quite difficult to compute, but when the temperature water is approaching 80 °C the system is at quasi stationary condition, so we have: Edisp (@80°C) = Ein – Eout. This value is comparable to the delta T between the Tout and Tamb. So, assuming in first approximation that the dispersed heat Pdisp is proportional to this delta T we get this very rough estimation: Pdisp(t) = Cdisp * (Tout-Tamb), where Cdisp = 1 W/°C. Finally, Edisp can be obtained by integrating Pdisp.


    At this point the difference (Ein – (Eout + Ewat + Edisp) ) gives a rough estimate of (Emet + Evap).


    Before trying to estimate Evap, and hence the mass of the evaporated water, it would be useful to have the graphs of the above quantities. Could you please compute and post them?


    You are a smart guy, so I decided to challange you as much as possible. :)

  • Ascoli65

    I'm afraid that this calculation might beyond my skill set and I'm not 100% sure of what I'm doing. I need some tips and confirmations, like for example (but not limited to):


    - What mass should I use for the metal and other solid parts in the Ecat? Heat capacity?

    - In general, how should temperature for those parts be expected to behave compared to the water?

    - etc

  • LDM ,

    If we, for sake of discussion, suppose that the recursive method did exaggerate the dummy reported rib area temperature, and so reduce it to the temperature it might have been without the exaggeration, might then re-calculated output and reported input become much closer together? I propose about 380 C, rather than about 450 C, simply by looking at table 2, and comparing to the value demonstrated there when using an emissivity of 1.0 . (The temperatures of the other parts can be left the same for now).


    As I already stated when I presented the results, this is indeed something I planned on doing.

    However such a recalculation does not make sense if we don't have a good explanation for the large increase in convective energy of the finned area.

    I need to know if there is an error in that calculation or not.

    What I already see is that there is a large difference between my convective heat transfer coefficient and that in the report.

    So I have to investigate this first before doing that next step.

    But I have to go slowly since my concussion is playing up.

  • - What mass should I use for the metal and other solid parts in the Ecat? Heat capacity?


    Oh, we don't need to know these data. For the moment we are only interested in computing the trend of the sum (Emet+Evap) by subtracting from Ein (already known) the sum of Eout (known as well) plus Ewat (1 L at Tout with the usual cp) plus Edisp (please, follows the hints I already gave to you).


    Quote

    - In general, how should temperature for those parts be expected to behave compared to the water?


    We have very few experimental data. For the Ecat device, we have only the temperature of the outlet water Tout, measured by a probe inserted on the vertical arm. So, we are obliged to make very rough approximations, assuming that all the water inside the Ecat is at the temperature measured by the above probe, and referring all the other temperatures to this only value.


    For the metal, we can think that the horizontal pipe - heated by the external band resistors - is hotter than Tout, and that the vertical pipe is at about Tout, apart during the rapid cool down after the starting of flooding. Anyway to perform the required calculations we don't need to know the specific temperature of the metals or other solid parts. We can use Tout to roughly estimate both Ewat and Edisp. In this last case, we know that the heat going to the ambient depends on the temperature of the external surface of the device, and that this temperature is far from being uniform. Anyway we can assume, in very first approximation, that this outer temperature is proportional to Tout so that the heat transfer coefficient I suggested before (1 W/°C) is referred to this value, more specifically to the difference Tout-Tamb.


    Ask me again if needed.

  • Is this correct? Is this what you're looking for?


    Yes. Very good and quick work. Thanks.


    However, I suggest you to update this first set of graphs with the following modification:


    - The title of the upper graph should be "Power (W)" and all the quantity labels should be modified from Ex to Px.


    - You probably computed the thin violet curve (Pmet+Pvap) by differentiating the corresponding (Emet+Evap) curve of the second graph, obtaining in this way wide and disturbing oscillations. I would suggest you to first smoothing the energy curve, and then differentiate it.


    I'll tell you later how to split Emet and Evap on the basis of the updated curves.

  • It isn't the final graph, so I haven't gone into putting the finishing touches yet.


    But more than this, personally I don't find reasonable that the mass of the metals and other solid parts (Emet) reach equilibrium in almost half as much time as the water, given the initial assumption of large thermal inertia.


  • But more than this, personally I don't find reasonable that the mass of the metals and other solid parts (Emet) reach equilibrium in almost half as much time as the water, given the initial assumption of large thermal inertia.


    I didn't say that the thermal inertia was large, I only said that it should be taken into account in evaluating the timing. It's not so strange that the metal reaches its equilibrium faster than water, because most of the electric heaters (4 on a total of 5) were band heaters placed at the external of the pipe.


    As for the calculation of (Pmet+Pvap), a much more simple way to compute it is just doing the same as for energy, that is (Pmet+Pvap) = Pin-(Pout+Pwat+Pdisp).


    Eta:

    Even using the above formula to obtain (Pmet+Pvap), you need to compute Pwat by differentiating the numerical series Ewat. So, you will get the same wide oscillations as before. To avoid them, you should use in any case a much more smoothed numerical series of Ewat.


    If there is any problem, the best thing to do is avoid to draw the Pwat and (Pmet+Pvap) curves in the Power graph. They are not necessary for estimating Evap.

  • Ascoli65

    Speaking of Tout, why would it be changing like that during the initial part of the experiment? This could be showing an important behavior characteristic of the system, it's not good to remove it by using artificially smoothed out data.



    It's even more spiky in the original photo, which means I haven't done a sufficiently good job at digitizing the data in my initial attempt:



  • The spiky look suggests the system is producing regular small bursts of XS heat - that would be typical of some types of LENR.


    There are very many ways in which waveforms look "spiky" which are not LENR.


    For example: the input power waveform is spiky (very obviously so). Is this LENR? No, it is some no doubt identifiable, but unidentified for us, mechanism.


    For this system, that has heated water in tubes, there are all sorts of nonlinear mechanisms to do with bubbles forming and breaking away that lead to apparent output temperature spikes.


    For other system, there will be other mechanisms. Spikes are indicative of some low-level unstable nonlinearity in the system and you can get them from almost anything.


    Beware of apophenia when looking at real-world experimental results. It is easy to see patterns in data that are not there, or due to some unexpected and unidentified low level mechanism.

  • Speaking of Tout, why would it be changing like that during the initial part of the experiment? This could be showing an important behavior characteristic of the system, it's not good to remove it by using artificially smoothed out data.


    That's the normal behavior due to the turbulence in a water flux, when the tip of T probe is wetted by eddies at different temperatures that flows through it. You can see that the oscillation are small until t=11s because the water temperature is still quite uniform. Then, up to t=15s, the rising trend stretches the oscillations making them less visible, but they become more evident as the mean temperature approaches the equilibrium value. After the stopping of the flow, they almost disappear because of the stillness of the water.


    So, don't worry, those oscillations don't show any special characteristic of the system. These are typical fantasies of the LENR world.


    So, it is my opinion that a properly smoothed Tout curve better represents the average temperature of the water, and its derivative provides a good and intelligible trend of Pwat, the heat exchanged between the metal and the water.


    In any case, as already said, this is not essential. Don't waste time in doing that. The graph with the curves of the cumulative energies, that you already provided, already allows a reasonable esteem of Evap. I'm just waiting its final version to suggest you how to do.