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

  • Lugano active run period 3 recalculation if temperatures where inflated.


    The attached spreadsheet contains a recalculation of the Lugano active run period 3 if the temperatures where inflated due to using broadband instead of in-band emissivities on the Optris thermal camera.


    The recalculation assumes that a factor of 2/3 was already included in the reported rod powers.

    However since it was not expliciy stated that the Lugano testers applied this factor 2/3, the sub page for the rod powers in the spreadsheet also includes the calculation for the case that the factor 2/3 was not applied.

    Note also that since the calculations are based on average temperatures, the results must be interpreted as approximate values.


    The recalculation with the established lower temperatures results in a total convective and radiated power of 1102 Watt assuming the factor of 2/3 was applied to the reported rod powers.


    If the factor of 2/3 was not applied to the rod powers the total calculated power would have been 1042 Watt


    The total applied electrical power for this run was 755 Watt.

  • LDM ,

    I will go over the spreadsheet, as well as the one before, soon.

    I have been a bit distracted by the SK presentation and the resurrection of Doral, but that will soon pass.


    BTW, I think that I have the world record for highest total emissivity conflated with IR measurement spectral emissivity fake COP. I challenge the rest world to beat it. (4.8 “COP”, using the Lugano Protocol). I think a 5 can be done with minor difficulty. 6 will take sone serious attention to details. Gold medal at the COPlylimpics!!! Yay!

  • Lugano heater configurations for an inflated dummy run


    For the Lugao dummy run we got the following values for the power of the different sections

    (see post #424)

    The powers for the different sections in case temperatures where inflated where


    Rods---------------119 Watt

    Caps-----------------87 Watt

    Ribs----------------202 Watt


    For the ECAT heater configuration we can investigate the possible power distribution options by varying the coil diameter and the coil length. The additional straigth wire length under the end caps and in the rods is then determined by the total coil resistance of .41 Ohm.

    The calculated power distrubution should then have appoximately the same values as reported above.

    The following heater configurations where evaluated


    coil 8 windings, diameter 9 mm, coil lengths 19 - 28 cm

    coil 8 windings, diameter 10 mm, coil lengths 19 - 28 cm

    coil 9 windings, diameter 10 mm, coil lengths 20 - 28 cm

    coil 9 windings, diameter 9 mm, coil lengths 20 -28 cm

    coil 9 windings, diameter 8.5 mm, coil lengths 20 -28 cm

    coil 9 windings, diameter 8.0 mm, coil lengths 20 -28 cm

    coil 10 windings, diameter 10 mm, coil lengths 20 - 28 cm

    coil 11 windings, diamter 10 mm, coil lengths 20 - 28 cm


    The evaluation of the power distributions for the above evaluated configurations did not give any one which approximates, even with a large margin, the calculated section powers for the inflated dummy run.

    This while for a non inlated dummy run there are several possible heater coil configurations which are close the the calculated values.


    This is one of the strongest indications that the Lugano dummy run temperatures where measured correctly and thus where not inflated.

  • This is one of the strongest indications that the Lugano dummy run temperatures where measured correctly and thus where not inflated.


    Interesting.

    I imagine that the Dummy has to be pretty close to reality, otherwise the Professors would have started diagnosing the problem right away.

    The method they ended up using seemed to come up with reasonable numbers, and off they went.

    However, the report explicitly goes through the inflation explanation and demonstrates it fully in action with several tables.

    Note that the overall uncertainty estimated for the Dummy in the Lugano report is fairly generous.


  • I felt the need to drag this out and go over it a bit more.

    It seems wrong that the ribs reduce the effective convection as compared to a regular cylinder (no ribs).

    How does the above Correction Factor method compare with a Fin Effectiveness factor (which compares the fin base area to the bare tube area)?

  • The problem with the control run in Lugano remains with the fact that Rossi shut it off while the temps were still climbing.

    A complete and total abort of the control run should have been called by any monitoring or controlling group if that was a truly independent test.

    The inventor interfered and control test integrity was lost then. We didn't know that at the time or we would have asked for a repeat of the control run without Rossi interference.

  • Was the Lugano broadband emissivity iteration done on the Optris ?


    From the text written in the Luagano report it has been concluded that broadband emissivities where used on the Optris and thus the measured temperatures where inflated.

    The Lugano report describes in detail how using an iterative procedure which uses emissivities from the broadband emissivity curve of alumina is used to arrive at the temperature and emissivity used.

    It has been assumed that this procedure was done on the Optris and if this has been the case then this would have led to inflated temperatures.


    On the other hand we have some indications that this has not been true.

    The following investigations contradict the broad band emissivity use on the Optris :


    1. Evaluation of the Lugano dummy run


    For a non inflated dummy run the difference between applied power and calculated power form radiation and convection was investigated in post #393.

    The result was that there was a difference between applied and calculated power of 1.6 %

    For a dummy run with inflated temperatures the difference between the powers was -13.5 % (see post #424)

    Since the error for a non inflated dummy run is much less then for an inflated one this indicates that the temperatures of the dummy run might not have been inflated.


    2. FEM simulation of the Lugano ECAT


    Using thermal finite element simulation on a model of the Luagano ECAT we see that the simulated temperatures are quite near those of the reported (Non inflated) temperatures , an other indication that the reported temperatures of the dummy run where correct (See post #465)


    3. Evaluation of the power distribution


    Comparing the powers dissipated in the rods, the end caps and the ribbed area of the ECAT with the powers dissipated by the heating wires in those sections revealed that for a non inflated dummy run there are several close matches, but there are no close matches for an inflated dummy run. (See post #543)


    So there are doubts if what has been concluded from the written text in the Lugano report, that broadband emissivities where used on the Optris, is correct.

    So I decided to repeat the different emissivity settings of the iterations shown in the Lugano report with the Optris software itself.

    As input the dogbone2_cal_full.ravi file from the second MFMP dogbone thermal test was used with the Lugano style profile for the different measurement areas.

    I used the temperature of measurement section 10 with an emissivity setting of 1.

    Also the background temperature was decreased from 23 degree C to the 21 degree C as reported in the Lugano report.

    The file was then started and then stopped at 24 minutes and 07 seconds at which time the temperature of section 10 was 366.5 degree C, almost the same temperature as the starting temperature of the first iteration in the Lugano report.

    Then for all broadband emissivities used during the iterations shown in the Lugano report the Optris temperatures where determined by changing the emissivities with the Optris PI connect software.

    The tabulated values together with those reported in the Lugano report are given in the following table :


    Emissivity----------Temperatures------------Temperatures

    ---------------------Lugano iteration----------Optris iteration

    ----1.00-------------------366.6------------------------366.5

    ----0.76-------------------426.6------------------------432.1

    ----0.71-------------------443.1------------------------450.3

    ----0.69-------------------450.3------------------------458.3

    ----0.68-------------------454.0------------------------462.4

    ----0.62-------------------478.3------------------------489.4

    ----0.50-------------------541.2------------------------559.7


    As can be seen there is a difference between the values reported in the Lugano report and those obtained when changing the emissivities on the Optris.

    This difference becomes larger for lower emissivity values. For the lowest used emissivity of .5 the difference is 18.5 degree C.

    Because the temperatures reported in the Lugano report are different from those obtained with the Optris software this leads to the conclusion that the iteration proces as outlined in the Lugano report can not have taken place with the Optris.

    Thus it is also not likely that broadband emissivities in the Lugano report where used with the Optris.This is in agreement with points 1, 2 and 3 above.


  • I can confirm that your above Optris conversions agree closely when I tried it using the .RAVI file suggested.

    You may have noted that the emissivity adjustment required to align the Optris-DB2 values to the Lugano numbers is only about an increase of 0.02, but ranges from 0.015 to 0.026 from E=0.76 to E= 0.5 .


    Unfortunately, when using the Padua Reheat 620 C file, I get different conversion numbers, as below:


    Emissivity----------Temperatures------------Temperatures-----------E adjustment to match Lugano

    ---------------------Lugano iteration----------Optris iteration-Padua

    ----1.00-------------------366.6------------------------366.5----------------------- 0

    ----0.76-------------------426.6------------------------429.3---------------------- +0.009

    ----0.71-------------------443.1------------------------446.7---------------------- +0.01

    ----0.69-------------------450.3------------------------454.3---------------------- +0.01

    ----0.68-------------------454.0------------------------458.2---------------------- +0.011

    ----0.62-------------------478.3------------------------483.9---------------------- +0.012

    ----0.50-------------------541.2------------------------550.4---------------------- +0.014


    The adjustment here required to match the Lugano temperatures is almost exactly half of that required for the DB2 file.


    The main differences between the Padua reheat and the DB2 Optris settings are the 150 - 900 C calibration zone rather than 200 - 1500 C zone, and the Optris bolometer emissivity setting of 0.95 instead of 1.0 . 450 C is almost in the middle of this calibration range, so it is the ideal range for this temperature.


    I note also that retroactively setting the background temperature or camera bolometer emissivity makes zero difference to the calculations the Optris performs on a stored .RAVI file. The original values are retained, as they are a critical part of the primary measurements.


    Based on the above numbers, I suggest that the Optris Device emissivity was set to 0.9 for the Lugano test. (Alternately the Device Transmissivity setting may have been changed).

    (The Optris fixed bolometer setting is different from the measurement area emissivity settings, and is located in the software Device settings folder, along with transmissivity settings, background settings etc.)

    That should result in correct conversions from one camera emissivity to another that are consistent with the Lugano report Table 2a and 2b.

    It may also explain the variances noted for the dummy.


    2246-pasted-from-clipboard-png

  • Paradigmnoia


    I can confirm that your above Optris conversions agree closely when I tried it using the .RAVI file suggested.


    Thank's for double checking !


    You may have noted that the emissivity adjustment required to align the Optris-DB2 values to the Lugano numbers is only about an increase of 0.02, but ranges from 0.015 to 0.026 from E=0.76 to E= 0.5 .


    Why would you need to adjust the emissivity values ?

    And while the changes are small in absolute values, as percentage changes they are significant

    See table below


    Emissivity Lugano----------Emissivity Optris corrected--------Difference--------Percentage

    --------1.00---------------------------------0.999----------------------------- -0.001------- ---- -0.10

    --------0.76---------------------------------0.776-------------------------------0.016--------------2.11

    --------0.71---------------------------------0.729-------------------------------0.019--------------2.68

    --------0.69---------------------------------0.710-------------------------------0.020--------------2.90

    --------0.68---------------------------------0.701-------------------------------0.021--------------3.09

    --------0.62---------------------------------0.644-------------------------------0.024--------------3.87

    --------0.50---------------------------------0.527-------------------------------0.027--------------5.40


    Unfortunately, when using the Padua Reheat 620 C file, I get different conversion numbers


    The adjustment here required to match the Lugano temperatures is almost exactly half of that required for the DB2 file.


    Any reasons which you can think off why these would differ ?


    The main differences between the Padua reheat and the DB2 Optris settings are the 150 - 900 C calibration zone rather than 200 - 1500 C zone, and the Optris bolometer emissivity setting of 0.95 instead of 1.0 .


    Since the full range is calibrated why would being in the middle of the range would make a difference.

    I expect that when saving to the file, the (corrected) sensor signal representing the amount of heatflux arriving at the sensor, possibly already corrected for the background temperature, is stored in order when changing emissivities afterwards to always start with the the correct flux.


    450 C is almost in the middle of this calibration range, so it is the ideal range for this temperature.


    I think that the full range is calibrated and that the error margins in the range are the same for all temperatures.


    I note also that retroactively setting the background temperature or camera bolometer emissivity makes zero difference to the calculations the Optris performs on a stored .RAVI file. The original values are retained, as they are a critical part of the primary measurements.


    For the background temperature I also noted that this makes no difference. (see remark above)

    Also since what I expect is that the corrected sensor signal representing the heat flux is stored, original emissivity settings will not make any difference. In that case it becomes only a question of afterwards calculating using the formula.


    Based on the above numbers, I suggest that the Optris Device emissivity was set to 0.9 for the Lugano test. (Alternately the Device Transmissivity setting may have been changed).


    Am curious how you arrive at that conclusion


    (The Optris fixed bolometer setting is different from the measurement area emissivity settings, and is located in the software Device settings folder, along with transmissivity settings, background settings etc.)


    Know where to find those settings


    That should result in correct conversions from one camera emissivity to another that are consistent with the Lugano report Table 2a and 2b.


    As what I expect that the corrected sensor signal is stored then indeed also the bolometer emissivity will not effect later calculations.


    It may also explain the variances noted for the dummy.


    ?


  • LDM ,

    I also yesterday tested the emissivity conversions using the GS5.3 .ravi files. In this case the Device emissivity was set at 0.95 and the calibration was for the 200 - 1500 C range. The results were similar to the Padua reheat one I did above, but slightly higher, typically by 1 C.


    The reason that being in the middle of the calibration range is important is that the values are well inside the calibration range. This is always the ideal measurement range whenever possible in almost all systems. It is hard to say how important that is with the Optris, but one can assume that since there are different temperature calibration ranges that some error creeps in near the edges of a given range that are less in a range that the temperature/radiance is more centered in. Therefore I would say that emissivity-temperature translations done in the more centered range are more likely to be better than ones near the edge of the temperature range. (Would you prefer to take you own temperature to test for a fever with a thermometer that starts at 35 C and goes to 500 C, or one that starts at 20 C and goes to 50 C?)


    Anyways, the general idea was to test the emissivity-temperature translations to see how variable they are. And they are variable, depending on something. It is a good point that the translations don’t seem to match the report values, and there should be a good reason for it.


    I think it is strange that the Device emissivity setting has an effect on the measurement boxes, however. That does seem to defeat the purpose. The Device emissivity and transmissivity setting is usually altered by a link to an outside system, like a factory computer that can change the emissivity required for some process need, like liquid vs solid targets, etc. (The transmissivity setting being used for windows or lenses interfering with the direct Optris view.)


    The other reason I was testing the emissivity adjustments required to match the Lugano temperature-emissivity translations is that I wanted to see if it was a constant shift, or as it is in this case, a non-linear adjustment. This helps define the source of the difference between our tests with Optris software and the the Lugano results that are posted in the report. The small power law adjustment suggests a minor global difference in the way radiant power is being measured from what we are doing.


    I admit that the Device emissivity adjustment seems to be unlikely to be causing the difference, (since the measurement boxes should be independent), but the Device Transmissivity adjustment is quite possible. I think that the effect of a transmissivity adjustment can be evaluated mathematically to compare to the differences we see, and whether it is possibly the culprit. A lens transmissivity adjustment should be equivalent to a bolometer emissivity adjustment. The fact that the calibration range +/- Device emissivity change seems to alter the emissivity difference between the Lugano and our own emissivity-temperature translations linearly, to a power law slope emissivity difference, suggests the answers are to be found in the way the device is “seeing” the radiated heat.


    It would be ideal if an owner of an Optris camera would simply film an object at 366 C , E measurement box = 1, and make a few short films with the transmissivity setting changed (1.0, 0.95, 0.9, 0.8) and the separately the Device emissivity changed (1.0, 0.95, 0.9, 0.8) so the effects of the changes can be considered. Once we enable our Optris software to do the temperature-emissivity translations the same as demonstrated in tables 2a and 2b, we can be more confident in what the higher temperatures were doing.

  • Clearly the Optris calibration range is important. Consider the data below. Although there are minor differences achievable due to mixing hot and cold pixels within a measurement area (emissivity changed here), in general the emissivity-temperature translations are consistent within calibration ranges (150-900 C or 200-1500 C).

    .


  • Cleaned up that plot a bit...

    Might be neat to see what happens in the 0-250 C calibration range. (If it can actually go significantly higher than 250 C in that calibration range, that is.)

    .


    .

    Edit. All new translations begin with a measurement box at 366.6 C or 366.5 C, set at an emissivity of 1. Lugano line is from the report Table 2a and 2b recursive method demonstration, page 15.

  • Paradigmnoia


    Thank's for putting all this effort into investigationg this in more detail.


    One thing what I see is that for both files with a 150-900 C calibration range the values are about equal.

    Also for both files of the extended calibration range of 200 - 1500 C the values are equal.

    But the 150 -900 range has lower values then those of the 200 - 1500 range.

    I wonder why this should be the case.


    When using the iterative approach based on the formula published by Optris and the published n values in post #412 I get when starting with a temperature of 366.6 degree C and an emissivity of 1, a temperature of 567.4 degree C with a n vale of 2.189.

    Using as the maximum error band the published values above and below the 2.189, then the target temperature must lay between 556 and 576 degree C.

    The iteration with the Optris for the used extended temperature range of 200 - 1500 C of 560 degree C is indeed in this range.

  • This is really interesting information. Does it prove that the Optris device was easy to spoof, that it's so complicated that 7 scientists misused it? Because this could become valuable information in a lawsuit against the 7 scientists who published their information and I know of some people who moved money into investments on the basis of that report. It could make Optris legally liable for the losses incurred.


    Once they file the lawsuit and use your information as the basis for moving against Optris, there could be money in it for you as consultants, that kind of thing. But if you're just playing around on a Playground, it would be easy for a counter deposition to show that you really weren't serious about this.


    Why did it take several years to come up with this? Is it real expensive, or is it difficult to generate this spoof?

    Like any equipment there is a right way and a whole bunch of wrong ways to use it. The Optris can be “spoofed” a whole bunch of ways, but spoofing people is a bit trickier. In the case of a group of people watching/using the Optris the wrong way must look like the right way in order to not be discovered to be wrong.


    In this case the error is substituting the Total normal emissivity (this is a single emissivity value that is equivalent to the integrated emissivity for each wavelength entire IR spectrum for a material, which in this case is alumina) for the camera emissivity setting necessary for the IR camera spectral sensitivity band, which just is a tiny part of the full IR spectrum. The IR camera does not see the entire IR spectrum, just a small window of it. So the emissivity setting for the camera is not the same as the Total emissivity. The Total emissivity value is correctly used and necessary for calculating the radiant Output power, however.


    The camera emissivity setting is only used for the IR camera/pyrometer in order to report the correct temperature of the item, and is not used for calculating power. There are several IR detection bands commonly used for IR cameras and pyrometers, and depending on the shape of the emissivty profile over the entire IR spectrum, the different detection bands could have wildly different camera/pyrometer emissivity settings for the same object. All of these could be different from the total emissivity value. For greybody and blackbody objects, the emissivity is the same at all wavelengths, and in that case the camera settings would be the same for all detection bands and the same as the Total emissivity value used separately for calculating radiant power. This can be a complicated subject and I don’t want to write a book on the explanation now if it can be avoided.


    Anyway, calibration is suggested by Optris (and all IR camera/pyrometer manufacturers) to ensure the IR camera/pyrometer is giving correct temperature information. Textbook values and chart summaries are a guide, not the law, about what emissivity value the IR camera/pyrometer needs in specific cases. A positive temperature error is multiplied 4 times when calculating power, and vice versa. This means that the IR camera/pyrometer can be remarkably accurate when calibrated properly since the device measures radiant power and derives a temperature as a result of 4th root of measured radiant power.


    The Lugano device Dummy was not operated to the temperature levels reported for the Active period. The emissivity problem would have shown up for sure if it had. In fact, I have shown that at about 450 C is where the 4th power really starts kicking in compared to a real dummy unit. (The T4 difference starts really going logarithmic at temperatures above 450 C relative to a real dummy temperature. This means that in Lugano the Dummy, even with wrong emissivity values used for the Optris, did not greatly differ from the power expected for the Dummy. Close enough to explain away the discrepancy in measurement uncertainty. At about 800 C the discrepancy in the Dummy power compared to the input power would have been very noticeable. Unfortunately about 450 C (wrong emissivity used) is as high a reported temperature as the Dummy was operated.


    The MFMP Dogbone experiments actually showed this effect, years ago, but it was not demonstrated to the level that it might have been. If all the Optris measurement boxes for the Dogbone are set to the emissivity settings from Plot 1, Lugano report, as was done at Lugano, and the radiant power calculated as done in the Lugano report, the MFMP Dogbone would have shown a COP of 3.8 (or something like that). The MFMP data and Optris files are there so the public can finish the job, but this doesn’t show up on any video stream so the public just didn’t notice it, or even ignored those that did do the final calculations. (I will have to find my spreadsheet to see what I actually came up with for Dogbone Lugano Style COP.)

  • A few notes on "An Indication of..." Levi et al 2012 March Test part.

    (I will add notes as I go over this report again)


    The length used in calculation [10] is 35 cm, rather than 33 cm described for the ecat length. This seems to be so that the 5 area sections can be 7 cm long. The extra adds roughly 42 W to the dummy output, however, due to 56.5 cm2 being added to the area. Subtracting the 42 W from the calculated power output puts the dummy about 119 W short of the reported input. Adding the final 17 W breech area power in (not evaluated yet), leads to a ~100 W shortfall for the dummy combined radiant and convective power.


    The report assumes that the total emissivity for power calculation is the same as the Optris emissivity setting. Hard to say how wrong this is. Might be close. Might not be.

    0.8 to 0.9 total emissivity seems to be typical for dark/high E enamel paint. (Using a total E of 0.95 gives a decent result)