The model I constructed is fairly simple in form. In this particular case I used a forth order function of power internally generated versus core internal temperature. I also tried many other functions, but felt that the magnitude of the nonlinearity was within reason with the forth order function. The assumed internally generated power begins at 0 watts and then rapidly increases with temperature as would be expected with the forth order relationship.
Could you offer a simple description of the behavior of the negative differential resistance function that you mention? My model also generates a negative resistance once a certain internal temperature is reached. The exact level at which this is reached depends also upon the thermal impendence that the core works into. I can adjust this factor fairly easily in the model and in real life I suspect that Rossi would likely reduce the coolant flow rate and hence raise its associated thermal resistance value at startup to reduce the power input required to enter into the positive resistance dominated region. Once this region is breeched, the positive feedback, as evidenced by the negative resistance calculation, takes over and brings the ECAT up to an active core temperature near the thermal runaway level. The control loop must rapidly begin to extract any excess power once this temperature is reached. A failure at that time will cause the ECAT to melt. It is evident from the model runs and common sense that the thermal runaway temperature can be modified on the fly by the settings of the coolant flow rate and input temperature. This was demonstrated in one of Rossi's earlier test runs where he upped the flow rate significantly to pull the early model into safe turn off. I suspect that even an intervention such as this has limitations unless applied soon enough. Rossi has numerous variables at his disposal that he can modify at startup, operation, and turn off. I hope that we can get more information from him before one of his final designs is thrown into our laps via production in volume. Dave -----Original Message----- From: Jones Beene <[email protected]> To: vortex-l <[email protected]> Sent: Thu, Jan 2, 2014 12:47 pm Subject: RE: [Vo]:Linear Operation of ECAT Modeled Dave, Did you consider anegative differential resistance scenario for the input? This would make fornonlinear operation but it is closer to what Rossi is suggesting. It implies a “sweetspot” in the parameters which should be easier to control since therewould be both positive and negative feedback. From:David Roberson Subject:[Vo]:Linear Operation of ECAT Modeled I have been toying with a new computer model of the ECAT that Iconstructed the other day. The concepts that are being presented arebased upon a simple model of the ECAT that has many assumptions since Rossi hasnot released many of the detailed technical information required to construct atruly accurate one. This particular model run assumes that the internally generated heat powerfollows a forth order function in the region around the thermal run awaytemperature. It can be adjusted to include any polynomial or otherfunction once that has been verified. The main idea at work is that theECAT must use positive feedback in order to operate at a reasonable COP. Negative internal feedback or no reinforcing heat from the powder will not workto a useful degree. The model suggests that Rossi must carefully set the thermal resistance intowhich heat is delivered by the device. If the coolant flow rate isexcessive, which would represent someone attempting to extract too much heatfrom the system, the positive feedback can be defeated and the temperaturewould collapse. This implies that there must be a tradeoff between thevariables which is most likely where a lot of Rossi's time is being expended. I did notice that under the ideal conditions operation slightly below the runaway core temperature can be theoretically controlled and the gain large. My model demonstrates this is possible, but the control system is subjected toa positive feedback behavior which it must overwhelm. Operation at thesetypes of location are tricky since any error in temperature of either directiontends to compound and the device heads ever stronger in that direction. If the core experiences a slight increase in temperature it heads towardthermal run away and must be reversed by the control loop. On the otherhand a tiny drop in core temperature leads to total cooling unlesscompensated. The control loop has to contend with environment changessuch as input coolant temperature and flow rate, or for example changes to theactivity of the powder with time. I am confident that there are manyother factors which attempt to influence the instantaneous balance required atthe chosen operation temperature and all of these require an excess of controlrange for proper allowance. The time constants associated with the device must also be contended with andof course these are not being revealed by Rossi at this time either. Anydelays built into the heat generation mechanism itself further complicate thecontrol system. For all of these reasons, a model such as the one I haveconstructed makes assumptions that will likely be found in error, but at leastthe trends should be revealed. One of the model runs that I conducted assumed that an input power set to aconstant 1000 watts(modified by the loop) could control a total output power of10000 watts for a net COP of 10. Other drives can of course be used whichyield higher or lower values of COP, but this value has a nice ring toit! The thermal run away trip point is within 5% of the absolutetemperature of operation in this particular case. I have noticed that mostany other polynomial relationship between core power generation and temperaturework in a similar fashion to the forth order where the higher ordered functionstend to be more critical. This is to be expected. Dave
