Re: EMC2 has published a polywell preprint on arXiv
Posted: Mon Jun 30, 2014 4:04 am
The energy distribution curve is wrong in several ways.I have addressed this before. Assuming I am not to far off, any graph that shows the electron numbers and ion numbers (positive charges) must have the same area under the curve, within one part per million. Unless you drew a graph several meters wide with fine lines, you would not percieve any difference in the sum of numbers of each. Also, the energy distribution is critical. The average temperature of the ions and electrons is the same (within some small variance). The graph of the electron energy would peak near the edge and be minimal near the center. The opposite is true for the ions. That is why any comment about electron versus ion energy must have the average energy nearly equal, and must reference the radius from the center to be meaningful.
The magnetic fields the particles see are from two sources. The electromagnets create the field which dominates everywhere except very near the center at low Beta. But, at high Beta this electromagnetic field is pushed outward. There is a much larger volume where the electromagnets derived fields are essentially absent. The other magnetic field effects is due to individual fields of the charged particles as they move around each other. This complex dance is generally described as B field free. This is of course not true. But the important point is that all of these very many local interactions and global interactions between the charged particles can be seen at the very small scales, but at larger scales, the effects cancel each other out so that the gross behavior of the plasma is as if there was no magnetic field at all. Note that the electromagnet effect does not have to be zero within the Wiffleball border, it is just that the strength becomes so small that the resultant lorentz force turning has negligible effects on the charged particles behavior relative to local and global Coulomb forces.Those electrons that are scattered deeper into the electromagnetic field (upscattered) along with similarly upscattered ions (more energy than the electrostatic potential well confinement) do enter the magnetic field domain and more typical spiraling along magnetic lines and ExB diffusion collisions become significant. It is generally considered that this happens only a little for the ions over their lifetime. And for the electrons, while this occurs , it is a minor consideration. The patent application mentioned that electrons that become trapped in the Magrid B field and undergo subsequent ExB drift, only leads to about 1 % of the losses compared to the cusp escape losses. A very through modeling of the system would account for these different behaviors of select populations of charged particles. But, for first order, or general estimates, the portion of the Wiffleball confined plasma that is magnatized can be ignored.
Alphas, or any charged particle product of fusion will have KE greater than the electrostatic potential well. As such they are confined by the Wiffleball confinement just like the electrons. They see the same holes and have about the same chance of hitting the holes as the electrons.
[EDIT- Oops, according to Grad the cusp hole size is limited to the charged particle gyro radius at high Beta. So a high energy alpha would have a larger gyroradus at the same B field strength region, so the hole would be much larger than the electron hole. An electron may make over a hundred thousand passes before hitting a cusp hole, a fusion alpha may make only a thousand passes (Nebel's quote) before hitting it's corresponding larger cusp hole.]
They naturally leave the reaction volume. There is no need for extractors, diverters, etc like in a Tokamak. Provided the alpha particle gyroradius is less than the distance from the Wiffleball border to magrid surface distance, the dominate and natural way they leave is through a cusp. These preferred escape paths lends themselves to direct conversion arrays better than generally isotropic escape trajectories. It is still a complex problem, but much less than with random distributions. An important consideration is that the alphas hit the Magrid only infrequently,and as such they will not heat the Magrid surfaces much. Without direct conversion, they will hit the vacuum vessel walls or other structures and transfer their energy as heat. But these walls are at a greater radius from the center and as such have greater surface area, the thermal loads are less. With alphas exiting through cusps, the cones and sheets of alphas tend to concentrate in certain areas of the wall. This provides the problem and opertunity for localized thermal wall loading- heat (and inpact ceramics in these areas, more cooling pipes, etc. For the magrid the heating concerns comes mostly from neutrons if D-D, and Bremsstruhlung x-rays, especially for P-B11.
Dan Tibbets
The magnetic fields the particles see are from two sources. The electromagnets create the field which dominates everywhere except very near the center at low Beta. But, at high Beta this electromagnetic field is pushed outward. There is a much larger volume where the electromagnets derived fields are essentially absent. The other magnetic field effects is due to individual fields of the charged particles as they move around each other. This complex dance is generally described as B field free. This is of course not true. But the important point is that all of these very many local interactions and global interactions between the charged particles can be seen at the very small scales, but at larger scales, the effects cancel each other out so that the gross behavior of the plasma is as if there was no magnetic field at all. Note that the electromagnet effect does not have to be zero within the Wiffleball border, it is just that the strength becomes so small that the resultant lorentz force turning has negligible effects on the charged particles behavior relative to local and global Coulomb forces.Those electrons that are scattered deeper into the electromagnetic field (upscattered) along with similarly upscattered ions (more energy than the electrostatic potential well confinement) do enter the magnetic field domain and more typical spiraling along magnetic lines and ExB diffusion collisions become significant. It is generally considered that this happens only a little for the ions over their lifetime. And for the electrons, while this occurs , it is a minor consideration. The patent application mentioned that electrons that become trapped in the Magrid B field and undergo subsequent ExB drift, only leads to about 1 % of the losses compared to the cusp escape losses. A very through modeling of the system would account for these different behaviors of select populations of charged particles. But, for first order, or general estimates, the portion of the Wiffleball confined plasma that is magnatized can be ignored.
Alphas, or any charged particle product of fusion will have KE greater than the electrostatic potential well. As such they are confined by the Wiffleball confinement just like the electrons. They see the same holes and have about the same chance of hitting the holes as the electrons.
[EDIT- Oops, according to Grad the cusp hole size is limited to the charged particle gyro radius at high Beta. So a high energy alpha would have a larger gyroradus at the same B field strength region, so the hole would be much larger than the electron hole. An electron may make over a hundred thousand passes before hitting a cusp hole, a fusion alpha may make only a thousand passes (Nebel's quote) before hitting it's corresponding larger cusp hole.]
They naturally leave the reaction volume. There is no need for extractors, diverters, etc like in a Tokamak. Provided the alpha particle gyroradius is less than the distance from the Wiffleball border to magrid surface distance, the dominate and natural way they leave is through a cusp. These preferred escape paths lends themselves to direct conversion arrays better than generally isotropic escape trajectories. It is still a complex problem, but much less than with random distributions. An important consideration is that the alphas hit the Magrid only infrequently,and as such they will not heat the Magrid surfaces much. Without direct conversion, they will hit the vacuum vessel walls or other structures and transfer their energy as heat. But these walls are at a greater radius from the center and as such have greater surface area, the thermal loads are less. With alphas exiting through cusps, the cones and sheets of alphas tend to concentrate in certain areas of the wall. This provides the problem and opertunity for localized thermal wall loading- heat (and inpact ceramics in these areas, more cooling pipes, etc. For the magrid the heating concerns comes mostly from neutrons if D-D, and Bremsstruhlung x-rays, especially for P-B11.
Dan Tibbets