Reduction of Electron Flow Current and Localized Anode Energy Deposition in Transitions from Coaxial Feeds to a Disk

Summary form only given. Many conceptual designs for fusion energy involve combining energy from a number of feed lines to a centrally located load. One such design with 70 coaxial lines (each delivering I_a ~ 1 MA at V ~ 7 MV) merging into a center disk has been shown to yield large (I_e ~ 10-40 MA) electron flow currents. The magnetic nulls inherent in this type of current adder geometry cause an extreme amount of energy to be deposited in localized areas of the anode. A revised design is proposed using 10 coaxial lines delivering I_a ~ 7 MA at V ~ 9 MV that provides better magnetic insulation and therefore lowers the electron flow currents. Computer simulations were used to examine these designs. The 3-D electromagnetic, particle-m-cell (PIC) code QUICKSILVER1 developed at Sandia National Laboratories was used for the numerical simulations. The capabilities of QUICKSILVER are especially well-suited to this problem since it has many features that were developed to model a variety of power flow issues in the Z and ZR accelerators. The simulation results show that switching to a 10-feed line structure (versus a 70-feed line structure) yields a vast improvement in (lowering of) the electron flow current (I_e ~ 1-2 MA). Unfortunately, this reduction in electron flow current only translates into a modest improvement in the energy deposition on the anode in the transition and disk regions due to the abrupt coax-disk transition and the necessarily small A-K gap in the disk region. These issues were addressed using two methods to attempt to lower this energy deposition. Smoothing the geometrical transition from the feed coaxial lines to the disk provides approximately a factor of 3 reduction in the peak local energy deposition. However, lengthening the A-K gaps in the localized areas where anode energy deposition is problematic provides more than a factor of 16 reduction in the peak energy deposition and the combination of the two m- ethods yields better than a 22 fold reduction in the peak energy deposition.