Three-dimensional MHD simulations of a plasma opening switch

Summary form only given. Pulsed power sources such as fast capacitor banks and flux compression generators can deliver tens of mega-amperes of current but require from a few to few tens of microseconds to reach that level of output. Plasma loads designed to achieve velocities of tens of centimeters per microseconds need current pulses of that magnitude but function in a few hundred nanoseconds. Using the former to drive the latter requires a transfer switch that can conduct those currents for the longer time scale with low losses and switch the current to the load in the shorter time scale. The plasma flow switch is just such a device. It consists of an annular plasma armature that is accelerated over the long time scale of the power source along a coaxial plasma gun toward an abrupt reduction in radius of the inner electrode. When the armature passes that end, an arc of low density plasma sweeps inward at very high velocity, quickly transferring the current to the load on its short time scale. It is important that the plasma armature move slowly, so that it does not extract too much inductive or kinetic energy from the source, and that the switching arc move quickly, so that as much current as possible is transferred while the load can make full use of it. Plasma armatures, whether formed from wire arrays, foils, or gas puffs, need to be symmetric, since they are subject to a variety of instabiliies during acceleration, a requirement is further complicated by the 1/r
²
variation in the magnetic driving force. Excessive armature asymmetry will allow the low density plasma behind the armature and its embedded field to break through the higher density part of the switch resulting in early opening before the desired current is reached. For over two decades, we have simulated plasma opening switch behavior using 2-d MHD in the R-Z plane. Those simulations have predicted experiments well even though they could not capture current filamentation. Our recent perturbed 3-d simulations show us that armatures can indeed filament along the direction of the current; however, that filamentation is Rayleigh-Taylor, not thermal. Thus, these filaments collect both mass and current density, so the force goes up with the mass, and on the whole, the filaments move like the unfilamented armature.