Models and Analysis of Wire Explosions Using TRAC II Simulations

Abstract

In order to understand the dynamics of Z-pinch imposions of thin wires in pulse-power accelerators, it is necessary to understand the physical process by which the initially solid wires are converted into plasma by rising current. For this purpose, we model wire explosions using TRAC II, a two-dimensional MHD code, in three distinct cases: pure tungsten, impure tungsten, and gold-plated tungsten. We compare our results--overall picture of the process, corona linear density, corona mass, and core expansion rate--to actual experiments performed at Sandia National Laboratory and Cornell University and present some explanations for the disagreements between our model and experimental observations. In Chapter 1, we discuss model results for several current waveforms (consisting of a 5 kA 50-150 ns pre-pulse and 80 kA 80 ns main pulse) for a pure tungsten wire, showing that the initial temperature of the wire does not affect the dynamics of the explosion. This suggests that different experimental results for unheated and preheated tungsten wires are due to the expulsion of impurities in the preheated wire and not to a change in the material properties of tungsten. To match the experimental set-up more accurately, we model the explosion of a tungsten wire with impurities in Chapter 2. The overall process predicted by the model agrees with experiment, namely the shunting of the current through the impurities region before tungsten expansion begins; however, quantitative results disagree with experimental observations mostly because of the extreme shunting of the current through the impurities in our model. Finally, in Chapter 3, we compare the explosions in gold-plated tungsten, pure tungsten, and pure gold wires under high (100 kA in 60 ns) and low (2 kA in 270 ns) currents, finding general agreement with experiment in the high-current case and a disagreement by a factor of ten in the low-current case. In addition, due to the similar properties of the two metals, we find no vast differences among the three cases in the high-current case, while the single-metal wire expand faster and farther than the gold-plated wire in the low-current case. We believe that the disagreement between our model and experiment can be decreased by better modeling of tungsten impurities and by improvements in the conductivity and bonding models.