Abstract and subjects
Summary form only given. Understanding the physical processes that can lead to the formation of plasma on the surface of metals subjected to megagauss magnetic fields and magnetic pressures of 0.1 Mbar and more is vital for both basic science and a wide variety of applications. "Thick" wire, i.e., rod, experiments on the University of Nevada, Reno (UNR) Zebra generator (2 TW, 1 MA, 100 ns) have provided an extensive data base on aluminum surface plasma formation. "Cold start" magnetohydrodynamic (MHD) computer models, one using a Lagrangian technique with an equation-of-state (EOS) that has VanderWaals loops and the second using an Eulerian technique with a Maxwell-construct EOS, have satisfactorily predicted many of the observations and trends in the observations as experimental parameters are varied. UNR Eulerian modeling has computationally predicted a magnetic field threshold for plasma formation and has led to a conclusion that the plasma formation in the Zebra experiments is predominantly a thermal process driven by Ohmic heating, although the modeling demonstrated significant dependence on the choice of equation-of-state (EOS) and resistivity models. In this paper, we examine the sensitivity of the computational results to various computational aspects such as physical model (e.g., with or without thermal conduction), computational approach (Eulerian or Lagrangian), computational grid size, time-step control, vacuum treatment, EOS (Maxwell construct or VanderWaals loops), and other computational issues. We also discuss the insight into experimental behavior that can be learned from the computations.