A numerical and experimental study of a shock-accelerated heavy gas cylinder

In this thesis we study the evolution of an SF₆ gas cylinder surrounded by air when accelerated by a planar Mach 1.2 shock wave. Vorticity generated by the interaction of the shock wave's pressure gradient with the density gradient at the air/SF₆ interface drives the evolution of the cylinder into a vortex pair.; This thesis contains two interrelated parts; the first part concerns the acquisition of experimental data and the second part concerns the use of this data to benchmark simulations of the experiment with the hydrodynamics code RAGE. RAGE, an adaptive-mesh Eulerian code, has had previous success in simulating shocked interfaces.; Improved experimental diagnostics were used to acquire data, which allowed us to perform a more stringent test of the code's capabilities. From each shock tube experiment, we obtained one image of the experimental initial conditions and six images of the time evolution of the cylinder. Moreover, the implementation of Particle Image Velocimetry (PIV) also allowed us to determine the velocity field at the last experimental time. This thesis is the first code validation study of a shocked flow to use two-dimensional velocity field data for comparison.; Simulations incorporating the two-dimensional image of the experimental initial conditions led to good qualitative agreement with the experimental images. Comparing length measurements of the evolving cylinder and velocity vectors at the last experimental time led to quantitative differences, particularly between the measured and computed velocity magnitudes. The computational study carried out in this thesis showed that agreement between the measured and the computed velocities could be achieved by decreasing the peak SF₆ concentration and diffusing the air/SF₆ interface in the experimental initial conditions. These modifications are consistent with the observation that the SF₆ gas diffuses faster than the glycol droplets used to track the gas.; This thesis demonstrates that quantitative measurements, in addition to qualitative images, should be examined when comparing experimental data and computational results. The quantitative differences between the measured and the computed velocity magnitudes led to the conclusion that the experimental initial conditions were not well characterized.