A Multiphase Fluid-Structure Computational Framework for Underwater Implosion Problems

The intense implosive collapse of an air-filled underwater structure can lead to ultra-high compressions. Shocks emanating from this process are a potential threat to a nearby submarine hull, particularly in the presence of UNDEX loading. Therefore, with the projected increase of the number of air-backed volumes external to a submarine hull, implosion has become a source of concern to the Navy who now requires an improved capability to design and qualify submarine external payloads for implosion avoidance and platform survivability. However, the development of such a computational model for simulating payload implosion and predicting platform damage is a formidable challenge. It requires accounting for all of the fluid, gas, and structural aspects of this complex problem as well as their interactions. It also requires incorporating in the computations material and failure models, UNDEX loading, and capturing the precise effects on the pressure peaks of many factors such as the rate of structural collapse, hydrodynamic instability at the fluid/bubble interface, and cavitation when it occurs in the fluid. To this effect, this thesis makes essential contributions to computational mechanics that pave the way for the simulation of implosion and the assessment of platform survivability. These include a novel finite volume scheme for the solution of two-phase gas/water problems characterized by strong contact discontinuities. This discretization scheme is equipped with an exact, local, two-phase Riemann solver for computing the fluxes at the material interface without crossing it, which makes it robust with respect to a large discontinuity of the density and a strong pressure jump at the material interface. This two-phase flow scheme is also proven to be contact preserving and nearly conservative. To achieve computational efficiency for complex equations of state such as that of Jones, Wilkins and Lee, this scheme is also equipped with a sparse grid method for tabulating and interpolating data relevant to the exact solution of a Riemann problem. The contributions of this thesis also include an exact local fluid-structure Riemann solver for embedded fluid methods, the implementation of all aforementioned CFD methods in a state-of-the-art fluid-structure analysis environment, and the application of the developed computational technology to the simulation of the implosion of aluminum cylinders with diameter-thickness ratio of approximately 54. The obtained numerical predictions are found to correlate well with the results of the corresponding experiments performed simultaneously at the University of Texas at Austin.