Effect of material strength models on numerical prediction of instability growth solids

Summary form only given, as follows. Considerable research has been devoted to the subject of the instability phenomena that appear on material interfaces. While instability phenomena in perfect fluids and gases are reasonably well understood, the problem of characterizing these phenomena in real materials possessing strength, viscosity, compressibility or phase transitions remains an open question. In 1999, researchers at the All Russian Scientific Institute of Experimental Physics (VNIIEF) conducted a series of tests on aluminum plates that had been given a small sinusoidal perturbation on one side. The plates in these tests are subjected simultaneously to a high pressure (/spl sim/15 GPa) and a high strain rate (/spl sim/10/sup 3/ - 10/sup 6/), a region for which testing mechanisms have only recently been developed. Results for a variety of wavelengths and amplitudes for the initial perturbations and of initial temperatures of the plate are presented in the report "Influence of Thermal Reduction of Strength on Instability Growth in Solids". The goal of my research is to use the data presented in this report in validating various material strength models, notably the Johnson-Cook, Steinberg-Guinan, Preston-Tonks-Wallace and mechanical threshold stress models. The first question of the simulations is whether the different models predict different behavior for a given initial wavelength, amplitude of perturbation and temperature. If so, the next step is to see if there is a consistent pattern to the predictions in terms of wavelength or amplitude of the perturbation or in terms of initial temperatures. CHAD (Computational Hydrodynamics for Advanced Design), a 3-dimensional, parallel, unstructured-grid, finite-volume, Lagrangian method, was used to do the numerical simulations. It is apparent that the various strength models predict different behaviors in the growth of the perturbation of the surface. The PTW and MTS models come closest to capturing the trends in th- data. The Steinberg-Guinan model tends to underestimate the growth in amplitude of the perturbation. The Johnson-Cook results tend to overestimate this growth. Numerical simulations such as this one can be used to help determine the kinds of experiments that might provide useful data for researchers attempting to develop more realistic material strength models, especially in strain and strain rate regimes where little or no data currently exists. In particular, upcoming efforts will include simulations of Atlas pulsed-power experiments.