Taylor Impact Tests with Copper Cylinders: Experiments, Microstructural Analysis and 3D SPH Modeling with Dislocation Plasticity and MD-Informed Artificial Neural Network as Equation of State

Taylor impact tests involving the collision of a cylindrical sample with an anvil are widely used to study the dynamic properties of materials and to test numerical methods. We apply a combined experimental-numerical approach to study the dynamic plasticity of cold-rolled oxygen-free high thermal conductivity OFHC copper. In the experimental part, impact velocities up to 113.6 m/s provide a strain up to 0.3 and strain rates up to 1.7 x 10(4) s(-1) at the edge of the sample. Microstructural analysis allows us to find out pore-like structures with a size of about 15-30 mu m and significant refinement of the grain structure in the deformed parts of the sample. In terms of modeling, the dislocation plasticity model, which was previously tested for the problem of a shock wave upon impact of a plate, is implemented in the 3D case using the numerical scheme of smoothed particle hydrodynamics (SPH). The model includes an equation of state implemented in the form of an artificial neural network (ANN) and trained according to molecular dynamics (MD) simulations of uniform isothermal stretching/compression of representative volumes of copper. The dislocation friction coefficient is taken from previous MD simulations. These two efforts are aimed at building a fully MD-based material model. Comparison of the final shape of the projectile, the reduction of the sample length and increase in the diameter of the impacted edge of the sample confirm the applicability of the developed model and allow us to optimize the model parameters for the case of cold-rolled OFHC copper.

Automated summary

Taylor impact tests and numerical methods were used to study the dynamic plasticity of cold-rolled OFHC copper. The experimental part revealed the existence of pore-like structures and refinement of grain structure in the deformed parts of the sample. The numerical modeling included a dislocation plasticity model and an artificial neural network to optimize the material model for cold-rolled OFHC copper.