Abstract
It is well-known that materials behavior changes with strain rate and temperature. The changes are described by an equation known as material model. The model involves a number of constants which are normally determined by experiment. In this study, the constants of Zerilli-Armstrong model are identified using dynamic indentation test combined with numerical simulation and an optimization technique. The specimens made of a steel alloy are subjected to indentation tests at four different strain rates and four temperatures and the experimental load-depth curve is recorded. The dynamic indentation test is simulated using Ls-dyna hydro code and the numerical load-depth is obtained. Attempts are made to optimize the error between the experimental and the numerical load-depth curves. This is accomplished using Surrogate model. The results are validated using the stress-strain curves obtained from Hopkinson bar tests. The study shows that the method yields acceptable results. The results obtained using artificial neural network and optimization technique, based on a quadratic polynomial, which have been reported in the previous works for Johnson-Cook model are reproduced here for Zerilli-Armstrong model for comparison purposes.