An improved multi-objective identification of Johnson–Cook material parameters

A weighted multi-objective identification strategy is presented for estimating Johnson–Cook material parameters often used in general purpose finite element models. The method is shown through illustrative examples for two different materials (Nitronic 33 super alloy and Ti–6Al–4V) using sets of data from quasi-static and split Hopkinson pressure bar (SHPB) compression experiments with varying temperature and strain rates. Results indicate an improved Johnson–Cook model predictability as compared to using conventional ad hoc methods. A sensitivity analysis is performed to reveal the variation of a normalized response rate with respect to the parameters. It is shown that the overall effect of some parameters in the model changes with the range of strain/rate/temperature employed in experiments. An interaction analysis is discussed and signifies the likelihood of high parameter interactions in the Johnson–Cook model owing to its multiplicative form. Using the proposed framework, it is possible to (a) reduce the number of experiments necessary for the identification of model parameters, particularly with respect to the strain rate hardening and thermal softening parameters, (b) consider all possible interactions among material parameters simultaneously, (c) account for variability in the regression degree-of-freedom in different tests by means of a proposed weighting scheme, and (d) arrive at a more generalized set of parameters. The performance of the multi-objective method is also checked with a modified constitutive relation where the number of parameters is increased.