Optimized prediction of strain distribution with crystal plasticity supported definition of yielding direction

Abstract

The predictability of strain distributions and the related prediction of hardening and failure plays a central role in tool and process design for any metal forming process. Studying yielding behaviour it was discovered, that for various physically motivated yield loci [2,5], no satisfying agreement between DIC measured strain distribution and simulation result could be obtained, even after optimization of parameters and for both associated and non-associated flow assumption (e.g. 8 or 16 parameters). In parallel, crystal plasticity simulations were investigated with the objective to predict the relation between stress and strain ratios for a large number of load cases based on texture measurement. The resulting relations were then applied as input parameters for the plastic yield description and without further optimization almost perfect agreement between forming experiment and simulation was reached. The output can be obtained with either free shape yield loci [13,19], or non-associated flow description [15]. This publication thus presents a novel approach to use micro scale predictions of plastic yielding behaviour to calibrate macroscopic models for metals on the example of an AA6016-T4 aluminium alloy.