An Optimisation Methodology for the Determination of Cyclic Plasticity Model Parameters

Dylan Agius, Mladenko Kajtaz, Kyriakos I. Kourousis (RMIT University, University of Limerick)

CHALLENGE - The incorporation of advanced cyclic plasticity modelling in the analysis of complex structures or structures subjected to complex loading can provide significant improvements in simulation accuracy. However, one of the difficulties with adopting advanced kinematic hardening models is the determination of the material parameters.

SOLUTION -  This study presents the improvements that can be achieved for the multi-component AF Model with Multiplier (MAFM), through a numerical optimization methodology. This approach aims to provide an overall improved simulation accuracy, not only for symmetric strain controlled loading but also for asymmetric strain and stress controlled loading.  A variety of strain/stress loading cases at different plasticity levels have been examined to verify the robustness of the simulation capacity of the parameters. The parameter determination methodology consists of coupling the MAFM material model with an optimization engine modeFRONTIER and its available genetic algorithm MOGA II.

BENEFITS - The simulation results obtained from this study highlight the significant benefit of employing a parameter optimization procedure for the case of the MAFM model. The simulation accuracy of cyclic phenomena occurring from a number of different loading cases was drastically improved with the application of MOGA II, in comparison to the standard protocols used for the determination of model parameters. Based on this analysis, it is recommended to adjust 1y , 2y and 4y when using the MAFM model, in order to achieve an improved simulation.