An Integrated Experimental-Simulation Calibration Method for the Constitutive Model of 6005A-T6 Aluminum Alloy Welds

Abstract Due to the different microstructures caused by the heat source effect, welding joints exhibit significant differences in mechanical properties compared to the base material. Precise characterization of the constitutive characteristics of the welded joint requires a large number of repetitive experiments, which are costly, inefficient, and have limited accuracy improvements. This paper proposes an integrated experimental-simulation-based inverse calibration method, which establishes a calibration optimization problem based on the corresponding constitutive model and a finite element calculation model built by the distribution of hardness in the weldment. Using the global tensile force-displacement curve of the MIG-welded 6005A-T6 aluminum alloy specimen and the experimental data of local deformation with time change obtained from DIC (Digital Image Correlation), the parameters involved in the constitutive models are optimized accordingly. This method can directly obtain the constitutive characteristics of the weldment under conditions of limited experiments and insufficient data. Additionally, the adaptability of the constitutive model to the calibration method and the influence of optimization results are discussed and analyzed. The results indicate that the global force-displacement response of the non-saturated Ramberg-Osgood (R-O) model is in the best agreement with that of the experimental data, and the energy error is only 2.62%, followed by the MPL model, while the saturation-based Voce model shows the largest simulation error in terms of the presented object. Furthermore, the simulation results of R-O, Voce, and MPL models in the local area are far superior to traditional fitting methods.