Calibration of a modified Johnson-Cook model using the Virtual Fields Method and a heterogeneous thermo-mechanical tensile test

•An innovative heterogeneous thermo-mechanical test covering a wide range of temperatures is introduced.•Novel VFM calibration methodology for thermo-elasto-viscoplasticity models is used with success.•Full integration of the VFM and full-field data of heterogeneous thermo-mechanical tests is achieved.•The modified Johnson-Cook model is efficiently calibrated for DP980 using only three tests at different displacement rates.•The positive strain rate sensitivity of DP980 steel is reasonably captured for warm temperatures. [Display omitted] Classical calibration procedures for phenomenological thermo-elasto-viscoplastic constitutive models usually involve a large number of tests to identify all the material parameters, leading to long experimental campaigns. In the present work, a novel calibration methodology that takes full advantage of full-field measurements of a heterogeneous test is proposed. It relies on an innovative combination of the Virtual Fields Method and a thermo-mechanical heterogeneous test carried out on a Gleeble 3500 system. To assess the feasibility of this novel calibration methodology, a modified version of the Johnson-Cook (J-C) model and dual-phase steel DP980 are selected. The three terms of the model that correspond to strain-hardening, temperature, and strain rate effects are calibrated simultaneously for the thermo-mechanical behaviour of this material. The calibration is attempted using a single test carried out at a constant displacement rate. This procedure is repeated for three tests at nominal strain rates of 10−4, 10−3 and 10−2 s−1. Accurate predictions of the flow stress are attained, but the information of a single test is insufficient to capture the positive strain rate sensitivity of the material. The three tests are then combined in an experimental database to calibrate the model. The results show that the positive strain rate sensitivity is reasonably predicted in the considered range of temperatures. These results unveil the potential of this methodology to simplify the calibration process of thermo-elasto-viscoplastic constitutive model.