On a novel heterogeneous mechanical test with continuous strain path changes for the inverse calibration of constitutive models

In recent years, the simulation of sheet metal forming has increasingly relied on calibrated constitutive models to predict material behaviour accurately under complex loading conditions. With the emergence of the Material Testing 2.0 paradigm, this calibration now uses full-field measurements combined with inverse identification techniques, reducing the overall experimental effort required with the classical calibration method. The accuracy of this process, however, depends significantly on the heterogeneity of the material behaviour during mechanical testing. Although heterogeneity brings the opportunity to capture multiple strain states in a single test, most current mechanical tests only provide information concerning continuous strain paths, offering limited insight into the material behaviour when there are strain path changes. Understanding the material behaviour when subjected to strain path changes is critical for accurately modelling sheet metal forming processes, such as deep drawing. This paper presents a mechanical test, inspired by the classical Arcan test, that introduces non-linear strain path changes without unloading the specimen. The proposed test aligns with the Material Testing 2.0 vision by increasing data richness and capturing multiple strain states in a single experiment. The new test is numerically analysed and the material’s deformation behaviour is evaluated using two heterogeneity indicators and the Schmitt parameter, a strain path change metric that quantifies the deviation in the strain path throughout the test. These results are then compared to those from the Arcan test without strain path changes and used to select the most appropriate test configuration for each mechanical test. Results demonstrate that incorporating strain path changes improves sensitivity to material anisotropy and contributes to a richer kinematic field dataset for more robust material characterisation. Material parameters for a dual-phase steel are identified using numerical data from both tests using the selected test configurations, allowing a direct comparison. To bridge the gap between numerical and experimental approaches, the numerical results are used to synthetically deform speckle pattern images, which are further processed by Digital Image Correlation. Finally, the full-field strain maps of both the reference and calibrated model tests are compared.