Investigation on geometric imperfections of tensile test specimens using optical full-field measurements and digital twin-based simulations

Standard-based evaluations of tensile tests assume ideal geometry and homogeneous, isotropic material. Based on the Digital Twin concept, a measurement and evaluation system has been built in recent years allowing the monitoring of tensile tests with a fine temporal resolution and full spatial data acquisition technology that provides significantly more detailed data than conventional measurement techniques. This paper investigates whether the theoretical model used in Digital Twin can capture differences between the realistic initial geometry of a specimen and its idealised model. High-precision machining of samples, combined with highly accurate coordinate measurements, results in a fine resolution coordinate map. The geometric imperfections of the finished samples are well within the allowed manufacturing tolerances. Digital Twins of the test specimens were built using two approaches. First, the initial geometry of the specimen's active zone was idealised. For the second, the shape of the test specimen was defined by the best fitting surfaces to the observed results. Simulation results show that computations, based on realistic initial geometry, i.e., considering geometric imperfections inherent in the initial geometry, are much more accurate in tracking time evolution of the specimen geometry –including necking zone location– than computations based on idealised geometry.