A metrological digital twin of nanoindentation for out-of-control measurements identification

Nanoindentation enables the high-resolution characterization of mechanical properties and the study of advanced materials. However, in the measurement practice, several influence factors affect nanoindentation, introducing bias or increasing measurement uncertainty. Some influence factors are internally sourced (force–displacement noise), others are induced by the environment (electro-mechanical vibrations, indenter-sample thermal equilibrium), and some depend on interactions with the measurand (sample surface integrity, indenter tip wear). Digital twins are simulation models that replicate physical systems in a virtual environment, dynamically updating the virtual model to match the observed state of its real counterpart, enabling physical control of the latter. Digital twins allow predicting system response and highlighting anomalies, while identifying critical influence factors causing such out-of-control conditions, and deploying control actions to avoid malfunctions through predictive maintenance, while avoiding wasting resources. However, to ensure appropriate equipment monitoring and decision-making, traceable digital twins that consider measurement uncertainty are needed. While discussing the advantages of alternative modelling pathways, the work demonstrates the creation of a traceable digital twin for nanoindentation on reference materials based on physics-based data-driven approach. Validation strategies for synthetic and real datasets are implemented, and the capability to detect the most common measurement errors in nanoindentation is highlighted. Strategies to eliminate or compensate for sources of identified measurement errors by the digital twin are discussed, outlining a route towards traceable digitization in nanoindentation.