Toward digital twin development for additively manufactured turbine blades with experimental and analytical methods

The development of a digital twin is supported by a validated digital replica. To create a digital replica, accurate knowledge of the part being digitally re-created is required. However, that does not automatically equate to an accurate representation of the component. This research applied Additive Manufacturing (AM) techniques to create ten turbine blades. By characterizing the material properties of the AM build plate and the geometric variations of each printed specimen through laboratory measurements, a digital replica Finite Element Model (FEM) was created for each specimen. The digital replica development applied state-of-the-practice methods to understand how AM variations impacted FEM predictions. The work focused on the ability to accurately predict stress values for the purpose of improving fatigue life predictions. The accuracy of the digital replicas were assessed by comparing the FEM predictions of mass, volume, natural frequencies, and location dependent strain values against the printed specimen test measurements. Comparisons between the digital replica models and the “as-designed” baseline model quantified the influences of material properties and small geometric variations in model predictions.