Multi-scale location-specific fatigue life prediction for additive propulsion components

To enable NASA’s technology roadmap for robust and efficient modeling and design of advanced propulsion system materials and structures, QuesTek (QT) will develop and demonstrate a multi-scale process-structure-property-performance modeling framework for aero propulsion component lifing. QT will develop a state-of-the-art integrated computational materials engineering toolkit and implement it in the ICMD® software platform to enable NASA, the aerospace industry, and broader materials intensive industries to leverage accurate and predictive component scale lifing in an intuitive and collaborative cloud-based graphical user interface. Additive manufacturing (AM) is key technology enabler for NASA to improve aero propulsion system efficiency, however, components in these systems are exposed to complex fatigue loading conditions, and paired with the inherent complexity of AM microstructures, durability is difficult to predict. This lack of predictive power that currently exists for fatigue of AM aero propulsion components is a barrier to the broader adoption of the technology. QT will enable composition, process history, microstructure, and component geometry/loading conditions to be captured in a multi-scale PSPP framework to provide robust and efficient component lifing predictions. This will enable concurrent engineering of propulsion systems and alloys. The framework will use CALPHAD based modeling to link composition and process history with microstructure evolution, physics-based mean field analytical modeling to link microstructure with tensile properties, crystal plasticity finite element method to link tensile properties and microstructure with fatigue properties, and Ansys’ nCode DesignLife, a component scale finite element method software to link fatigue properties with component scale lifing. Machine learning will be used for reduced order surrogate modeling to enable uncertainty quantification and propagation through the model framework.