Fatigue life of parts produced by metal additive manufacturing is determined by the complex interaction of defects, surface properties, and material microstructure. Each of these constituents is affected by the choice of processing parameters, as well as the feedstock, machine performance, etc. In addition, stochastic events often contribute significantly to the fatigue life of individual samples or components. Incorporating all of this variability and complexity towards achieving the best fatigue life requires the use of machine learning and artificial intelligence to adapt to changing conditions and extend knowledge gained during expensive experimentation from part to part, machine to machine, and material to material. With this in mind, the proposed effort will demonstrate the feasibility of
Benefit: The lack of understanding of fatigue life for AM components limits the scope of adoption within high performance and high value systems, such as aerospace and medical. By reducing the need for expensive trial and error and integrating knowledge gained through prior efforts, the proposed work would reduce the cost and lead time associated with fielding critical components made using additive manufacturing. This would allow designers to leverage the attendant benefits of AM to the highest degree, including reduced requirements for tooling, novel geometries, unique microstructures and material properties, etc, which will in turn produce cost savings for sustainment, customization, and production of novel designs.
Keywords: Artificial Intelligence, Artificial Intelligence, Fatigue, Machine Learning, Sensors, laser powder bed fusion, additive manufacturing, ICME