skip to main content
Caltech

Microstructural effects during fatigue and fracture of additively manufactured metals

Friday, November 19, 2021
3:00pm to 4:00pm
Add to Cal
Online Event
John Lambros, Donald B. Willett Professor of Engineering, Aerospace Engineering, University of Illinois Urbana-Champaign,

Additive manufacturing (AM) of metallic components is increasing in popularity in the aerospace industry and elsewhere. Several major aerospace companies have already begun implementing AM technology into the production of metallic components, such as the fuel nozzle tips in GE Aviation's LEAP engines, and SpaceX's SuperDraco rocket engines. However more widespread use of AM technology has been hindered by certification, reliability, and repeatability issues. Additively manufactured alloys often exhibit imperfections such as voids and inclusions resulting from their manufacturing process. They also possess unique microstructures, similar to those seen in welded components, when compared to their traditionally manufactured counterparts. In this work we explore microscale effects related both to the unique microstructure of Direct Metal Laser Melted (DMLM) AM metals and to the defects that are invariably present in their manufacture. Using a combination of high-resolution digital image correlation measurements, optical microscopy, and electron backscatter diffraction (EBSD) we identify mechanisms of microscale strain accumulation in AM metals under fracture and fatigue loading conditions, and relate these mechanisms to observed crack growth in the materials. Using an experimentally-informed dual-scale porosity model in 3D finite element simulations we also investigate the influence of defects on crack path selection and compare predicted vs. measured crack paths and strain fields. The 3D numerical simulations also allow us to explore the interaction of the dual-scale voids in the sample interior, and reveal avenues of possible optimal design of AM metallic components. Finally, some recent results in the use of machine learning approaches to bypass explicit microscale modeling of strain accumulation are explored.

For more information, please contact Michael Stramenga by email at [email protected] or visit https://caltech.zoom.us/j/81983472182?pwd=RGE1ZjFCQ0pDT09MTTduL2EyVyt3dz09.