Additive Manufacturing (AM) has proven advantages to enable highly complex designs with customized structural and functional properties. However, variabilities in feedstock, machine, process, and environment can lead to inconsistency in part quality. While tremendous efforts have been devoted to thermal-based online monitoring systems, ultrasound-based methods for in situ process monitoring in metals-based AM are still in their infancy.
Sponsored by NIST and Led by Dr. Lang Yuan, the University of South Carolina (USC) collaborates with the University of Illinois at Urbana-Champaign (UIUC), and Purdue University (PU) to develop a hardware and software architecture platform that characterizes the microstructure and defects of metallic parts in real-time using ultrasound non-destructive evaluation (NDE) in the Laser Powder Bed Fusion AM (LPBFAM) process.
This proposed work will primarily focus on the use of linear in situ ultrasound responses and real-time signal processing during the printing process. Ex situ non-linear ultrasound testing (NLUT) will be leveraged throughout the project to establish the basic principles to build up the knowledge base and connect the nonlinear elasticity to nano- and micro-scale defects. This work will set up the foundation for in situ NLUT characterization for micro-cracks, precipitates, grain boundary, and dislocation density in future projects. The successful execution of this work will enhance the understanding of ultrasound NDE on AM-specific microstructure and address the key puzzles in Measurement Science for real-time in situ characterization of metallic parts in LPBFAM.
Project Goals
- An ultrasound NDE system integrated with an LPBFAM machine to enable the acquisition of ultrasound responses and edge computing for data analysis in real-time
- A comprehensive knowledge base using ML to bridge the ultrasound responses to microstructures and defects quantitively
- Physics-based computational models for solidification microstructures and ultrasound propagation to establish a theoretical baseline and a fundamental understanding of ultrasound responses
- A data science tool to deploy the models from the knowledge base onto real-time computing systems to enable the timing deterministic in situ characterization of microstructure