Skip to Content

College of Engineering and Computing

  • Mechanical Engineering Associate Professor Lang Yuan

Utilizing microstructure modeling for eliminating defects

Lang Yuan’s research aims to produce quality parts in 3D printing for functionally graded materials

“What is additive manufacturing?” 3D printing is the simple answer. By creating alloys and metal prints with 3D printing, the metal additive manufacturing process creates parts and prototypes while improving design, production and supply chain. 

Mechanical Engineering Associate Professor Lang Yuan’s lab at the University of South Carolina’s McNair Center focuses on developing new materials and processes through additive manufacturing, assisted by advanced computational modeling. For the last two years, he has been working on a collaborative project to develop a machine learning enhanced understanding of functionally gradient materials (FGMs) enabled by directed energy deposition. 

Yuan’s project is part of a nearly $5 million grant funded award from the Department of Energy and expected to be completed this coming February. He is partnering with Edison Welding Institute (EWI) and General Electric Research to optimize the directed energy deposition process for a new nickel-based superalloy FGM. 

FGMs are innovative materials in which their final properties vary by dimensions. They are tailored to satisfy perfectly to the working environment, maintaining their strengths and eliminating weaknesses. Direct energy deposition is a category of the additive manufacturing process that forms 3D objects by melting material as it is being deposited using focused thermal energy, such as a laser beam. 

“This project is aligned with the kind of work we're doing in my lab. Our work has been focusing on the computational side, developing a unique set of tools to predict microstructure and defects during the rapid solidification process,” Yuan says. 

Yuan’s research investigated the microstructure variation as alloy composition transition from one alloy to another during printing by evaluating the process parameters and ensuring satisfactory material structure (microstructure) and the ability to prevent defects. The project began in September 2020, and the first stage focused on utilizing machine learning and physics-based modeling to drive alloy design and process optimization, including printing and heat treatment, while demonstrating successful FGMs for mechanical tasks.

Our model is helping the team to build up a fundamental understanding of why hot cracks form and their connections to the process condition and microstructure.

- Associate Professor Lang Yuan

Since the solidification microstructures are determined by process parameters and alloy composition, Yuan’s team took the Integrated Computational Materials Engineering approach and extended their unique high performance microstructure code to take inputs from thermodynamics modeling for alloy composition variation. The directed energy deposition process was utilized for modeling of transient temperature distribution. This approach effectively evaluates the grain size and element segregation in both monolithic alloys and the FGMs. 

The project is currently focused on predicting and controlling defects, such as cracking, which are harmful to material properties. Yuan’s team has been examining key parameters that drive the formation of defects through microstructure models. Machine learning, supported by microstructure characteristics and experiments, will also be applied for predicting and controlling defects. 

“The second period is translating our work into an industrial application. We’re trying to print a demo part that can be tested or later used in the actual application,” Yuan says. 

Yuan’s post-doctoral researcher Hamed Seyyedhosseinzadeh is working on predicting cracking in this specific alloy.

“It’s a physics-based computational model and method to correctly predict defects and trends. We have at least three different super alloys that we are analyzing the cross sections and measuring crack density. A couple of our numerical and analytical methods have produced the correct density so far,” Seyyedhosseinzadeh says.

Yuan’s team aims to demonstrate cost reduction by 10%-20% and durability improvement by 20%-30%. The nickel-based super alloys will be designed, printed and evaluated for fatigue, strength and corrosion resistance. Direct energy deposition process microstructure modeling, wrapped around machine learning, will also be e-applied to support successful printing. Led by EWI, the team is currently preparing for a technology demonstration for a typical engine case, which includes the design, direct energy deposition process simulations for stress and distortions, case printing, and post-treatment and examination. Plans will also be developed for commercializing the technology.

“The concept of FGM was introduced in the 1980s, but the alloy design and fabrication have been challenging due to the complication of varying alloy composition,” Yuan says. “In the project, we targeted the hot-gas path (HGP) in gas turbines to demonstrate the technology and benefits. One key outcome is that FGMs allow for cheaper materials at non-critical sections and provide non-abrupt interfaces that reduce stresses, thus having a great potential to reduce cost and improve durability for HGP parts.”

Strengthened nickel-base super alloys, such as Rene 41 and Rene 80, commonly reveal hot cracks. 

“Our model is helping the team to build up a fundamental understanding of why hot cracks form and their connections to the process condition and microstructure. The final goal is to accurately predict the conditions that will avoid defects with a desired grain size distribution and accelerate the material process development,” Yuan says.

Yuan is excited about the project’s progress, especially his team’s effort and high productivity working with world-leading industrial partners. 

“It is fortunate to work with such capability teams and participate in such a holistic effort in combined modeling and experiments to drive the development of FGM,” Yuan says. “This effort brings together essential materials science and additive manufacturing knowledge and physics-based computational model and machine learning tools from different scales. Everything is integrated to find the right process and material.”

Challenge the conventional. Create the exceptional. No Limits.