Reducing Defects in 3D Metal Printing by Scientific Machine Learning
Qizhi He in his office

New Assistant Professor Qizhi He received a U of M Informatics Institute (UMII) Seed Grant to develop an AI computational tool to improve 3D metal printing. He is working in partnership with Assistant Professor Ju Sun from Computer Science and Engineering and the Minnesota Supercomputing Institute (MSI).

3D printing, more technically known as Additive Manufacturing (AM), shows great promise. It allows greater design freedom and less material waste than conventional manufacturing methods. However, the process can induce defects that affect the quality of the final products. In order to expand AM for large-scale manufacturing, these process-induced defects must be eliminated.

Qizhi He and Ju Sun are developing a physics aware machine learning framework to help predict defects in metal AM. Their novel, knowledge-augmented, machine learning tool will quickly and reliably predict thermal mechanical behavior and the induced defects by utilizing both thermomechanical models and process monitoring data. Their project focuses on a type of AM technology called laser powder bed fusion (LPBF). LPBF has been used in industries for a wide spectrum of materials including metals, polymers, and ceramics. He and Sun’s research will advance understanding of the process-structure-properties relation and hidden defect mechanisms in metal AM. Their research will also promote the application of AI technology and information science to real-time data assimilation for extreme manufacturing conditions.

Their UMII Seed Grant project is called “Multiphysics Data Assimilation Framework Based on Process-Aware Neural Operator for Failure Prediction in Additive Manufacturing.” UMII Seed Grant funds promote, catalyze, accelerate, and advance UMN-based informatics research in areas related to the MnDRIVE initiative. This Seed Grant falls under the Robotics, Sensors, and Advanced Manufacturing research area of the MnDRIVE initiative.  

Qizhi He works at the intersection of computational mechanics, materials modeling, and scientific machine learning. His research focuses on developing computational tools that synergize mechanistic, data-driven, and artificial intelligence approaches to modeling the mechanics of complex multiphysical processes. The goal of his research is to improve fundamental understanding of multiscale materials and structures, and advance the predictive simulation and design capabilities for complex engineered and natural systems. Qizhi He will be teaching a 5000-level course in Applied Structural Mechanics fall semester 2022.

Share