Agentic AI-enabled Ditigal and Cyber Manufacturing Systems: Predictive Modeling, Generative Design, Resilience, and Security

We are leading the research on AI agents for (i) predictive modeling of manufacturing processes, (ii) defect detection and classification, and (iii) Adversarial resilience of cyber manufacturing systems.

Mixed stochastic system design

Our goal is to establish the first-of-its-kind design methodology that (i) tailors the structural stochasticity and morphology simultaneously to achieve optimal performances and (ii) enables designing mixed stochasticity structural/microstructural systems. In our vision, this research will also lead to the automation of the nature/bio-inspired design process.

 

Stochastic reconstruction and computational design of microstructures

Our goal is to develop computational methods to enable microscopic image-based statistical characterization, stochastic reconstruction, numerical modeling, and uncertainty quantification of the heterogeneous microstructural materials. We are also interested in discovering the process-microstructure-property relationship for computational material design. Our research has had a significant impact and has been applied to a wide range of microstructural materials.

Computational modeling of microstructural materials in energy storage systems
Our goal is to establish a digital twin of Li-ion battery microstructures to enable performance prediction and inverse optimization.
– Statistical characterization and stochastic reconstruction of digital microstructures to predict mechanical and electrochemical properties;
– Property prediction and inverse design based on machine learning;
– Manufacturing process simulation for electrode material design.
Deep learning-driven design of metamaterial systems considering manufacturing uncertainties
– Deep learning-based discovery and design of cellular metamaterials to achieve superior performances (e.g. elastic wave trapping, impact energy absorption);
– Manufacturing process modeling to understand its impact on material microstructures and properties;
– Uncertainty quantification of the spatially-correlation random quantities in complex topological structures.
Uncertainty quantification, uncertainty propagation, and design for reliability/robustness
The purpose is to provide computational methods (enablers) for the aforementioned research topics.