Materials Frontier 2025 ISSUE 22(Total ISSUE 129)
July 10, 2025 15:00 ~ 16:30

Third-Generation Pandat Software: A High-Performance Materials Data Engine for ICME and AI

Guest SpeakerVice President Weisheng CaoCompuTherm LLCUSA

Inviter: Prof. Liming Peng

Date&Time: Thursdayy, 10 July, 15:00-16:30

Venue: Yiucheng Lecture Hall (500), Xu Zuyao Building

 

Biography:

 Dr. Weisheng Cao earned his Ph.D. from the Univer Vice President of CompuTherm LLC sity of Wisconsin–Madison in 2006. He is the  and leads Pandat software design and development. His recent work focuses on integrating CALPHAD-based modeling with large-scale ICME and AI applications. Dr. Cao is passionate about bridging cutting-edge research with industrial innovation and looks forward to sharing Pandat’s latest progress and exchanging ideas with the community.

 Abstract:

Pandat, developed based on the CALPHAD method, has evolved through three generations. The first generation laid the foundation by focusing on thermodynamic and phase diagram calculations rooted in the CALPHAD methodology. The second generation of Pandat has been re-designed as an integrated platform by coupling CALPHAD with kinetic simulations, enabling predictions of microstructure evolution during processes such as precipitation, diffusion, and solidification. The third-generation advances toward large-scale Integrated Computational Materials Engineering (ICME) and AI model integration, aiming to build a high-performance materials data engine for both research and industrial applications.

 

Key to this advancement are the new modules PanLink and PanDataNet, which overcome the computational limitations of traditional direct-coupling approaches. By dynamically grouping, managing and reusing calculation data, they enable thermodynamic/kinetic simulations at billion-scale speeds. When coupled with PanEvolution and PanSolidification, the system supports large-scale FEM simulations of heat treatment and casting processes. Several case studies will be presented to demonstrate these capabilities, along with the integration of Pandat with AI models for accelerated materials discovery. These advancements position Pandat as a leading solution for seamless integration with FEM software and AI-driven platforms.