Microstructure-based knowledge systems for capturing process-structure evolution linkages

Area of Science:

• Materials Science
• Computational Science
• Data Science

Background:

• Multiscale simulations are crucial for understanding material behavior.
• Existing frameworks primarily focused on spatial linkages, limiting temporal analysis.
• Capturing the evolution of material structure over space and time remains a challenge.

Purpose of the Study:

• To generalize the Materials Knowledge Systems (MKS) framework for spatiotemporal multiscale simulations.
• To incorporate diverse basis functions for enhanced process-structure-property (PSP) linkage analysis.
• To demonstrate computational advantages over traditional numerical integration methods.

Main Methods:

• Advancement of the MKS data science framework.
• Introduction of generalized basis functions for MKS.
• Application to a Cahn-Hilliard simulation for structure evolution prediction.

Main Results:

• The generalized MKS framework successfully captures spatiotemporal material structure evolution.
• Structure evolution was predicted three orders of magnitude faster than optimized numerical integration.
• Demonstrated the framework's ability to learn embedded physics via influence kernels.

Conclusions:

• The extended MKS framework provides a comprehensive approach for PSP linkages in multiscale materials modeling.
• Offers significant computational advantages, particularly for complex simulations.
• Represents an alternative to differential equations for understanding material physics in simulations.