Materials

Problem

Prediction of crack growth, fatigue life, and multiscale material behavior.

Why Important

Enables material discovery, reliability engineering, and safety assurance.

Physics / Chemistry / Biology involved

Solid mechanics
Fracture mechanics
Microstructure evolution

Mathematical Model

Phase-field models
Damage and fracture evolution equations
Multiscale constitutive models

AI Method

Graph Neural Networks
Operator learning
Data-assimilation enhanced models

HPC Method

Sparse matrix parallel solvers
Multi-GPU workload partitioning
Mixed precision acceleration

Software Implementation

Target tool: CrackNet (planned)

Benchmark & Validation

Standard crack propagation benchmarks
Experimental fit and generalization tests

Future Tool

A full material performance and failure-warning software pipeline.