High-fidelity grain growth modeling: Leveraging deep learning for fast computations

Grain growth simulation is crucial for predicting metallic material microstructure evolution during annealing and resulting final mechanical properties, but traditional partial differential equation-based methods are computationally expensive, creating bottlenecks in materials design and manufacturing. In this work, we introduce a machine learning framework that combines a Convolutional Long Short-Term Memory networks with an Autoencoder to efficiently predict grain growth evolution. Our approach captures both spatial and temporal aspects of grain evolution while encoding high-dimensional grain structure data into a compact latent space for pattern learning, enhanced by a novel composite loss function combining Mean Squared Error, Structural Similarity Index Measurement, and Boundary Preservation to maintain structural integrity of grain boundary topology of the prediction. Results demonstrated that our machine learning approach accelerates grain growth prediction by up to 89 × faster, reducing computation time from 10 min to approximately 10 s while maintaining high-fidelity predictions. The best model (S-30-30) achieving a structural similarity score of 86.71 % and mean grain size error of just 0.07 %. All models accurately captured grain boundary topology, morphology, and size distributions. This approach enables rapid microstructural prediction for applications where conventional simulations are prohibitively time-consuming, potentially accelerating innovation in materials science and manufacturing. • This work modeled grain growth evolution as a spatio-temporal task in machine learning framework and developed a hybrid neural networks that combine an Autoencoder and a Convolutional Long Short-Term Memory for predicting grain growth evolution. • A custom adaptive weight loss function combining Mean Squared Error, Structural Similarity Index Measurement, and Boundary Preservation is introduced to maintain structural integrity of grain boundary topology. • The developed machine learning framework successfully accelerates the prediction of grain growth evolution up to 89 × faster compared to conventional simulation that based on partial-differential equations, while maintaining high-fidelity. • Computation time reduced from approximately 10 min to around 10 s. • Grain boundary topology, morphology and size distributions are accurately captured across all models. • Best model achieved 86.71 % structural similarity score with only 0.07 % error in mean grain size.