# GrainGNN: A dynamic graph neural network for predicting 3D grain microstructure

arXiv: https://arxiv.org/abs/2401.03661 · HTML: https://mlsanity.com/paper/2401.03661/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 8 Jan 2024.

## Built on by (most cited first, 8 of 8)

- [Learning Metal Microstructural Heterogeneity through Spatial Mapping of Diffraction Latent Space Features](https://mlsanity.com/paper/2501.18064/) (2025)
- [Scaling Kinetic Monte-Carlo Simulations of Grain Growth with Combined Convolutional and Graph Neural Networks](https://mlsanity.com/paper/2511.17848/) (2025)
- [Fully Convolutional Spatiotemporal Learning for Microstructure Evolution Prediction](https://mlsanity.com/paper/2602.19915/) (2026)
- [Computational, Data-Driven, and Physics-Informed Machine Learning Approaches for Microstructure Modeling in Metal Additive Manufacturing](https://mlsanity.com/paper/2505.01424/) (2025)
- [Physics-Informed GCN-LSTM Framework for Long-Term Forecasting of 2D and 3D Microstructure Evolution](https://mlsanity.com/paper/2509.15029/) (2025)
- [Physics-Guided Fully Convolutional Spatiotemporal Learning Toward Digital-Twin-Enabled Microstructure Evolution Prediction](https://mlsanity.com/paper/2606.20983/) (2026)
- [Machine-learning-enabled interpretation of tribological deformation patterns in large-scale MD data](https://mlsanity.com/paper/2512.05818/) (2025)
- [Efficient and Scalable Physics-Guided Fully Convolutional Spatiotemporal Learning for 3D Microstructure Evolution Prediction](https://mlsanity.com/paper/2609.36504/) (2026)

## Abstract (arXiv)

We propose GrainGNN, a surrogate model for the evolution of polycrystalline grain structure under rapid solidification conditions in metal additive manufacturing. High fidelity simulations of solidification microstructures are typically performed using multicomponent partial differential equations (PDEs) with moving interfaces. The inherent randomness of the PDE initial conditions (grain seeds) necessitates ensemble simulations to predict microstructure statistics, e.g., grain size, aspect ratio, and crystallographic orientation. Currently such ensemble simulations are prohibitively expensive and surrogates are necessary. In GrainGNN, we use a dynamic graph to represent interface motion and topological changes due to grain coarsening. We use a reduced representation of the microstructure using hand-crafted features; we combine pattern finding and altering graph algorithms with two neural networks, a classifier (for topological changes) and a regressor (for interface motion). Both networks have an encoder-decoder architecture; the encoder has a multi-layer transformer long-short-term-memory architecture; the decoder is a single layer perceptron. We evaluate GrainGNN by comparing it to high-fidelity phase field simulations for in-distribution and out-of-distribution grain configurations for solidification under laser power bed fusion conditions. GrainGNN results in 80%--90% pointwise accuracy; and nearly identical distributions of scalar quantities of interest (QoI) between phase field and GrainGNN simulations compared using Kolmogorov-Smirnov test. GrainGNN's inference speedup (PyTorch on single x86 CPU) over a high-fidelity phase field simulation (CUDA on a single NVIDIA A100 GPU) is 150×--2000× for 100-initial grain problem. Further, using GrainGNN, we model the formation of 11,600 grains in 220 seconds on a single CPU core.