# Graph model simulates grain growth faster than phase-field calculations

> Researchers at the Hong Kong University of Science and Technology, HKUST Guangzhou and Wuhan University developed a graph-based grain-growth simulation. It updates grain areas using the von Neumann–Mullins law and reconstructs the microstructure with a neural network.

- Canonical URL: https://newsverum.com/en/news/graph-grain-growth-simulation-vn-law-microstructure-2026
- Publisher: Newsverum
- Section: Science
- Country: Hong Kong SAR China
- Published: 2026-10-13T00:51:40+00:00
- Updated: 2026-10-13T00:51:40+00:00
- Event / data period: 2026-10-09
- Other editions: [Türkçe](https://newsverum.com/tr/news/graph-grain-growth-simulation-vn-law-microstructure-2026.md), [العربية](https://newsverum.com/ar/news/graph-grain-growth-simulation-vn-law-microstructure-2026.md), [Français](https://newsverum.com/fr/news/graph-grain-growth-simulation-vn-law-microstructure-2026.md), [Español](https://newsverum.com/es/news/graph-grain-growth-simulation-vn-law-microstructure-2026.md), [Deutsch](https://newsverum.com/de/news/graph-grain-growth-simulation-vn-law-microstructure-2026.md), [Português](https://newsverum.com/pt/news/graph-grain-growth-simulation-vn-law-microstructure-2026.md)

The method, called VN-GraphGG, converts a two-dimensional polycrystalline microstructure into a graph with grain and junction nodes. At each step, grain areas are updated by a physical law tied to the number of sides, then a graph neural network reconstructs the next microstructure field. Separating these stages allows kinetic coefficients to change without retraining the reconstruction component.

On a 6,000-example test set, the authors report a mean pixel-classification error of 3.54% and a mean grain-area error of 2.04 pixels per grain. They also report similar bulk coarsening behaviour to phase-field simulations at two orders of magnitude lower computational cost. These are simulation benchmarks; the paper does not report experimental validation in a real production process.

## Data

### Reported study result

Test examples used by the authors to evaluate reconstruction of a microstructure field from a grain graph.

Methodology: The paper reports a mean pixel-classification error of 3.54% and pixel accuracy of 96.46% on 6,000 graph-field pairs. These are test-set results, not measured prediction accuracy in a real material.

| Reported quantity | Value | Study-specific scope |
|---|---|---|
| Graph-field pairs in the reconstruction test set | 6000 | Test examples used by the authors to evaluate reconstruction of a microstructure field from a grain graph. |

Sources: <https://arxiv.org/abs/2610.13418> · Download data: [JSON](https://newsverum.com/api/v1/articles/graph-grain-growth-simulation-vn-law-microstructure-2026/data/study-result?lang=en&format=json) · [CSV](https://newsverum.com/api/v1/articles/graph-grain-growth-simulation-vn-law-microstructure-2026/data/study-result?lang=en&format=csv)

## Sources

- [arXiv](<https://arxiv.org/abs/2610.13418>) (Source checked: 2026-10-13; Published: 2026-10-09)

---

Cite the canonical URL above. [Editorial standards and corrections](https://newsverum.com/en/editorial)