AI forecasts grain growth under non-isothermal conditions
Original title: Thermally Conditioned Time-Series Forecasting of Grain Growth in Polycrystalline Materials
Recurrent neural networks conditioned on thermal data predict grain growth with mean relative errors below 5%.
Grain growth affects material properties during heat treatment, but costly simulations limit analysis of non-isothermal cycles. The framework encodes grain size distributions with CNNs, embeds thermal descriptors via MLP, and predicts evolution with RNNs using instantaneous temperature inputs. k-fold cross-validation on front-tracking simulations yields mean relative errors of 2.99% for one-hour cycles and 4.51% for four-hour cycles.
- Trained on wide range of non-isothermal thermal paths
- Mean relative errors 2.99% one-hour and 4.51% four-hour cycles
- Coefficients of determination exceeding 0.99
From the original article
1 min read at arxiv.orgGrain growth during heat treatment strongly affects the microstructure of polycrystalline materials and, consequently, their mechanical properties. Although numerical methods such as front-tracking or front-capturing simulations can accurately describe grain growth, their high computational cost limits their use for long or non-isothermal thermal cycles.
By Eliane Younes · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.