Surrogate Models for Stochastic Al-Cu Phase-Field Solidification
Abstract
Phase-field models resolve microstructural evolution during rapid alloy solidification, but their computational cost limits long trajectories, stochastic ensembles, and studies across process conditions. We present a variance-preserving denoising diffusion bridge model (VP-DDBM) for state-conditioned prediction of two-dimensional Al-Cu solidification. Given the current phase, concentration, and prescribed temperature fields, the model predicts the subsequent state and is applied autoregressively. Across two held-out thermal gradients, VP-DDBM produces lower long-horizon errors than deterministic U-Net and SI-ConvNeXt baselines in domain-mean phase, domain-mean Cu concentration, and leading-tip position. It also retains fine dendritic structure over long rollouts, whereas the deterministic U-Net progressively smooths the morphology. VP-DDBM predicts the phase and concentration evolution approximately 73× faster than the phase-field solver, showing potential for faster process-condition screening and ensemble-based microstructure studies.
Related
- Response-function-optimized phase field modeling of solute trapping and solute drag in rapid alloy solidification
- Deep-learning surrogate model for rapid quantum-noise computation in dual-pump fiber-optic phase-sensitive amplifiers
- Effective field equations with stochastic initial conditions
- Stochastic inflation as an open quantum system II: open effective field theory and stochastic matching
- Phase-locked beam splitting with continuously tunable intensity ratio using stochastic phase masks
BibTeX
@misc{savukoski2026surrogate,
title = {{Surrogate Models for Stochastic Al-Cu Phase-Field Solidification}},
author = {Simon Savukoski and Tatu Pinomaa},
howpublished = {arXiv},
year = {2026},
month = nov,
doi = {10.48550/arxiv.2611.04695},
eprint = {2611.04695},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2611.04695},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Surrogate Models for Stochastic Al-Cu Phase-Field Solidification
AU - Simon Savukoski
AU - Tatu Pinomaa
PY - 2026
DA - 2026/11/03
PB - arXiv
DO - 10.48550/arxiv.2611.04695
UR - https://arxiv.org/abs/2611.04695
AB - Phase-field models resolve microstructural evolution during rapid alloy solidification, but their computational cost limits long trajectories, stochastic ensembles, and studies across process conditions. We present a variance-preserving denoising diffusion bridge model (VP-DDBM) for state-conditioned prediction of two-dimensional Al-Cu solidification. Given the current phase, concentration, and prescribed temperature fields, the model predicts the subsequent state and is applied autoregressively. Across two held-out thermal gradients, VP-DDBM produces lower long-horizon errors than deterministic U-Net and SI-ConvNeXt baselines in domain-mean phase, domain-mean Cu concentration, and leading-tip position. It also retains fine dendritic structure over long rollouts, whereas the deterministic U-Net progressively smooths the morphology. VP-DDBM predicts the phase and concentration evolution approximately 73× faster than the phase-field solver, showing potential for faster process-condition screening and ensemble-based microstructure studies.
ER -