Surrogate Models for Stochastic Al-Cu Phase-Field Solidification
What it is
This paper introduces a variance-preserving denoising diffusion bridge model (VP-DDBM) to accelerate phase-field simulations of Al-Cu solidification. It predicts microstructural evolution autoregressively from current state and temperature fields, achieving significantly lower long-horizon errors and better retention of dendritic structures compared to deterministic baselines.
The problem it solves
The high computational cost of phase-field models limits the ability to simulate long trajectories, generate stochastic ensembles, and study a wide range of process conditions.
What you could use it for
- Rapid screening of process conditions for alloy solidification
- Generating stochastic ensembles of microstructures for statistical analysis
- Accelerating long-horizon phase-field simulations by a factor of 73
Why it matters
Enables high-throughput study of microstructural evolution and stochastic variability in alloy solidification, which is currently computationally prohibitive with direct phase-field solvers.
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.