FlowRefiner: Flow Matching-Based Iterative Refinement for 3D Turbulent Flow Simulation
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866913061522112512 |
|---|---|
| author | Dai, Yilong Sun, Yiming Chen, Yiheng Chen, Shengyu Jia, Xiaowei Yu, Runlong |
| author_facet | Dai, Yilong Sun, Yiming Chen, Yiheng Chen, Shengyu Jia, Xiaowei Yu, Runlong |
| contents | Accurate autoregressive prediction of 3D turbulent flows remains challenging for neural PDE solvers, as small errors in fine-scale structures can accumulate rapidly over rollout. In this paper, we propose FlowRefiner, a flow matching-based iterative refinement framework for 3D turbulent flow simulation. The method replaces stochastic denoising refinement with deterministic ODE-based correction, uses a unified velocity-field regression objective across all refinement stages, and introduces a decoupled sigma schedule that fixes the noise range independently of refinement depth. These design choices yield stable and effective refinement in the small-noise regime. Experiments on large-scale 3D turbulence with rich multi-scale structures show that FlowRefiner achieves state-of-the-art autoregressive prediction accuracy and strong physical consistency. Although developed for turbulent flow simulation, the proposed framework is broadly applicable to iterative refinement problems in scientific modeling. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_17149 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | FlowRefiner: Flow Matching-Based Iterative Refinement for 3D Turbulent Flow Simulation Dai, Yilong Sun, Yiming Chen, Yiheng Chen, Shengyu Jia, Xiaowei Yu, Runlong Fluid Dynamics Machine Learning Accurate autoregressive prediction of 3D turbulent flows remains challenging for neural PDE solvers, as small errors in fine-scale structures can accumulate rapidly over rollout. In this paper, we propose FlowRefiner, a flow matching-based iterative refinement framework for 3D turbulent flow simulation. The method replaces stochastic denoising refinement with deterministic ODE-based correction, uses a unified velocity-field regression objective across all refinement stages, and introduces a decoupled sigma schedule that fixes the noise range independently of refinement depth. These design choices yield stable and effective refinement in the small-noise regime. Experiments on large-scale 3D turbulence with rich multi-scale structures show that FlowRefiner achieves state-of-the-art autoregressive prediction accuracy and strong physical consistency. Although developed for turbulent flow simulation, the proposed framework is broadly applicable to iterative refinement problems in scientific modeling. |
| title | FlowRefiner: Flow Matching-Based Iterative Refinement for 3D Turbulent Flow Simulation |
| topic | Fluid Dynamics Machine Learning |
| url | https://arxiv.org/abs/2604.17149 |