Deep Learning-based Algebraic Reynolds Stress Closures for RANS Simulations of Turbulent Flows
arXiv:2605.26358v1 Announce Type: cross Abstract: Turbulence is ubiquitous in engineering and science, yet direct simulation is prohibitively expensive. The Reynolds-averaged Navier-Stokes (RANS) equations provide savings exceeding ten orders of magnitude but introduce unclosed terms (the closure problem). Offline-trained machine-learning (ML) closures suffer distribution shift in predictive simulations, while ML methods that bypass the governing equations struggle to generalise from scarce high-fidelity data. We develop a physics-derived deep learning closure model for RANS, the Deep Algebrai
This development arises as deep learning techniques mature and are increasingly applied to complex scientific and engineering problems where traditional methods are computationally prohibitive, such as turbulence modeling.
Improving RANS simulations through AI-driven closures can significantly reduce computational costs and enhance accuracy in fields from aerospace engineering to climate modeling, accelerating research and development cycles.
The ability to accurately model turbulent flows with reduced computational overhead via Deep Learning-based closures will enable faster and more reliable simulations, potentially revolutionizing design processes in many industries.
- · Aerospace engineering
- · Automotive industry
- · Climate modeling
- · AI/ML research
- · Computational fluid dynamics software reliant on older closure models
More accurate and faster simulations will lead to optimized designs across various engineering disciplines.
Reduced reliance on expensive physical prototypes due to improved simulation fidelity could accelerate product development cycles and lower costs.
The widespread adoption of such AI-enhanced simulation tools could democratize advanced engineering capabilities, fostering innovation in smaller firms and research groups.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG