Acceleration of RANS Solver Convergence via Initialization with Wake Extension Models

Kazuko W. Fuchi, Eric M. Wolf, Christopher R. Schrock, Philip S. Beran · 2025

Read Acceleration of RANS Solver Convergence via Initialization with Wake Extension Models by Kazuko W. Fuchi, Eric M. Wolf, Christopher R. Schrock, Philip S. Beran (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-14699.txt.

Description

Kazuko Fuchi, Eric Wolf, Christopher Schrock, and Philip Beran sped up turbulence simulations in twenty twenty-five by giving the solver a smarter starting point. Reynolds-averaged Navier-Stokes computations converge slowly from a naive uniform start. Their idea: prefill the domain with a modeled solution, splitting it into near-body, wake, and off-body regions.

A convolutional neural network predicts downstream wake development, trained on data from just one simulation. Combined with an accurate near-body prediction, the warm start slashed the needed iterations by a factor of about twenty-six in their test. Wall-clock time fell by a factor of about sixteen versus freestream initialization. The work shows cheap learned models can pay for themselves inside classical solvers.

Catalog record

CREATORKazuko W. Fuchi, Eric M. Wolf, Christopher R. Schrock, Philip S. Beran
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE32 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-14699

Browse the full library · About the Tower of Babel