Assessment of non-intrusive sensing in wall-bounded turbulence through explainable deep learning
A. Cremades 1, R. Freibergs 1, S. Hoyas 2, A. Ianiro 3, S. Discetti 3, R. Vinuesa 1 · 2025
Read Assessment of non-intrusive sensing in wall-bounded turbulence through explainable deep learning by A. Cremades 1, R. Freibergs 1, S. Hoyas 2, A. Ianiro 3, S. Discetti 3, R. Vinuesa 1 (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-2502-07610.txt.
Description
Cremades and five co-authors open the black box of a neural network that senses turbulence from a wall, in a 2025 paper. Engineers would love to infer flow away from a surface using only wall sensors, and deep learning can do it, but not explain itself. The team trained a model to predict velocity fluctuations just above a wall from wall pressure and shear stress in a simulated channel. Then they applied an explanation method that scores how much each input point sways the prediction.
The important spots clustered into streak-like patches, tens to a few hundred wall units wide. Those patches showed strong, but not the strongest, pressure and shear swings, typically one to two standard deviations. The authors say knowing which wall features the model trusts helps judge when such sensing can be believed. They frame the work as a step toward explainable tools for flow control.
Catalog record
| CREATOR | A. Cremades 1, R. Freibergs 1, S. Hoyas 2, A. Ianiro 3, S. Discetti 3, R. Vinuesa 1 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-07610