Surrogate model benchmark for kω-SST RANS turbulence closure coefficients

Philipp Schlichter; Michaela Reck; Jutta Pieringer; Thomas Indinger

doi:10.1016/j.jweia.2024.105678

Surrogate model benchmark for kω-SST RANS turbulence closure coefficients

Philipp Schlichter, Michaela Reck, Jutta Pieringer, Thomas Indinger

Default

Research output: Contribution to journal › Article › peer-review

2 Scopus citations

Abstract

AI-based methods show immense potential to assist engineers in further improving vehicle aerodynamics. It is vital to assess the performance of different surrogate models based on the provided training data size and shape to aid future model selections. This study uses data from the simulated flow around the 2D NACA 8810 airfoil. The closure coefficients of the kω-SST RANS turbulence model are varied by Design of Experiment to achieve the desired amount of varying data for the training and validation datasets. Each dataset uses Principal Component Analysis to generate various levels of dimensional reduction.

Original language	English
Article number	105678
Journal	Journal of Wind Engineering and Industrial Aerodynamics
Volume	246
DOIs	https://doi.org/10.1016/j.jweia.2024.105678
State	Published - Mar 2024

Keywords

Benchmark
Machine learning
Regression methods

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1016/j.jweia.2024.105678

Cite this

@article{844e3621e7b941c9a1e9fdd00c849b1f,

title = "Surrogate model benchmark for kω-SST RANS turbulence closure coefficients",

abstract = "AI-based methods show immense potential to assist engineers in further improving vehicle aerodynamics. It is vital to assess the performance of different surrogate models based on the provided training data size and shape to aid future model selections. This study uses data from the simulated flow around the 2D NACA 8810 airfoil. The closure coefficients of the kω-SST RANS turbulence model are varied by Design of Experiment to achieve the desired amount of varying data for the training and validation datasets. Each dataset uses Principal Component Analysis to generate various levels of dimensional reduction.",

keywords = "Benchmark, Machine learning, Regression methods",

author = "Philipp Schlichter and Michaela Reck and Jutta Pieringer and Thomas Indinger",

note = "Publisher Copyright: {\textcopyright} 2024 The Authors",

year = "2024",

month = mar,

doi = "10.1016/j.jweia.2024.105678",

language = "English",

volume = "246",

journal = "Journal of Wind Engineering and Industrial Aerodynamics",

issn = "0167-6105",

publisher = "Elsevier",

}

TY - JOUR

T1 - Surrogate model benchmark for kω-SST RANS turbulence closure coefficients

AU - Schlichter, Philipp

AU - Reck, Michaela

AU - Pieringer, Jutta

AU - Indinger, Thomas

PY - 2024/3

Y1 - 2024/3

N2 - AI-based methods show immense potential to assist engineers in further improving vehicle aerodynamics. It is vital to assess the performance of different surrogate models based on the provided training data size and shape to aid future model selections. This study uses data from the simulated flow around the 2D NACA 8810 airfoil. The closure coefficients of the kω-SST RANS turbulence model are varied by Design of Experiment to achieve the desired amount of varying data for the training and validation datasets. Each dataset uses Principal Component Analysis to generate various levels of dimensional reduction.

AB - AI-based methods show immense potential to assist engineers in further improving vehicle aerodynamics. It is vital to assess the performance of different surrogate models based on the provided training data size and shape to aid future model selections. This study uses data from the simulated flow around the 2D NACA 8810 airfoil. The closure coefficients of the kω-SST RANS turbulence model are varied by Design of Experiment to achieve the desired amount of varying data for the training and validation datasets. Each dataset uses Principal Component Analysis to generate various levels of dimensional reduction.

KW - Benchmark

KW - Machine learning

KW - Regression methods

UR - http://www.scopus.com/inward/record.url?scp=85185844324&partnerID=8YFLogxK

U2 - 10.1016/j.jweia.2024.105678

DO - 10.1016/j.jweia.2024.105678

M3 - Article

AN - SCOPUS:85185844324

SN - 0167-6105

VL - 246

JO - Journal of Wind Engineering and Industrial Aerodynamics

JF - Journal of Wind Engineering and Industrial Aerodynamics

M1 - 105678

ER -

Surrogate model benchmark for kω-SST RANS turbulence closure coefficients

Abstract

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this