Skip to main content
OSTI.GOVU.S. Department of Energy Office of Scientific and Technical Information
U.S. Department of Energy
Office of Scientific and Technical Information
 

Regularizing Invertible Neural Networks for Airfoil Design Through Dimension Reduction

Conference ·
DOI:https://doi.org/10.2514/6.2022-1098· OSTI ID:1845889
; ; ;

This work examines how dimension reduction can improve the performance of invertible neural networks (INN) for airfoil design. Design workflows are typically expensive, relying on many evaluations of high fidelity computational fluid dynamics (CFD) models. Furthermore, the inverse design problem is typically ill-posed. That is, multiple valid solutions exist that satisfy the design criteria. Regularization can reduce this inverse design space and simplify the problem. We study the use of subspace-based input dimension reduction to act as a regularizer for the INN model and improve the recovery of new airfoil shapes with desired performance characteristics. We find that the dimension reduction identifies two dominant modes, relating to airfoil thickness and camber, that optimally determine the airfoil's aerodynamics. We demonstrate the capability of the proposed INN model to generate 100 airfoils that satisfy the specific aerodynamic and structural characteristics.

Research Organization:
National Renewable Energy Laboratory (NREL), Golden, CO (United States)
Sponsoring Organization:
USDOE Advanced Research Projects Agency - Energy (ARPA-E)
DOE Contract Number:
AC36-08GO28308
OSTI ID:
1845889
Report Number(s):
NREL/CP-2C00-82137; MainId:82910; UUID:9df74d56-0fc9-4d3c-8910-0989de837455; MainAdminID:63876
Resource Relation:
Conference: Presented at the AIAA SCITECH 2022 Forum, 3-7 January 2022
Country of Publication:
United States
Language:
English

References (20)

ECN-G1-21 Airfoil: Design and Wind-Tunnel Testing journal September 2016
An Introduction to Data Analysis and Uncertainty Quantification for Inverse Problems January 2017
Application of Convolutional Neural Network to Predict Airfoil Lift Coefficient conference January 2018
Neural network topology for wind turbine analysis
  • Gaymann, Audrey; Schiaffini, Giorgio; Massini, Michela
  • European Conference on Turbomachinery Fluid Dynamics and Thermodynamics, European Conference on Turbomachinery Fluid Dynamics and hermodynamics https://doi.org/10.29008/ETC2019-174
conference January 2019
Universal Parametric Geometry Representation Method journal January 2008
γ−Reθt¯ Spalart–Allmaras with Crossflow Transition Model Using Hamiltonian–Strand Approach journal May 2019
Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations journal February 2019
Reynolds averaged turbulence modelling using deep neural networks with embedded invariance journal October 2016
Deep learning in fluid dynamics journal January 2017
Deep learning for in situ data compression of large turbulent flow simulations journal November 2020
Adversarial super-resolution of climatological wind and solar data journal July 2020
Multilayer feedforward networks are universal approximators journal January 1989
Universal approximation of an unknown mapping and its derivatives using multilayer feedforward networks journal January 1990
Active Subspaces book January 2015
Inverse regression for ridge recovery: a data-driven approach for parameter reduction in computer experiments journal May 2019
Data-Driven Polynomial Ridge Approximation Using Variable Projection journal January 2018
Hit-and-Run Algorithms for Generating Multivariate Distributions journal May 1993
Definition of a 5-MW Reference Wind Turbine for Offshore System Development report February 2009
Active Subspaces of Airfoil Shape Parameterizations journal May 2018
Data-Based Approach for Fast Airfoil Analysis and Optimization journal February 2019

Similar Records

Related Subjects

WIND ENERGY
airfoil design
dimension reduction
invertible neural networks