Neural Network-Based Parametric Model Reduction for Predicting Turbulent Flow for Different Vehicle Geometries

2026-06-23

AI Summary

High-fidelity aerodynamic simulations for parametric design across multiple vehicle configurations incur prohibitive computational costs, hindering their adoption in industrial workflows. This work proposes a nonlinear reduced-order model that integrates neural networks with a variational autoencoder to efficiently represent high-Reynolds-number turbulent flow fields in a low-dimensional latent space, enabling rapid prediction of flow characteristics for diverse vehicle geometries. The approach leverages the variational autoencoder to enhance reconstruction robustness of complex, multiscale spatiotemporal vortex structures—particularly those in the wake region—and incorporates distributed parallel training with high-resolution flow field data. Experimental results demonstrate that the model significantly reduces computational overhead while accurately recovering key flow features, such as trailing vortical structures.

Abstract

Numerical simulations in industrial applications often require performing numerous high-precision computations parameterized by specific experimental conditions. For instance, in vehicle body design, aerodynamic simulations are essential for evaluating the aerodynamic characteristics of various proposed body geometries. However, computational resource constraints often become a bottleneck. Therefore, achieving the desired accuracy while minimizing computational cost is crucial. To address this challenge, model reduction methods have been developed to decrease the degrees of freedom by constraining the possible states of a physical system to a lower-dimensional subspace. In particular, reduction techniques that project the system onto a nonlinear subspace using neural networks have been actively studied. Our previous research developed a reduced-order model that integrates neural-network-based model reduction with a time-evolution method, implemented as a distributed parallel training framework to process high-resolution flow field data efficiently. In this study, we extend this reduction approach by incorporating a variational autoencoder to assess its robustness in high-Reynolds-number flows around multiple vehicle bodies with varying geometries. Specifically, we evaluate the reconstruction accuracy of vortex generation across different spatial and temporal scales using a compact latent representation, with a particular focus on the flow behavior near the rear end of the vehicle body.

Problem

• turbulent flow
• vehicle geometries
• computational cost
• model reduction
• high-Reynolds-number flows

Innovation

• neural network-based model reduction
• variational autoencoder
• turbulent flow prediction
• parametric reduced-order modeling
• high-Reynolds-number flows