Generalization capabilities of conditional GAN for turbulent flow under changes of geometry

TL;DR

This paper evaluates how well a conditional GAN can generate realistic turbulent flow fields around turbine components when the flow configuration changes, aiming to reduce computational costs compared to traditional simulation methods.

Contribution

It demonstrates the potential and limitations of using conditional GANs for generalizing turbulent flow simulations across different geometries, based on LES training data.

Findings

GAN can generalize to unseen wake positions with increasing accuracy

Statistical properties of generated flows closely match LES results for certain configurations

Limitations observed in generalization when geometric changes are large or complex

Abstract

Turbulent flow consists of structures with a wide range of spatial and temporal scales which are hard to resolve numerically. Classical numerical methods as the Large Eddy Simulation (LES) are able to capture fine details of turbulent structures but come at high computational cost. Applying generative adversarial networks (GAN) for the synthetic modeling of turbulence is a mathematically well-founded approach to overcome this issue. In this work, we investigate the generalization capabilites of GAN-based synthetic turbulence generators when geometrical changes occur in the flow configuration (e.g. aerodynamic geometric optimization of structures such as airfoils). As training data, we use the flow around a low-pressure turbine (LPT) stator with periodic wake impact obtained from highly resolved LES. To simulate the flow around a LPT stator, we use the conditional deep convolutional GAN framework pix2pixHD conditioned on the position of a rotating wake in front of the stator. For the generalization experiments we exclude images of wake positions located at certain regions from the training data and use the unseen data for testing. We show the abilities and limits of generalization for the conditional GAN by extending the regions of the extracted wake positions successively. Finally, we evaluate the statistical properties of the synthesized flow field by comparison with the corresponding LES results.