Parameterization of Forced Isotropic Turbulent Flow using Autoencoders and Generative Adversarial Networks
This addresses the problem of reducing processing time and dependency on high-fidelity CFD simulations for turbulent flow generation in fluid mechanics, though it appears incremental as it builds on existing neural network methods.
The study tackled generating forced isotropic turbulent flow by parameterizing it into basic statistical characteristics and using autoencoders and generative adversarial networks (GANs) trained on pre-simulated data, resulting in flow generation that varies with these parameters and removes dependency on classical mesh-based Navier-Stokes equation estimation.
Autoencoders and generative neural network models have recently gained popularity in fluid mechanics due to their spontaneity and low processing time instead of high fidelity CFD simulations. Auto encoders are used as model order reduction tools in applications of fluid mechanics by compressing input high-dimensional data using an encoder to map the input space into a lower-dimensional latent space. Whereas, generative models such as Variational Auto-encoders (VAEs) and Generative Adversarial Networks (GANs) are proving to be effective in generating solutions to chaotic models with high 'randomness' such as turbulent flows. In this study, forced isotropic turbulence flow is generated by parameterizing into some basic statistical characteristics. The models trained on pre-simulated data from dependencies on these characteristics and the flow generation is then affected by varying these parameters. The latent vectors pushed along the generator models like the decoders and generators contain independent entries which can be used to create different outputs with similar properties. The use of neural network-based architecture removes the need for dependency on the classical mesh-based Navier-Stoke equation estimation which is prominent in many CFD softwares.