English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

With the recent rapid developments in machine learning (ML), several attemptshave been made to apply ML methods to various fluid dynamics problems. However,the feasibility of ML for predicting turbulence dynamics has not yet beenexplored in detail. In this study, PredictionNet, a data-driven ML frameworkbased on generative adversarial networks (GANs), was developed to predicttwo-dimensional (2D) decaying turbulence. The developed prediction modelaccurately predicted turbulent fields at a finite lead time of up to half theEulerian integral time scale. In addition to the high accuracy in pointwisemetrics, various turbulence statistics, such as the probability densityfunction, spatial correlation function, and enstrophy spectrum, were accuratelycaptured by the employed GAN. Scale decomposition was used to interpret thepredictability depending on the spatial scale, and the role of latent variablesin the discriminator network was investigated. The good performance of the GANin predicting small-scale turbulence is attributed to the scale-selectioncapability of the latent variable. Results also revealed that the recursiveapplications of the prediction model yielded better predictions than singlepredictions for large lead times. Furthermore, by utilizing PredictionNet as asurrogate model, a control model named ControlNet was developed to identifydisturbance fields that drive the time evolution of the flow field in thedirection that optimises the specified objective function. Therefore, anillustrative example in which the evolution of 2D turbulence can be predictedwithin a finite time horizon and controlled using a GAN-based deep neuralnetwork is presented.

Prediction and control of two-dimensional decaying turbulence using generative adversarial networks