Understanding Bursting Dynamics in Two-dimensional Turbulence Using Deep Convolutional Autoencoders

  • Holey, Joe (University of Cambridge)
  • Page, Jacob (University of Edinburgh)
  • Brenner, Michael (Harvard University)
  • Kerswell, Rich (University of Cambridge)

Please login to view abstract download link

Understanding turbulence is a long-standing problem in the field of fluid dynamics. Here we attempt to further this understanding and classify bursting events in the setting of two-dimensional Kolmogorov flow. To do this we utilise deep convolutional autoencoders and symmetry properties of Kolmogorov flow to construct data-driven reduced order bases in the latent space of the autoencoders.