Replication data for : Autoencoder-Based Dimensionality Reduction of Turbulent Channel Flow Under Spanwise Wall Oscillations

The present dataset contains 2D slices of a turbulent channel flow under spanwise wall oscillations, obtained from simulations with varying actuation amplitudes. The dataset features velocity and temperature fields (u, v, w, and Θ) as well as corresponding latent variables obtained through various dimensionality reduction techniques. The dimensionless numbers for this case are Reτ = 200 and Pr = 1. The dataset includes ground-truth data, reconstructed fields, and latent variables for multiple models: Convolutional AutoEncoder (CAE), β-Variational AutoEncoder (VAE), and extended Proper Orthogonal Decomposition (POD). For a detailed description of the dataset structure, file naming conventions, and instructions on how to read and process the data, please refer to the accompanying README file. For a full description of the configuration and methodology, please check the corresponding paper: "Autoencoder-Based Dimensionality Reduction of Turbulent Channel Flow Under Spanwise Wall Oscillations" (not yet published).