Kolmogorov flow test case

该数据集包含了受不可压缩纳维耶-斯托克斯方程控制的强迫湍流流动——科尔莫戈罗夫流的模拟轨迹。这些高保真度的模拟在一个2048×2048的网格上进行，并被下采样至128×128的网格以生成滤波后的直接数值模拟（DNS）解，这些解作为比较的基准真值。数据集同时包含了高保真度模拟和下采样轨迹，用于评估各种生成模型在长时间范围内准确模拟湍流流动的性能。模拟的规模从2048×2048的高保真度网格下采样至128×128，任务涉及湍流模拟和概率预测。

This dataset contains simulated trajectories of forced turbulent flow (Kolmogorov flow) governed by the incompressible Navier-Stokes equations. These high-fidelity simulations are conducted on a 2048×2048 grid, then downsampled to a 128×128 grid to generate filtered direct numerical simulation (DNS) solutions, which serve as the ground truth for comparison. The dataset includes both high-fidelity simulations and downsampled trajectories, which are used to evaluate the performance of various generative models in accurately simulating turbulent flows over long-time horizons. The simulations are downsampled from the 2048×2048 high-fidelity grid to 128×128, and the tasks involve turbulent flow simulation and probabilistic forecasting.