Pretrained Video Models as Differentiable Physics Simulators for Urban Wind Flows

TL;DR

This paper presents WinDiNet, a pretrained video diffusion model that acts as a fast, differentiable surrogate for urban wind flow CFD simulations, enabling efficient design optimization.

Contribution

It introduces a novel neural surrogate model trained on CFD data, outperforming traditional solvers and enabling gradient-based urban layout optimization.

Findings

WinDiNet generates 112-frame wind flow simulations in under a second.

The model outperforms purpose-built neural PDE solvers in accuracy.

Gradient-based optimization with WinDiNet improves urban wind safety and comfort.

Abstract

Designing urban spaces that provide pedestrian wind comfort and safety requires time-resolved Computational Fluid Dynamics (CFD) simulations, but their current computational cost makes extensive design exploration impractical. We introduce WinDiNet (Wind Diffusion Network), a pretrained video diffusion model that is repurposed as a fast, differentiable surrogate for this task. Starting from LTX-Video, a 2B-parameter latent video transformer, we fine-tune on 10,000 2D incompressible CFD simulations over procedurally generated building layouts. A systematic study of training regimes, conditioning mechanisms, and VAE adaptation strategies, including a physics-informed decoder loss, identifies a configuration that outperforms purpose-built neural PDE solvers. The resulting model generates full 112-frame rollouts in under a second. As the surrogate is end-to-end differentiable, it doubles as a physics simulator for gradient-based inverse optimization: given an urban footprint layout, we optimize building positions directly through backpropagation to improve wind safety as well as pedestrian wind comfort. Experiments on single- and multi-inlet layouts show that the optimizer discovers effective layouts even under challenging multi-objective configurations, with all improvements confirmed by ground-truth CFD simulations.