Anchored-Branched Steady-state WInd Flow Transformer (AB-SWIFT): a metamodel for 3D atmospheric flow in urban environments

2026-03-26

AI Summary

This work addresses the challenge of accurately modeling three-dimensional steady-state wind fields in urban environments, where complex geometries and multi-scale atmospheric stability conditions pose significant difficulties for conventional deep learning approaches. To this end, the authors propose AB-SWIFT, an anchored-branch Transformer-based surrogate model specifically designed for atmospheric flow modeling. AB-SWIFT employs an embedded branching architecture to efficiently integrate urban geometric features with atmospheric stratification information. The model is trained on a high-resolution computational fluid dynamics (CFD) dataset comprising randomized urban scenarios under neutral, unstable, and stable atmospheric conditions. Experimental results demonstrate that AB-SWIFT consistently outperforms existing Transformer and graph neural network methods across all flow field prediction metrics, achieving state-of-the-art accuracy and generalization capability.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Planning, Routing, and Scheduling: Planning with Language Models
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Air flow modeling at a local scale is essential for applications such as pollutant dispersion modeling or wind farm modeling. To circumvent costly Computational Fluid Dynamics (CFD) computations, deep learning surrogate models have recently emerged as promising alternatives. However, in the context of urban air flow, deep learning models struggle to adapt to the high variations of the urban geometry and to large mesh sizes. To tackle these challenges, we introduce Anchored Branched Steady-state WInd Flow Transformer (AB-SWIFT), a transformer-based model with an internal branched structure uniquely designed for atmospheric flow modeling. We train our model on a specially designed database of atmospheric simulations around randomised urban geometries and with a mixture of unstable, neutral, and stable atmospheric stratifications. Our model reaches the best accuracy on all predicted fields compared to state-of-the-art transformers and graph-based models. Our code and data is available at https://github.com/cerea-daml/abswift.

Problem

• urban airflow modeling
• computational fluid dynamics
• deep learning surrogate models
• atmospheric flow
• urban geometry

Innovation

• Transformer-based model
• urban airflow modeling
• deep learning surrogate
• branched architecture
• atmospheric stratification