Technology Category
Planning, Routing, and Scheduling: Optimization of Spatio-temporal SystemsSearch and Optimization: Local SearchMachine Learning: Feature Construction/Reformulation
Data-efficient rapid prediction of urban airflow and temperature fields for complex building geometries
🤖 AI Summary
To address the high computational cost of CFD simulations, poor generalization of data-driven methods, and difficulty in jointly modeling local and far-field aerodynamic dependencies in urban microclimate modeling, this paper proposes a lightweight, real-time 3D wind-temperature field prediction framework requiring only building geometry as input. Our key contributions are: (1) a novel Multi-Directional Distance Feature (MDDF) that explicitly encodes buildings’ multi-directional influence on airflow; and (2) a Localized Fourier Neural Operator (Local-FNO) coupled with a CFD-driven dimensionality-reduction data generation strategy, enabling strong generalization from only 24 CFD simulations. Experiments demonstrate that our method achieves prediction times under one minute—500× faster than CFD—while attaining mean absolute errors of 0.3 m/s for wind speed and 0.3°C for temperature. Notably, it significantly overcomes the long-standing challenge of far-field wake modeling in urban flow prediction.
Technology Category
Application Category
📝 Abstract
Accurately predicting urban microclimate, including wind speed and temperature, based solely on building geometry requires capturing complex interactions between buildings and airflow, particularly long-range wake effects influenced by directional geometry. Traditional methods relying on computational fluid dynamics (CFD) are prohibitively expensive for large-scale simulations, while data-driven approaches struggle with limited training data and the need to model both local and far-field dependencies. In response, we propose a novel framework that leverages a multi-directional distance feature (MDDF) combined with localized training to achieve effective wind field predictions with minimal CFD data. By reducing the problem's dimensionality, localized training effectively increases the number of training samples, while MDDF encodes the surrounding geometric information to accurately model wake dynamics and flow redirection. Trained on only 24 CFD simulations, our localized Fourier neural operator (Local-FNO) model generates full 3D wind velocity and temperature predictions in under one minute, yielding a 500-fold speedup over conventional CFD methods. With mean absolute errors of 0.3 m/s for wind speed and 0.3 $^{circ}$C for temperature on unseen urban configurations, our method demonstrates strong generalization capabilities and significant potential for practical urban applications.
Problem
Research questions and friction points this paper is trying to address.
Predict urban microclimate using building geometry efficiently
Overcome high computational costs of traditional CFD methods
Address limited training data in data-driven approaches
Innovation
Methods, ideas, or system contributions that make the work stand out.
Multi-directional distance feature for geometric encoding
Localized training to increase sample efficiency
Fourier neural operator for rapid 3D predictions
🔎 Similar Papers
No similar papers found.
