Using diffusion models for reducing spatiotemporal errors of deep learning based urban microclimate predictions at post-processing stage
- 1 March 2025
- journal article
- Published by AIP Publishing in Physics of Fluids
- Vol. 37 (3)
Abstract
Computational fluid dynamics (CFD) is a powerful tool for modeling turbulent flow and is commonly used for urban microclimate simulations. However, traditional CFD methods are computationally intensive, requiring substantial hardware resources for high-fidelity simulations. Deep learning (DL) models are becoming popular as efficient alternatives, requiring less computational resources to model complex non-linear interactions in fluid flow simulations. A major drawback of DL models is that they are prone to error accumulation in long-term temporal predictions, often compromising their accuracy and reliability. To address this shortcoming, this study investigates the use of a denoising diffusion probabilistic model (DDPM) as a novel post-processing technique to mitigate error propagation in DL models' sequential predictions. To address this, we employ convolutional autoencoder (CAE) and U-Net architectures to predict airflow dynamics around a cubic structure. The DDPM is then applied to the model's predictions, refining the reconstructed flow fields to better align with high-fidelity statistical results from large-eddy simulations. Results demonstrate that, although deep learning models provide significant computational advantages over traditional numerical solvers, they are susceptible to error accumulation in sequential predictions; however, utilizing DDPM as a post-processing step enhances the accuracy of DL models by up to 65% while maintaining a three times speedup compared to traditional numerical solvers. These findings highlight the potential of integrating denoising diffusion probabilistic models as a transformative approach to improving the reliability and accuracy of deep learning-based urban microclimate simulations, paving the way for more efficient and scalable fluid dynamics modeling.Keywords
All Related Versions
Funding Information
- Consortium de Recherche et d'Innovation en Aérospatiale au Québec (051811)
- Natural Sciences and Engineering Research Council of Canada (RGPIN-2024- 06297)
- Canada First Research Excellence Fund (Volt-Age)
This publication has 77 references indexed in Scilit:
- A Note on Friction Velocity and Viscous Effect for Idealized Urban Canopy FlowsBoundary-Layer Meteorology, 2023
- Vortex-U-Net: An efficient and effective vortex detection approach based on U-Net structureApplied Soft Computing, 2022
- The budgets of turbulence kinetic energy and heat in the urban roughness sublayerEnvironmental Fluid Mechanics, 2021
- An introduction to deep generative modelingGAMM-Mitteilungen, 2021
- Turbulence-Kinetic-Energy Budget in the Urban-Like Boundary Layer Using Large-Eddy SimulationBoundary-Layer Meteorology, 2020
- Machine-learning-based reduced-order modeling for unsteady flows around bluff bodies of various shapesTheoretical and Computational Fluid Dynamics, 2020
- Modeling building resilience against extreme weather by integrated CityFFD and CityBEM simulationsApplied Energy, 2019
- A high‐order backward forward sweep interpolating algorithm for semi‐Lagrangian methodInternational Journal for Numerical Methods in Fluids, 2017
- U-Net: Convolutional Networks for Biomedical Image Segmentation2015
- Recent advances on the numerical modelling of turbulent flowsApplied Mathematical Modelling, 2015