Simulating The Urban Canopy's Impact on Wind-Driven Natural Ventilation
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.
Kai: We're looking at the paper: "Simulating The Urban Canopy's Impact on Wind-Driven Natural Ventilation".
Mira: Here's the gist of the paper. The urban canopy affects wind in complex ways, making it challenging to predict wind-driven natural ventilation and cooling in buildings Using large eddy simulations of coupled outdoor and indoor airflow, we study how the surrounding urban canopy and wind angle influence ventilation rates through four ventilation configurations: cross, corner, dual-room, and single-sided Flow visualizations demonstrate how both large-scale flow patterns and local interference effects can influence ventilation rates by 50-85% In general, lower density canopies give higher ventilation rates and wind angles that align with a direct path between two openings also lead to higher ventilation rates However, interference effects from surrounding buildings can significantly change the local wind speed and direction, thus also changing ventilation rates The magnitude of these interference effects depends on both the wind angle and surrounding building geometries Simulation Setup and Model Description The study employs large eddy simulations (LESs) to predict coupled outdoor and indoor airflow through specific canopy and building geometries The computational domain includes 140 buildings distributed over four quadrants, with each quadrant containing an identical group of 35 houses forming a five-by-seven staggered grid Two domains representing different canopy densities are considered, with house centers being 50% further apart in the low density domain compared to the high density domain The building interiors include four representative ventilation configurations: cross, corner, dual-room, and single-sided The mesh resolution varies across the domain, extending for six elements around interior and exterior walls for houses with simulated interiors Flow Field Analysis Domain-scale flow patterns reveal that lower density domains support a faster in-canopy flow, consistent with smaller frontal area fractions (lambda f) The aligned quadrants, which have a smaller lambda f, support an overall faster in-canopy flow with building wakes confined to the smaller side yards Diagonal flow cases also reveal domain-scale variability in wind speed stemming from the anisotropic nature of the drag exerted by a quadrant on the flow Ventilation Rate Sensitivity The dual-room configuration experiences the largest median ventilation rate at 0.22, followed by the cross (10% less), corner (35% less), and single-sided (85% less) ventilation types On average, the higher density canopy gives 15% lower ventilation rates and 40% lower IQRs compared to the low density canopy Conditioning on both density and wind angle provides a significant decrease in IQRn for multi-window ventilation types, showing that natural ventilation models should represent the combined effect of these two parameters Implications for Modeling The results indicate that existing models do not accurately capture the interference effects that arise in real urban canopies, as changes in flow rates depend on both the wind angle and surrounding building geometries To improve accuracy, natural ventilation models must represent both the local and larger-scale effects of a specific urban canopy geometry Performing large-scale LESs for every building designed to use natural ventilation will be out of reach for the foreseeable future The LES data presented can provide a starting point for developing such models This study reveals a nuanced interaction between urban canopy geometry and wind angle, which combine to determine the effectiveness of different natural ventilation configurations The importance of the specific configuration of surrounding buildings was further highlighted by considering the ventilation rates obtained in houses near quadrant edges, where unique interference effects alter the ventilation rate by up to 60% compared to houses in quadrant centers Future work should focus on developing new models that can better represent these geometry-specific canopy effects to accurately assess natural ventilation and cooling at larger scales The LES data presented in this paper can provide a starting point for developing such models The authors acknowledge that the results reinforce the effects observed in the flow field visualizations, showing that interference effects from surrounding buildings can play an equally important role in determining ventilation rates for specific wind angles The dependency of these local and large-scale effects on specific urban geometry poses a significant modeling challenge given the diverse geometrical configurations in realistic urban areas The LES data presented in this paper can provide a starting point for developing such models The authors thank Stanford University and the Stanford Research Computing Center for providing computational resources and support that contributed to these research results The raw data are available at https://doi.org/10.25740/vb900bf9392 and from the corresponding author (N.B.) The paper is 2508.04091 The authors declare no conflict of interest N.B.
Kai: performed all simulations and data analysis C.G.
Mira: and H.S.
Kai: contributed guidance and advice during the research process The paper is 17 The authors declare no conflict of interest N.B.
Paper summary: Kai: Building on that, this paper investigates how different configurations—cross, corner, dual-room, and single-sided—affect ventilation rates under these conditions. They use large eddy simulations of coupled outdoor and indoor airflow to study this.
Mira: The study looks at how these canopy geometries influence the results across those four setups. It’s important because existing models often rely on one-dimensional airflow models, which don't account for this complexity well when you look at diverse urban areas.
Lev: When we think about running this on real hardware, the simulation setup involves one hundred forty buildings spread over four quadrants, with each quadrant having thirty-five houses in a staggered grid. They are comparing low density versus high density domains to see how that affects things.
Kai: They specifically looked at how the house centers are spaced—the centers being fifty percent further apart in the low density domain compared to the high density domain.
Mira: And they found that domain-scale flow patterns show that lower density domains support a faster airflow within the canopy, which makes sense because of those smaller frontal area fractions.
Lev: They also observed that aligned quadrants, which have a smaller lambda f value, support an overall faster in-canopy flow with building wakes confined to the smaller side yards.
Kai: But they also saw domain-scale variability in wind speed coming from how anisotropic the drag is exerted by a quadrant on the flow, especially in diagonal flow cases.
Mira: Now let's look at what they actually measured regarding ventilation rates. They found that the dual-room configuration had the largest median ventilation rate at zero point two two, followed by the cross configuration with ten percent less, corner with thirty-five percent less, and single-sided with eighty-five percent less.
Lev: And when you look at density alone, they found that the higher density canopy gives fifteen percent lower ventilation rates and forty percent lower IQRs compared to the low density canopy.
Kai: What’s interesting is that conditioning on both the density and the wind angle provides a significant decrease in IQRn for multi-window ventilation types. This suggests that natural ventilation models should account for both of these parameters working together.
Conclusion: Mira: So, looking at this paper, "Simulating The Urban Canopy's Impact on Wind-Driven Natural Ventilation," the core message is that the interaction between urban canopy geometry and wind angle determines how effective different natural ventilation setups actually are.
Kai: It’s not just about having a big canopy or a small one; it’s about the specific arrangement of those surrounding buildings that really dictates your airflow. They highlight how local interference effects can alter the rates by up to sixty percent in houses near the edges of quadrants compared to those in the centers.
Lev: From an engineering standpoint, this suggests that current models aren't capturing these interference effects well because they treat them separately. The paper points out that existing natural ventilation models need to represent both the local and larger-scale effects of a specific urban canopy geometry for better accuracy.
Mira: And they also noted something about scaling up—performing large eddy simulations for every building designed to use natural ventilation is not feasible right now. So, the LES data they present serves as a starting point for developing those improved models.
Kai: Ultimately, this work shows that the urban canopy effect on ventilation is much more complex than what many existing natural ventilation parameterizations suggest. It underscores that we need better models to accurately assess cooling in realistic urban areas based on these specific geometry effects.
Stanford University · Autodesk Research
physics.flu-dyn
Submitted: 2025-08-06
Updated: 2026-11-10
Comments: 19 pages, 11 figures
Journal ref: Flow 6, E1 (2026)
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 85/100
The gist: The urban canopy affects wind in complex ways, making it challenging to predict wind-driven natural ventilation and cooling in buildings Using large eddy simulations of coupled outdoor and indoor…
Key concepts
- Large Eddy Simulations (LES)
- A computational method used to model complex fluid dynamics, like wind flow around buildings. LES breaks down turbulent airflow into different scales, allowing researchers to simulate how wind interacts with detailed urban geometries and building features.
- Ventilation Configurations
- Different ways a building can be ventilated, such as cross-room or single-sided setups. The study tested four configurations (cross, corner, dual-room, single-sided) to see how the surrounding environment impacts the airflow for each specific design.
- Interference Effects
- The influence of neighboring buildings on local wind speed and direction. These effects are highly dependent on both the wind angle and the specific geometry of nearby structures, making them a major, complex factor in determining actual ventilation rates.
Terminology
Summary
The urban canopy affects wind in complex ways, making it challenging to predict wind-driven natural ventilation and cooling in buildings Using large eddy simulations of coupled outdoor and indoor airflow, we study how the surrounding urban canopy and wind angle influence ventilation rates through four ventilation configurations: cross, corner, dual-room, and single-sided Flow visualizations demonstrate how both large-scale flow patterns and local interference effects can influence ventilation rates by 50-85% In general, lower density canopies give higher ventilation rates and wind angles that align with a direct path between two openings also lead to higher ventilation rates However, interference effects from surrounding buildings can significantly change the local wind speed and direction, thus also changing ventilation rates The magnitude of these interference effects depends on both the wind angle and surrounding building geometries
Simulation Setup and Model Description
The study employs large eddy simulations (LESs) to predict coupled outdoor and indoor airflow through specific canopy and building geometries The computational domain includes 140 buildings distributed over four quadrants, with each quadrant containing an identical group of 35 houses forming a five-by-seven staggered grid Two domains representing different canopy densities are considered, with house centers being 50% further apart in the low density domain compared to the high density domain The building interiors include four representative ventilation configurations: cross, corner, dual-room, and single-sided The mesh resolution varies across the domain, extending for six elements around interior and exterior walls for houses with simulated interiors
Flow Field Analysis
Domain-scale flow patterns reveal that lower density domains support a faster in-canopy flow, consistent with smaller frontal area fractions (lambda f) The aligned quadrants, which have a smaller lambda f, support an overall faster in-canopy flow with building wakes confined to the smaller side yards Diagonal flow cases also reveal domain-scale variability in wind speed stemming from the anisotropic nature of the drag exerted by a quadrant on the flow
Ventilation Rate Sensitivity
The dual-room configuration experiences the largest median ventilation rate at 0.22, followed by the cross (10% less), corner (35% less), and single-sided (85% less) ventilation types On average, the higher density canopy gives 15% lower ventilation rates and 40% lower IQRs compared to the low density canopy Conditioning on both density and wind angle provides a significant decrease in IQRn for multi-window ventilation types, showing that natural ventilation models should represent the combined effect of these two parameters
Implications for Modeling
The results indicate that existing models do not accurately capture the interference effects that arise in real urban canopies, as changes in flow rates depend on both the wind angle and surrounding building geometries To improve accuracy, natural ventilation models must represent both the local and larger-scale effects of a specific urban canopy geometry Performing large-scale LESs for every building designed to use natural ventilation will be out of reach for the foreseeable future The LES data presented can provide a starting point for developing such models
Conclusion
This study reveals a nuanced interaction between urban canopy geometry and wind angle, which combine to determine the effectiveness of different natural ventilation configurations The importance of the specific configuration of surrounding buildings was further highlighted by considering the ventilation rates obtained in houses near quadrant edges, where unique interference effects alter the ventilation rate by up to 60% compared to houses in quadrant centers Future work should focus on developing new models that can better represent these geometry-specific canopy effects to accurately assess natural ventilation and cooling at larger scales The LES data presented in this paper can provide a starting point for developing such models The authors acknowledge that the results reinforce the effects observed in the flow field visualizations, showing that interference effects from surrounding buildings can play an equally important role in determining ventilation rates for specific wind angles The dependency of these local and large-scale effects on specific urban geometry poses a significant modeling challenge given the diverse geometrical configurations in realistic urban areas The LES data presented in this paper can provide a starting point for developing such models The authors thank Stanford University and the Stanford Research Computing Center for providing computational resources and support that contributed to these research results The raw data are available at https://doi.org/10.25740/vb900bf9392 and from the corresponding author (N. p. 3 Figure 2: Canopy horizontal dimensions with 𝐿𝑅 = 4 m. Table 2 summarizes the corresponding frontal area fractions (𝜆 𝑓 ) of the urban canopies. Often, in the evening or at night, these forces exchange cooler outdoor air with warmer indoor
air. This outside air cools building surfaces, which function as thermal storage, absorbing heat and
1
arXiv:2508. Section 4 summarizes the conclusions. 2 Description of computational model and quantities of interest
This section describes the governing equations, computational domain, mesh, boundary conditions, and
other forcings. Based on this analysis,
we selected two typical wind and temperature scenarios: 10 m height wind speeds of 2 and 4 m/s and
indoor-outdoor temperature differences of 0 and 5 ◦C. We performed simulations for the four possible combinations of these conditions, assuming neutral
conditions for the outdoor wind flow. Based on this analysis,
we selected two typical wind and temperature scenarios: 10 m height wind speeds of 2 and 4 m/s and
indoor-outdoor temperature differences of 0 and 5 ◦C. We performed simulations for the four possible combinations of these conditions, assuming neutral
conditions for the outdoor wind flow. Simulating The Urban Canopy’s Impact on Wind-Driven Natural
Ventilation
Nicholas Bachand1
, Hesam Salehipour2
, Catherine Gorlé1
1Civil and Environmental Engineering Department, Stanford University, Stanford, USA
2Autodesk Research, Toronto, Canada
Keywords: Natural Ventilation, Natural Cooling, Large Eddy Simulation, Urban Canopy Flow
Abstract
The urban canopy affects wind in complex ways, making it challenging to predict wind-driven
natural ventilation and cooling in buildings. Using large eddy simulations of coupled outdoor and
indoor airflow, we study how the surrounding urban canopy and wind angle influence ventilation rates
through four ventilation configurations: cross, corner, dual-room, and single-sided.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Pruning redundant reasoning tokens
The urban canopy affects wind in complex ways, making it challenging to predict wind-driven natural ventilation and cooling in buildings Flow visualizations demonstrate how both large-scale flow patterns and local interference effects can influence ventilation rates by 50-85% The magnitude of these interference effects depends on both the wind angle and surrounding building geometries
How the training signal is built
The instantaneous ventilation rate Q(t) is calculated as a time-averaged, non-dimensional ventilation rate from Q(t) = 1/2 sum over Nw i=1 integral Ai u(t) · ni dA The time-averaged, non-dimensional ventilation rate is then obtained from Qn = (1/Nt) sum i=1 Q(t) / U10A To analyze and interpret the variability in the calculated ventilation rates, the median and IQR are used because these estimators are interpretable and robust to outliers The conditional sample-normalized IQR is defined as IQRn(Qn Pi:j) = IQR(Qn Pij) / Eboot[IQRn(Qn)]
Where the gains come from
Wind angles that align with a room’s ventilation axis or place one window on the upwind side of the building often drive high ventilation rates In both cases, the inflow window is likely to be located in a high-pressure stagnation region, whereas the outflow window will be in a lower pressure region Interference effects from surrounding buildings can play an equally important role in determining ventilation rates for specific wind angles The importance of the specific configuration of surrounding buildings was further highlighted by considering the ventilation rates obtained in houses near quadrant edges The results demonstrate that the canopy’s effect on ventilation rates is much more complex than those suggested by existing natural ventilation parameterizations The results present a first step towards developing improved models that can efficiently represent the dominant urban canopy effects The results provide a first step towards developing improved models that can efficiently represent the dominant urban canopy effects N.B. wrote the manuscript, which H.S. and C.G.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved AI system can predict natural ventilation rates with higher accuracy by incorporating local flow field interference effects, as noted in
interference effects from surrounding buildings can significantly change the local wind speed and direction, thus also changing ventilation rates.
-
Improved AI system can better characterize urban canopy impact by using a model that accounts for both domain-scale variability and local flow patterns, addressing the finding that
domain-scale flow patterns reveal that both lambda f and anisotropy in the urban canopy drag can cause large-scale variations in wind speed that will impact interior flow.
-
Improved AI system can improve design recommendations by moving beyond single correction coefficients to represent
the combined effect of these two parameters [canopy density and wind angle] in order to make accurate predictions,
as suggested by Figure 10.
Abstract
The urban canopy affects wind in complex ways, making it challenging to predict wind-driven natural ventilation and cooling in buildings. Using large eddy simulations of coupled outdoor and indoor airflow, we study how the surrounding urban canopy and wind angle influence ventilation rates through four ventilation configurations: cross, corner, dual-room, and single-sided. Flow visualizations demonstrate how both large-scale flow patterns and local interference effects can influence ventilation rates by 50-85%. In general, lower density canopies give higher ventilation rates and wind angles that align with a direct path between two openings also lead to higher ventilation rates. However, interference effects from surrounding buildings can significantly change the local wind speed and direction, thus also changing ventilation rates. The magnitude of these interference effects depends on both the wind angle and surrounding building geometries. The effect of wind angle is less pronounced in a higher density canopy, where the urban canopy geometry more strongly guides the flow. The results demonstrate that the canopy's effect on ventilation rates is much more complex than those suggested by existing natural ventilation parameterizations.
Related papers
- Emergent Transfer of a Physics Foundation Model from Simulation to Laboratory Turbulence
- Early onset of secondary shear instability in Kelvin-Helmholtz braids at high Reynolds number
- Rotational covariance, continuous spectrum and non-modal growth of three-dimensional planar weak viscous shock layers
- Deep Learning-based Algebraic Reynolds Stress Closures for RANS Simulations of Turbulent Flows
- Neptuna: A Comprehensive Machine Learning Framework for Benchmarking Complex Multiphase Flows
- Size and shape of condensing steam jets and plumes