Pith. sign in

REVIEW 5 major objections 5 minor 1 cited by

From large-eddy simulations to deep learning: A U-net model for fast urban canopy flow predictions

T0 review · 5 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A U-Net trained on LES of 252 synthetic city layouts predicts urban canopy wind fields in about one second, with mean relative errors near 9% for wind speed and 5% for turbulence.

desk verdict Honest, workmanlike U-Net surrogate for urban canopy flow; the accuracy claims hold for its synthetic distribution, but the abstract and per-height reporting need tightening before the headline numbers can be taken at face value. read the letter →

arxiv 2507.06533 v1 pith:PJXPYWR7 submitted 2025-07-09 physics.comp-ph cs.LGphysics.flu-dyn

Themistoklis Vargiemezis , Catherine Gorlé This is my paper · ORCID
classification physics.comp-phcs.LGphysics.flu-dyn
keywords urbancanopyflowU-Netlarge-eddysimulationsurrogatepedestrianwindcomfortturbulenceintensitypredictionsigneddistancefunctiondeeplearningforCFD
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper aims to establish that a U-Net can act as a fast surrogate for large-eddy simulation in urban canopies. Trained on 252 synthetic city layouts simulated with LES at wind directions from $0^\circ$ to $90^\circ$, the network takes a $256\times256\times9$ tensor built from a binary building mask, a signed distance function, and its gradient at three heights, and outputs the mean velocity magnitude $U_{\mathrm{mag}}$ and the streamwise turbulence intensity $I_u$ at those heights. On 50 held-out test cases the reported mean relative error is 9.3\% for $U_{\mathrm{mag}}$ and 5.2\% for $I_u$ over the full field, while the evaluation time drops from roughly 10 hours on 32 CPUs to about one second on one GPU. If this holds, urban wind assessment becomes cheap enough to screen many design variants before committing to expensive simulations.

What carries the argument

The central object is a U-Net whose encoder compresses the geometry tensor into a latent space and whose decoder reconstructs the two quantities of interest, with skip connections preserving spatial detail. A Spatial Attention Module re-weights the skip-connection features before decoding, and a binary mask from the input is added at the output so the network learns the flow field rather than the building surfaces. The loss combines RMSE of the quantities, RMSE of their Sobel-computed gradient magnitude, and L2 weight regularization; this gradient term is what pushes the predictions to keep sharp edges at building faces and shear layers.

What would settle it

Evaluate the model on a real urban district with non-rectilinear buildings, variable roof heights, and an oblique street network, comparing pedestrian-height predictions against LES or field measurements; if the mean relative error for $U_{\mathrm{mag}}$ in the pedestrian zone substantially exceeds the reported 9.3\% or the hit rate falls below 90\%, the central claim of practical, geometry-general accuracy is contradicted.

Watch

Extended reading notes

Core claim

The central claim is that the spatial structure of the two quantities that matter for pedestrian wind comfort can be learned from geometry alone. The U-Net reproduces wake formation, flow separation, and downstream recovery across layouts and wind directions, with errors concentrated in building wakes at the highest of the three canopy heights, where three-dimensional effects are strongest. The paper reads this as evidence that the model induces a generalizable geometry-to-flow mapping rather than memorizing the training layouts, and it reports accurate predictions even in a central region of interest where wakes and shear layers dominate.

Load-bearing premise

The riskiest premise is that three horizontal slices of 2D geometry carry enough information for the network to infer the three-dimensional flow structures that set wind speed and turbulence; the paper itself finds the largest errors at the top height, where vertical shear and wake interactions dominate, and all training cases are synthetic, rectilinear cities under neutral atmospheric conditions.

Editorial extensions

If this is right

  • Pedestrian comfort maps based on the effective wind $U_e = U_{\mathrm{mag}} + k\,\sigma_u$ follow directly from the predicted fields, so a full comfort assessment can be produced in about a second per configuration.
  • Accuracy is highest for wind aligned with the building grid ($0^\circ$ and $90^\circ$) and lowest at $45^\circ$, so adding more diagonal wind directions to the training data is a concrete route to better generalization.
  • The largest errors occur at the top height slice, where three-dimensional wake interactions matter most; the paper's own suggested remedy is to add more vertical slices or 3D spatial features to the input.
  • The speed of the model makes iterative urban design and layout screening practical, with LES kept as the verification tool for the final configuration.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A stricter generalization test would evaluate on 50 physically distinct layouts rather than 25 layouts plus their vertical flips; the augmentation doubles the test set and shares symmetry with the training augmentation, so the reported 50-case metrics may overstate true diversity.
  • Because the model already receives distance-to-building information, adding a building-height map as a fourth input channel is a cheap, directly testable way to attack the top-height errors without changing the architecture.
  • The same geometry-to-statistics pipeline should extend to other scalar quantities with LES training data, such as pressure coefficients or pollutant concentrations, but the paper does not demonstrate this.
  • The reported speed-up of roughly four orders of magnitude compares inference on one GPU with LES on 32 CPUs; the total cost of producing the 252 training simulations is the amortized price of that speed, so the surrogate pays off only when many evaluations are needed.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. The paper introduces a U-Net-based deep learning surrogate trained on 252 synthetic urban geometries simulated with large-eddy simulation (LES) at seven wind directions. The model takes three 2D input representations per height (binary building mask, signed distance function, and its gradient) and predicts mean velocity magnitude and streamwise turbulence intensity at three heights (0.1, 0.3, and 0.5 m). A spatial attention module is inserted in the skip connections, and the loss combines RMSE, gradient-magnitude RMSE, and L2 regularization. The authors report a speedup from about 10 hours on 32 CPUs for LES to about 1 second on a GPU, with overall mean relative errors of 9.3% (velocity magnitude) and 5.2% (turbulence intensity) on the test set, and higher errors in a central region of interest. Hyperparameters were tuned on a dev set, and the code is publicly available.

Significance. If the reported performance holds, this is a practically useful fast surrogate for preliminary pedestrian-level wind assessment in synthetic urban geometries. The strengths are the clean train/dev/test protocol with hyperparameter tuning on the dev set, the held-out test evaluation, the public code release, and an unusually candid limitations section. The U-Net architecture and input representation are incremental rather than radically novel, but the scale of the LES dataset and the reported speedup give the contribution practical value. The main weakness is that the headline metrics are aggregates over heights and over the full domain; the text itself shows that errors are larger at the highest height and in the region of interest, so the current presentation makes the central accuracy claim hard to evaluate precisely. These issues are fixable with more detailed reporting rather than requiring a change of approach.

major comments (5)
  1. [Abstract; Section 5] The abstract and conclusions state an overall mean relative error of 9.3% for velocity magnitude, but Table 3 reports 9.1% for the same quantity over the same test set. Since this is the headline accuracy figure, the discrepancy must be resolved: either the 9.3% is an error and should be corrected to 9.1%, or the two numbers come from different weighting/masking conventions and the convention should be stated.
  2. [Section 4.4, Table 3; Sections 4.2.3 and 4.3.3] The paper claims accurate predictions at 'multiple heights', but Table 3 reports metrics only after averaging over the three heights. The qualitative results in Sections 4.2.3 and 4.3.3 and the line plots in Figures 7 and 11 show that errors are largest at Height 3, with localized relative errors of 10-25% for Umag and 15-20% for Iu, and RoI hit rates dropping to 89-90% at that height. Since Section 4.5 itself acknowledges that the 2D input representation limits the capture of 3D flow dynamics at higher heights, the authors should report per-height MRE, NRMSE, and hit rate over the full test set, at least for the full domain and the region of interest, so the central 'multiple heights' claim can be verified.
  3. [Eq. (10); Section 3.2] The MRE definition in Eq. (10) is computed over all N pixels, but the paper never specifies whether building pixels are excluded from the metric. Inside buildings the QoIs are not defined by the LES, and the denominator |Q_i^T| can be zero, making the per-pixel relative error undefined. The binary skip connection described in Section 3.2 forces the prediction to match the building mask, but no corresponding statement is made for the evaluation metrics. The authors should state explicitly how building pixels are treated in Eq. (10) and in the NRMSE and hit-rate computations.
  4. [Section 3.4; Section 4.4] The test set is described as 25 cases augmented to 50 by flipping each image vertically. The abstract and conclusions refer to '50 test cases' as the evaluation sample, but the flipped versions are deterministic transformations of the same 25 geometries. The effective number of independent test configurations is therefore 25, not 50. This matters because no confidence intervals are reported, and the effective sample size should be stated clearly to avoid overstating the breadth of the evaluation.
  5. [Section 4.4, Table 3] The text says the results are compared 'across all wind directions' and that metrics are provided for each wind direction case, but Table 3 contains only '0 deg', '45 deg', and 'Overall' rows. The dataset covers seven wind directions (15, 30, 60, 75, and 90 degrees are absent from the table). Without per-direction metrics for all seven directions, the claims about generalization across wind directions in Sections 4.4 and 5 are only supported for two selected cases, and the reader cannot see which directions contribute most to the overall error.
minor comments (5)
  1. [Section 2.2] The phrase 'identity regions that may pose risks' should read 'identify regions that may pose risks'.
  2. [Section 4.2.3] There are several typos: 'indicitating' should be 'indicating', 'ad 93%' should be 'and 93%', and 'Similarlry' should be 'Similarly'.
  3. [Section 4.3.3] The typo 'Similarlry' appears again, and 'To asses the overall performance' should be 'To assess the overall performance'.
  4. [Abstract; Section 4.5] The notation 'O10 hours' and 'O1 seconds' should be written as 'O(10 hours)' and 'O(1 second)' for consistency with standard asymptotic notation.
  5. [Figure 7] The caption states that 'pixel 0 lies downstream and pixel 250 upstream', but the images are 256x256 pixels, so the upstream end should be pixel 255; the description of the x-axis direction would also benefit from a clearer wording.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the U-net accuracy claims are empirical measurements on a held-out test set, with no equation reducing to a fitted value or self-citation chain.

full rationale

The paper's predictive claim is evaluated by comparing U-net outputs against held-out LES test cases. The loss function in Eq. 4 is a standard supervised RMSE plus gradient and regularization terms, and the evaluation metrics in Eqs. 8-10 are standard error measures; none of these equations is constructed from a fitted parameter in a way that forces the reported accuracy. Hyperparameters in Table 2 were selected using the development set, and the reported 50-case metrics come from test cases not used in training or hyperparameter tuning. Data augmentation by vertical flipping is applied separately to train, dev, and test sets, so the augmented test cases are not reused for training. The self-citations in Section 2.3, where the LES inflow condition, grid-sensitivity analysis, and setup are referred to Vargiemezis and Gorle (2024), transfer a previously established simulation configuration rather than importing the target prediction or a uniqueness claim; that prior work includes wind-tunnel comparisons and is not a circular justification of the present U-net results. Section 4.5's acknowledged limitations regarding synthetic geometries, 2D input slices, and neutral boundary-layer conditions concern external validity and generalization, not circularity. The minor numerical inconsistency between the abstract's 9.3% MRE and Table 3's 9.1% overall MRE is an internal consistency or reporting issue, not a circular-reasoning issue. No load-bearing step in the paper reduces an equation or metric to its own input by construction, and the accuracy numbers are genuine measurements on independent test data.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central claim is empirical and rests on the LES dataset, the chosen geometry distribution, and the model architecture. No new physical entities are introduced.

free parameters (6)
  • Reference channels C = 128
    Tuned via stochastic grid search on development set; controls model capacity.
  • Kernel size k = 5
    Tuned via stochastic grid search.
  • Gradient loss weight λ1 = 1e-3
    Tuned to balance RMSE and gradient magnitude loss.
  • L2 regularization weight λ2 = 7e-5
    Tuned to prevent overfitting.
  • Batch size = 16
    Tuned via stochastic grid search.
  • Learning rate = 5e-4
    Tuned via stochastic grid search.
assumptions (4)
  • domain assumption LES ground truth from CharLES with Vreman model is sufficiently accurate for this surrogate training.
    Section 2.3; the paper relies on prior validation of CharLES and a grid convergence study, but does not validate against wind tunnel data for these synthetic configurations.
  • domain assumption The flow can be represented by 2D horizontal slices at three heights; 3D effects are inferred from these slices.
    Section 3.1; the model uses a 256x256x9 tensor with three slices per height. The paper itself identifies this as a limitation at the highest height.
  • domain assumption Spanwise symmetry about the y-axis is valid for wind directions 0-90 degrees, justifying data augmentation by flipping.
    Section 4.4; used to double the dataset, but this reduces the effective number of independent test cases.
  • domain assumption The synthetic city geometries (70 randomly placed buildings, 10-50 m dimensions) are representative of real urban layouts.
    Section 2.1; the paper acknowledges in Section 4.5 that real cities have more complexity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of From large-eddy simulations to deep learning: A U-net model for fast urban canopy flow predictions." pith.science (2026). https://pith.science/paper/PJXPYWR7

@misc{pith2026250706533,
  author       = {Pith},
  title        = {Pith review of: From large-eddy simulations to deep learning: A U-net model for fast urban canopy flow predictions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PJXPYWR7}},
  note         = {Machine review of arXiv:2507.06533}
}
abstract

Accurate prediction of wind flow fields in urban canopies is crucial for ensuring pedestrian comfort, safety, and sustainable urban design. Traditional methods using wind tunnels and Computational Fluid Dynamics, such as Large-Eddy Simulations (LES), are limited by high costs, computational demands, and time requirements. This study presents a deep neural network (DNN) approach for fast and accurate predictions of urban wind flow fields, reducing computation time from an order of 10 hours on 32 CPUs for one LES evaluation to an order of 1 second on a single GPU using the DNN model. We employ a U-Net architecture trained on LES data including 252 synthetic urban configurations at seven wind directions ($0^{o}$ to $90^{o}$ in $15^{o}$ increments). The model predicts two key quantities of interest: mean velocity magnitude and streamwise turbulence intensity, at multiple heights within the urban canopy. The U-net uses 2D building representations augmented with signed distance functions and their gradients as inputs, forming a $256\times256\times9$ tensor. In addition, a Spatial Attention Module is used for feature transfer through skip connections. The loss function combines the root-mean-square error of predictions, their gradient magnitudes, and L2 regularization. Model evaluation on 50 test cases demonstrates high accuracy with an overall mean relative error of 9.3% for velocity magnitude and 5.2% for turbulence intensity. This research shows the potential of deep learning approaches to provide fast, accurate urban wind assessments essential for creating comfortable and safe urban environments. Code is available at https://github.com/tvarg/Urban-FlowUnet.git

Figures

Figures reproduced from arXiv: 2507.06533 by the authors.

Figure 1
Figure 1. Examples of 3D urban models (top) and their top views (bottom) [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Computational domain (top) and zoom-in on the buildings with grid [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Geometry representation and data preparation for the U-net inputs. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Model architecture for predicting velocity magnitude, [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: shows the contour plot comparison of the velocity magnitude, while [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Turbulence intensity Iu at 0˝ wind incidence. Left: LES ground truth, middle: U-net prediction, right: absolute error. Height 1, 2, and 3 correspond to 0.1m, 0.3m, and 0.5m. 6 [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]
Figure 7
Figure 7. Figure 7: Line plot comparisons of Umag and Iu at 0˝ wind incidence. 8 [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Error histogram at 0˝ wind incidence. 4.3. Flow predictions at 45˝ wind incidence 4.3.1. Contour plots of Umag and Iu In [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Velocity magnitude Umag at 45˝ wind incidence. Left: LES ground truth, middle: U-net prediction, right: absolute error. Height 1, 2, and 3 corre￾spond to 0.1m, 0.3m, and 0.5m [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]
Figure 10
Figure 10. Figure 10: Turbulence intensity Iu at 45˝ wind incidence. Left: LES ground truth, middle: U-net prediction, right: absolute error. Height 1, 2, and 3 corre￾spond to 0.1m, 0.3m, and 0.5m. Despite these localized discrepancies in magnitude, the predic￾tions follow the overall tren…
Figure 11
Figure 11. Figure 11: Line plot comparisons of Umag and Iu at 45˝ wind incidence. 11 [PITH_FULL_IMAGE:figures/full_fig_p011_11.png]
Figure 12
Figure 12. Figure 12: Error histogram at 45˝ wind incidence. symmetry about the y-axis is a reasonable assumption for our setup. To evaluate the performance of the U-net model, we use the following metrics: the hit rate, the NRMSE, and the MRE. All metrics are computed for the two QoIs. Th…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Large Eddy Simulation of Plunging Flows in Laboratory-Scale Bedrock Rivers

    physics.flu-dyn 2026-07 conditional novelty 6.0 of 10

    LES simulations reproduce laboratory plunging flows in bedrock-river constriction-pool-widening morphologies and indicate discharge-dependent optimal constrictions (~35% at low flow, ~50% at high flow) for velocity in...

Reference graph

Works this paper leans on

43 extracted references · 38 canonical work pages · cited by 1 Pith paper

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  3. [3]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in ":" * " " * FUNCTION f...

  4. [4]

    , author Vasan, N

    author Adamek, K. , author Vasan, N. , author Elshaer, A. , author English, E. , author Bitsuamlak, G. , year 2017 . title Pedestrian level wind assessment through city development: A study of the financial district in toronto . journal Sustainable cities and society volume 35 , pages 178--190

    work page 2017

  5. [5]

    , author Nagaoka, H

    author Ambo, K. , author Nagaoka, H. , author Philips, D.A. , author Ivey, C. , author Br \`e s, G.A. , author Bose, S.T. , year 2020 . title Aerodynamic force prediction of the laminar to turbulent flow transition around the front bumper of the vehicle using dynamic-slip wall model les , in: booktitle AIAA Scitech 2020 Forum , p. pages 0036

    work page 2020

  6. [6]

    , author Andrianne, T

    author Ba s , H. , author Andrianne, T. , author Reiter, S. , year 2024 . title City configurations to optimise pedestrian level ventilation and wind comfort . journal Sustainable Cities and Society volume 114 , pages 105745

    work page 2024

  7. [7]

    , author Bengio, Y

    author Bergstra, J. , author Bengio, Y. , year 2012 . title Random search for hyper-parameter optimization. journal Journal of machine learning research volume 13

    work page 2012

  8. [8]

    , year 2015

    author Blocken, B. , year 2015 . title Computational fluid dynamics for urban physics: Importance, scales, possibilities, limitations and ten tips and tricks towards accurate and reliable simulations . journal Building and Environment volume 91 , pages 219--245

    work page 2015

Show all 43 references
  1. [9]

    , author Stathopoulos, T

    author Blocken, B. , author Stathopoulos, T. , author Van Beeck, J. , year 2016 . title Pedestrian-level wind conditions around buildings: Review of wind-tunnel and cfd techniques and their accuracy for wind comfort assessment . journal Building and Environment volume 100 , pa...

    2016

  2. [10]

    , year 1999

    author Bottema, M. , year 1999 . title Towards rules of thumb for wind comfort and air quality . journal Atmospheric Environment volume 33 , pages 4009--4017

    1999

  3. [11]

    , year 2000

    author Bottema, M. , year 2000 . title A method for optimisation of wind discomfort criteria . journal Building and Environment volume 35 , pages 1--18

    2000

  4. [12]

    title Fidelity charles

    author Cadence Design Systems , year 2022 . title Fidelity charles . https://community.cadence.com/. note accessed: April 14, 2025

    2022

  5. [13]

    , author Gao, X

    author Chen, D. , author Gao, X. , author Xu, C. , author Wang, S. , author Chen, S. , author Fang, J. , author Wang, Z. , year 2022 . title Flowdnn: a physics-informed deep neural network for fast and accurate flow prediction . journal Frontiers of Information Technology & El...

    2022

  6. [14]

    , author Huang, Z

    author Ciarlatani, M.F. , author Huang, Z. , author Philips, D. , author Gorl \'e , C. , year 2023 . title Investigation of peak wind loading on a high-rise building in the atmospheric boundary layer using large-eddy simulations . journal Journal of Wind Engineering and Indust...

    2023

  7. [15]

    , author Welling, M

    author Cohen, T. , author Welling, M. , year 2016 . title Group equivariant convolutional networks , in: booktitle International conference on machine learning , organization PMLR . pp. pages 2990--2999

    2016

  8. [16]

    , author Hellsten, A

    author Franke, J. , author Hellsten, A. , author Schlunzen, K.H. , author Carissimo, B. , year 2011 . title The cost 732 best practice guideline for cfd simulation of flows in the urban environment: a summary . journal International Journal of Environment and Pollution volume ...

    2011

  9. [17]

    , author Gorl \'e , C

    author Hwang, Y. , author Gorl \'e , C. , year 2022 . title Large-eddy simulations of wind-driven cross ventilation, part1: validation and sensitivity study . journal arXiv preprint arXiv:2204.00786

    2022 arXiv

  10. [18]

    , author Gorl \'e , C

    author Hwang, Y. , author Gorl \'e , C. , year 2023 . title Large-eddy simulations to define building-specific similarity relationships for natural ventilation flow rates . journal Flow volume 3 , pages E10

    2023

  11. [19]

    , author Blocken, B

    author Janssen, W. , author Blocken, B. , author van Hooff, T. , year 2013 . title Pedestrian wind comfort around buildings: Comparison of wind comfort criteria based on whole-flow field data for a complex case study . journal Building and Environment volume 59 , pages 547--562

    2013

  12. [20]

    , author Yan, C

    author Jin-Yu, Z. , author Yan, C. , author Xian-Xiang, H. , year 2009 . title Edge detection of images based on improved sobel operator and genetic algorithms , in: booktitle 2009 International Conference on Image Analysis and Signal Processing , organization IEEE . pp. pages 31--35

    2009

  13. [21]

    , author Kim, J.J

    author Kang, G. , author Kim, J.J. , author Choi, W. , year 2020 . title Computational fluid dynamics simulation of tree effects on pedestrian wind comfort in an urban area . journal Sustainable Cities and Society volume 56 , pages 102086

    2020

  14. [22]

    , author Ba, J

    author Kingma, D.P. , author Ba, J. , year 2014 . title Adam: A method for stochastic optimization . journal arXiv preprint arXiv:1412.6980

    2014 arXiv

  15. [23]

    , author Blocken, B

    author Montazeri, H. , author Blocken, B. , author Janssen, W. , author van Hooff, T. , year 2013 . title Cfd evaluation of new second-skin facade concept for wind comfort on building balconies: Case study for the park tower in antwerp . journal Building and Environment volume...

    2013

  16. [24]

    , author Gromke, C

    author Moonen, P. , author Gromke, C. , author Dorer, V. , year 2013 . title Performance assessment of large eddy simulation (les) for modeling dispersion in an urban street canyon with tree planting . journal Atmospheric environment volume 75 , pages 66--76

    2013

  17. [25]

    , author Soligo, M.J

    author Phillips, D.A. , author Soligo, M.J. , et al., year 2019 . title Will cfd ever replace wind tunnels for building wind simulations? journal International journal of high-rise buildings volume 8 , pages 107--116

    2019

  18. [26]

    , author Kiefer, H

    author Plate, E.J. , author Kiefer, H. , year 2001 . title Wind loads in urban areas . journal Journal of Wind Engineering and Industrial Aerodynamics volume 89 , pages 1233--1256

    2001

  19. [27]

    , author Tominaga, Y

    author Potsis, T. , author Tominaga, Y. , author Stathopoulos, T. , year 2023 . title Computational wind engineering: 30 years of research progress in building structures and environment . journal Journal of Wind Engineering and Industrial Aerodynamics volume 234 , pages 105346

    2023

  20. [28]

    , author Li, Z

    author Qin, Z. , author Li, Z. , author Zhang, Z. , author Bao, Y. , author Yu, G. , author Peng, Y. , author Sun, J. , year 2019 . title Thundernet: Towards real-time generic object detection on mobile devices , in: booktitle Proceedings of the IEEE/CVF international conferen...

    2019

  21. [29]

    , author Fischer, P

    author Ronneberger, O. , author Fischer, P. , author Brox, T. , year 2015 . title U-net: Convolutional networks for biomedical image segmentation , in: booktitle Medical image computing and computer-assisted intervention--MICCAI 2015: 18th international conference, Munich, Ger...

    2015

  22. [30]

    , author Tognaccini, R

    author Saetta, E. , author Tognaccini, R. , author Iaccarino, G. , year 2024 . title Uncertainty quantification in autoencoders predictions: Applications in aerodynamics . journal Journal of Computational Physics volume 506 , pages 112951

    2024

  23. [31]

    , author Buccolieri, R

    author Salim, S.M. , author Buccolieri, R. , author Chan, A. , author Di Sabatino, S. , year 2011 . title Numerical simulation of atmospheric pollutant dispersion in an urban street canyon: Comparison between rans and les . journal Journal of Wind Engineering and Industrial Ae...

    2011

  24. [32]

    , author Guastoni, L

    author Srinivasan, P.A. , author Guastoni, L. , author Azizpour, H. , author Schlatter, P. , author Vinuesa, R. , year 2019 . title Predictions of turbulent shear flows using deep neural networks . journal Physical Review Fluids volume 4 , pages 054603

    2019

  25. [33]

    , author Wei enow, K

    author Thuerey, N. , author Wei enow, K. , author Prantl, L. , author Hu, X. , year 2020 . title Deep learning methods for reynolds-averaged navier--stokes simulations of airfoil flows . journal AIAA Journal volume 58 , pages 25--36

    2020

  26. [34]

    , year 2003

    author Tieleman, H.W. , year 2003 . title Wind tunnel simulation of wind loading on low-rise structures: a review . journal Journal of wind engineering and industrial aerodynamics volume 91 , pages 1627--1649

    2003

  27. [35]

    , author Blocken, B

    author Van Hooff, T. , author Blocken, B. , year 2010 . title Coupled urban wind flow and indoor natural ventilation modelling on a high-resolution grid: A case study for the amsterdam arena stadium . journal Environmental Modelling & Software volume 25 , pages 51--65

    2010

  28. [36]

    , author Gorl \'e , C

    author Vargiemezis, T. , author Gorl \'e , C. , year 2024 . title A predictive large-eddy simulation framework for the analysis of wind loads on a realistic low-rise building geometry . journal Journal of Wind Engineering and Industrial Aerodynamics

    2024

  29. [37]

    , author Gorle, C

    author Vargiemezis, T. , author Gorle, C. , year 2024 . title Wind tunnel experiments on realistic low-rise building in urban area . https://www.designsafe-ci.org/data/browser/public/designsafe.storage.published/PRJ-2571/#detail-4070145275427230191-242ac118-0001-012/?version=2...

    2024 doi

  30. [38]

    , author Gorl \'e , C

    author Vargiemezis, T. , author Gorl \'e , C. , year 2025 . title Predicting wind-induced interference effects on a low-rise building in a realistic urban area using large-eddy simulations. journal Journal of Wind Engineering and Industrial Aerodynamics

    2025

  31. [39]

    , year 2004

    author Vreman, A. , year 2004 . title An eddy-viscosity subgrid-scale model for turbulent shear flow: Algebraic theory and applications . journal Physics of fluids volume 16 , pages 3670--3681

    2004

  32. [40]

    , author Mochida, A

    author Yoshie, R. , author Mochida, A. , author Tominaga, Y. , author Kataoka, H. , author Harimoto, K. , author Nozu, T. , author Shirasawa, T. , year 2007 . title Cooperative project for cfd prediction of pedestrian wind environment in the architectural institute of japan . ...

    2007

  33. [41]

    , author Zhang, S

    author Zhai, J. , author Zhang, S. , author Chen, J. , author He, Q. , year 2018 . title Autoencoder and its various variants , in: booktitle 2018 IEEE international conference on systems, man, and cybernetics (SMC) , organization IEEE . pp. pages 415--419

    2018

  34. [42]

    , author Gu, M

    author Zhang, A. , author Gu, M. , year 2008 . title Wind tunnel tests and numerical simulations of wind pressures on buildings in staggered arrangement . journal Journal of Wind Engineering and Industrial Aerodynamics volume 96 , pages 2067--2079

    2008

  35. [43]

    , author Kwok, K.C

    author Zhang, S. , author Kwok, K.C. , author Liu, H. , author Jiang, Y. , author Dong, K. , author Wang, B. , year 2021 . title A cfd study of wind assessment in urban topology with complex wind flow . journal Sustainable Cities and Society volume 71 , pages 103006

    2021

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.