A cloud-based platform to predict wind pressure coefficients on buildings

• Published in: Building simulation Vol. 15; no. 8; pp. 1507 - 1525
• Main Authors: Bre, Facundo, Gimenez, Juan M.
• Format: Journal Article
• Language: English
• Published: Beijing Tsinghua University Press 2022; Springer Nature B.V
• Subjects: Air quality; Atmospheric boundary layer; Atmospheric models; Atmospheric Protection/Air Quality Control/Air Pollution; Building Construction and Design; Building design; Buildings; Cloud computing; Coefficients; Computational fluid dynamics; Engineering; Engineering Thermodynamics; Floorplans; Green buildings; Heat and Mass Transfer; Indoor air pollution; Indoor air quality; Mathematical models; Monitoring/Environmental Analysis; Research Article; Urban environments; Wind pressure; Wind tunnels; Workflow; EnergyPlus; wind pressure coefficient; natural ventilation; building simulation; computational fluid dynamics; airflow network model
• ISSN: 1996-3599; 1996-8744
• DOI: 10.1007/s12273-021-0881-9

Natural ventilation (NV) is a key passive strategy to design energy-efficient buildings and improve indoor air quality. Therefore, accurate modeling of the NV effects is a basic requirement to include this technique during the building design process. However, there is an important lack of wind pressure coefficients (
C
p
) data, essential input parameters for NV models. Besides this, there are no simple but still reliable tools to predict
C
p
data on buildings with arbitrary shapes and surrounding conditions, which means a significant limitation to NV modeling in real applications. For this reason, the present contribution proposes a novel cloud-based platform to predict wind pressure coefficients on buildings. The platform comprises a set of tools for performing fully unattended computational fluid dynamics (CFD) simulations of the atmospheric boundary layer and getting reliable
C
p
data for actual scenarios. CFD-expert decisions throughout the entire workflow are implemented to automatize the generation of the computational domain, the meshing procedure, the solution stage, and the post-processing of the results. To evaluate the performance of the platform, an exhaustive validation against wind tunnel experimental data is carried out for a wide range of case studies. These include buildings with openings, balconies, irregular floor-plans, and surrounding urban environments. The
C
p
results are in close agreement with experimental data, reducing 60%–77% the prediction error on the openings regarding the EnergyPlus software. The platform introduced shows being a reliable and practical
C
p
data source for NV modeling in real building design scenarios.