A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
A zoning method for the extreme wind pressure coefficients of buildings based on weighted K-means clustering
https://orcid.org/0000-0001-8541-4592
https://orcid.org/0000-0003-0473-2631
Abstract The wind pressure distribution on building surface is uneven and it varies considerably due to wind turbulence. The engineering design of cladding and components of a building often requires knowledge of the envelope of extreme wind pressure coefficients (Cp) over all directions at different locations. This study proposes a zoning method based on weighted K-Means cluster analysis with the criterion of similar extreme Cp and their proximal spatial locations within the zone. An optimal set of weighting factors is estimated by a new algorithm to balance the influences of the extreme Cp and their spatial locations. The sum of squares of distance and the mean silhouette coefficient are adopted to reduce the range of cluster number studied to improve the efficiency of ranking the clustering results. The proposed wind pressure zoning method based on clustering is further modified to avoid multiple folds or zones with too small an area. An evaluation criterion is proposed to finalize the wind pressure zoning results. The proposed method is then applied to the roof/walls of low-rise and high-rise buildings. Results are noted consistent with the wind pressure distribution of the structures, which is helpful for determining the design wind load of cladding & component.
Highlights The methodology estimated the optimal weighting factors of the extreme Cp and spatial locations with a new algorithm. The reasonable clustering validity indices are adopted to improve the efficiency of comparing different zoning results. The zoning evaluation criteria is newly proposed to consider the influence of subjective cooperation from the clustering result to the zoning result.
A zoning method for the extreme wind pressure coefficients of buildings based on weighted K-means clustering
https://orcid.org/0000-0001-8541-4592
https://orcid.org/0000-0003-0473-2631
Abstract The wind pressure distribution on building surface is uneven and it varies considerably due to wind turbulence. The engineering design of cladding and components of a building often requires knowledge of the envelope of extreme wind pressure coefficients (Cp) over all directions at different locations. This study proposes a zoning method based on weighted K-Means cluster analysis with the criterion of similar extreme Cp and their proximal spatial locations within the zone. An optimal set of weighting factors is estimated by a new algorithm to balance the influences of the extreme Cp and their spatial locations. The sum of squares of distance and the mean silhouette coefficient are adopted to reduce the range of cluster number studied to improve the efficiency of ranking the clustering results. The proposed wind pressure zoning method based on clustering is further modified to avoid multiple folds or zones with too small an area. An evaluation criterion is proposed to finalize the wind pressure zoning results. The proposed method is then applied to the roof/walls of low-rise and high-rise buildings. Results are noted consistent with the wind pressure distribution of the structures, which is helpful for determining the design wind load of cladding & component.
Highlights The methodology estimated the optimal weighting factors of the extreme Cp and spatial locations with a new algorithm. The reasonable clustering validity indices are adopted to improve the efficiency of comparing different zoning results. The zoning evaluation criteria is newly proposed to consider the influence of subjective cooperation from the clustering result to the zoning result.
A zoning method for the extreme wind pressure coefficients of buildings based on weighted K-means clustering
https://orcid.org/0000-0001-8541-4592
https://orcid.org/0000-0003-0473-2631
Abstract The wind pressure distribution on building surface is uneven and it varies considerably due to wind turbulence. The engineering design of cladding and components of a building often requires knowledge of the envelope of extreme wind pressure coefficients (Cp) over all directions at different locations. This study proposes a zoning method based on weighted K-Means cluster analysis with the criterion of similar extreme Cp and their proximal spatial locations within the zone. An optimal set of weighting factors is estimated by a new algorithm to balance the influences of the extreme Cp and their spatial locations. The sum of squares of distance and the mean silhouette coefficient are adopted to reduce the range of cluster number studied to improve the efficiency of ranking the clustering results. The proposed wind pressure zoning method based on clustering is further modified to avoid multiple folds or zones with too small an area. An evaluation criterion is proposed to finalize the wind pressure zoning results. The proposed method is then applied to the roof/walls of low-rise and high-rise buildings. Results are noted consistent with the wind pressure distribution of the structures, which is helpful for determining the design wind load of cladding & component.
Highlights The methodology estimated the optimal weighting factors of the extreme Cp and spatial locations with a new algorithm. The reasonable clustering validity indices are adopted to improve the efficiency of comparing different zoning results. The zoning evaluation criteria is newly proposed to consider the influence of subjective cooperation from the clustering result to the zoning result.
A zoning method for the extreme wind pressure coefficients of buildings based on weighted K-means clustering
Yang, Qingshan (author) / Yin, Jiaqi (author) / Liu, Min (author) / Law, Siu-seong (author)
2022-08-03
Article (Journal)
Electronic Resource
English
Extreme wind prediction and zoning
| Online Contents | 2005
Extreme wind prediction and zoning
| British Library Online Contents | 2005
Zoning of reservoir water temperature field based on K-means clustering algorithm
| DOAJ | 2022
Zoning of reservoir water temperature field based on K-means clustering algorithm
| Elsevier | 2022