基于k-均值聚类算法的高层建筑表面风压分区研究  

Research on Wind Pressure Zoning of High-rise Buildings Based on k-Means Clustering Algorithm

作　　者：王健 陈统岳 朱杰 WANG Jian;CHEN Tongyue;ZHU Jie Shanghai(Construction Group Co.,Ltd.,Shanghai 200080,China)

机构地区：[1]上海建工集团股份有限公司,上海200080

出　　处：《建筑施工》2024年第7期1001-1004,共4页Building Construction

基　　金：国家重点研发计划(2022YFC3802200)。

摘　　要：为分析高层建筑表面的风压特征和关键区域,以高宽比为4∶1的高层建筑风洞试验模型为对象,采用k-均值聚类算法,对0°风向角下模型各个面的风压测压管时程数据进行分析,研究结果表明:建筑左、右侧面以强烈的负压为主导,且角点附近存在负压极值;k-均值聚类算法可以有效地识别不同表面风压场的特征,风压的聚类结果与平均风压系数的分布较为吻合,且能得到代表性的风压测压管。To analyze the wind pressure characteristics and key areas on the surface of a high-rise buildings,a wind tunnel test model of a high-rise building with a height-width ratio of 4:1 was taken as the object,and the k-means clustering algorithm was used to analyze the time history data of wind pressure piezometers on each side of the model under a wind direction angle of 0°.The results show that the left and right sides of the building are dominated by strong negative pressure,with negative pressure extremes near the corners;the k-means clustering algorithm can effectively identify the characteristics of different surface wind pressure fields,and the clustering results of wind pressure are consistent with the distribution of average wind pressure coefficient,and the representative wind pressure piezometer tubes can be obtained.

关 键 词：高层建筑 K-均值聚类 风压分布 风洞试验 

分 类 号：TU971[建筑科学—建筑理论]