Wind Load On High-rise Building And Extreme Value Estimation Method

Compared to the main structure,the high-rise building envelope is prone to the winddamaging,and wind-resistant design of envelope structure is an important and difficult problem in wind engineering research.The extreme wind pressure coefficient of the high-rise building surface is one of the key variables of the envelope structure design wind load.Currently,the common wind pressure coefficient extreme value estimation method is block maximum extreme value estimation method（BM）,also known as a classical extreme value estimation method.Exceeding threshold extreme value estimation method（POT）can make better use of sample information,and can overcome the shortcomings of BM method,which is timeconsuming,laborious and requires a large number of samples.However,the traditional POT method has some shortcomings,such as the difficulty of threshold selection,the ambiguity of the benchmark compared with the classical extreme value estimation methods,the correlation of samples and so on.Based on the rigid model pressure test of CAARC（Commonwealth Advisory Aeronautical Research Council）high-rise building,this article makes a systematic study on the surface wind load characteristics of high-rise building and the POT method in the surface extreme wind pressure coefficient estimation method.The main research contents and conclusions are as follows:1)A CAARC standard tall building model with a scale of 1:400 is designed and made to simulate exposure categories D.Under the typical wind direction of 0° and 45°,wind tunnel tests were carried out on the CAARC standard tall building model,and 200 independent repeated samples were carried out.This article studies the spatial distribution of average,fluctuating and extreme wind pressure coefficients of high-rise buildings and their nonGaussian characteristics,and gives the best probability distribution model of single and multiple wind pressure coefficients.2)Aiming at the demand of wind pressure coefficient extreme value estimate with small sample size,POT method was used to estimate extreme wind pressure coefficient.The changepoint theory was introduced,and the parameter stability of GPD probability distribution model was taken as the judging benchmark.An automatic threshold selection method based on local comparison method was proposed to construct an improved POT extreme value estimate method.By comparing with BM extreme estimation method,the applicability of the improved POT extreme value estimation method for wind pressure coefficient with small sample size is studied,and the minimum sample size requirement of the improved POT extreme value estimation method is given.At the same time,the uncertainty of the improved POT extreme value estimation method is studied.The results show that the accuracy and stability of the improved POT extreme value estimation method meet the expectation when the sample size is small.3)In this article,the important factors（independent peak extraction method,parameter estimation method,threshold selection method）that affect the POT extreme value estimation method are studied.（1）The comparison of the number of extracted peaks and autocorrelation shows that the Dat Duthinh peak extraction method based on mean exceeding can obtain better independent and identically distributed peak samples.（2）Monte Carlo method is used to study the uncertainty of GPD parameter estimation method under different shape parameters,size parameters and sample size.the results show that the best parameter estimation method is probability weighted moments estimation method.（3）The sensitivity of different threshold selection methods to sample size and non-Gaussian characteristics is studied.The experimental analysis of small sample size shows that the threshold selection method based on W statistics and the change-point threshold selection method based on shape parameter stability are suitable for tap of low kurtosis-low skewness and middle kurtosis-middle skewness.The change-point threshold selection method based on extreme stability is suitable for high kurtosis-high skewness tap.