Study On Non-stationary Wind Spectral Model Of Recorded Typhoon At Sutong Bridge Site

In twenty-first century,bridge engineering field raise the high tide period of construction of sea-crossing and island-linking projects.As one kind of long span bridge types,the development of cable-stayed bridge is concerned greatly.However,with the increasing length of main span of the cable-stayed bridge,the structural stiffness of the bridge will decrease.This will lead to the increase of the structural sensibility to the wind load.In addition,the wind effects on long span cable-stayed bridges are no negligible.Existing studies have shown that the recorded wind speeds exhibits obvious non-stationary features,which deviates from the stationary hypothesis in traditional study on wind engineering.Further study on the non-stationary features of typhoon is of great significance to refinement study on wind-induced responses and develop wind resistant theory.However,field measurement study on the non-stationary features of typhoon at long span bridge site is relatively few and the research needs to go further.In this paper,the Sutong Bridge is selected to be the engineering background.Field measurement study is conducted to research the non-stationary features of typhoon using structural health monitoring system(SHMS).The evolutionary power spectral density(EPSD)of non-stationary turbulence during typhoon is calculated using wavelet transform.And parameter fitting studies of the proposed EPSD model are conducted.The main contents are as follows.(1)The stationarity study on recorded turbulence of typhoon.A local recurrence loss value is defined to quantify the stationarity of recoded data based on recurrence plot(RP)and recurrence quantification analysis(RQA).A program to calculate the local recurrence loss value is finished with Matlab.The local recurrence loss value is evaluated with a numerical experiment.Rusults show that the non-stationarity of recorded typhoon data can be evaluated reasonably by the proposed local recurrence loss value.(2)Comparison study on stationary and non-stationary features of the recorded typhoon at Sutong Bridge site.The recorded wind speed data at the bridge site during typhoon "Kalmaegi" and "Morakot" are taken as examples.Their stationarity is evaluated.A comparison study on the mean and turbulent wind characteristics is conducted based on stationary and non-stationary wind speed model.Results show that the recorded wind data contain non-stationary components,the calculated turbulent wind characteristics are different when using different wind speed models.(3)The study on EPSD of recorded turbulence during typhoon.A GHWT-based EPSD estimation method is elucidated clearly.A numerical experiment is conducted to verify the validity of the EPSD estimating program.Then,the EPSD of recorded non-stationary turbulence during typhoon is estimated and compared with the deduced Kaimal EPSD.Results show that the GHWT-based EPSD estimation method is more appropriate for the EPSD estimation of non-stationary recorded turbulence data during typhoon.In general,the time slices of Kaimal EPSD agree well with the measured ones.However,the kaimal EPSD model still needs to be improved so that it can capture the time-frequency features of the recorded turbulent data well.(4)The parameter fitting study on EPSD model of turbulence during typhoon.The modified Quasi-Newton method is elucidated and a numerical experiment is conducted to verify the validity of the program.An estimation flowchart is proposed to estimate parameters of EPSD model using Quasi-Newton method.The measured EPSD of typhoon "Kalmaegi" is taken as the fitting data and the parameters of EPSD model are estimated using proposed estimation flowchart.A comparative analysis between the fitted and measured EPSD is conducted.Results show that the proposed fitting process of the parameter fitting problem reduces the calculation scale and improves the computational efficiency of the iterative algorithm.The time slices of the fitted EPSD agree well with the measured ones near the extreme points of the measured EPSD.