| The number of thunderstorms has increased dramatically throughout the world as a result of global climate change.When a thunderstorm develops to the mature cumulus stage,the cold air mass at its top will destabilize and collapse to form local strong downdrafts.After hitting the ground,the downdrafts will spread around and turn into strong jet streams,namely downbursts.Downbursts are characterized by abruptness,randomness,and strong destructiveness and the outflows of downbursts present prominent non-stationarity.When downbursts occur in the sites of long-span bridges out of the blue,the service performance and the driving safety of bridges will be seriously threatened.Therefore,based on this,this thesis takes the Sutong Cable-stayed Bridge(SCB),the first cable-stayed bridge in China with a span of more than 1,000 meters,as the engineering background and carries out the measured analysis and prediction of buffeting response of long-span cable-stayed bridge under the action of downburst winds.The main contents of this thesis mainly include:(1)Field measurement and analysis of nonstationary characteristics of the downburst.Based on the measured data of the downburst,the time-frequency analysis method has been utilized to automatically extract the components of time-varying trend and nonstationary turbulence from the original wind speed.Since the time-varying mean wind speed has been isolated,the wind characteristics of the residual turbulence,e.g.,turbulence intensity,gust factor,turbulence integral scale,and power spectral density(PSD),are emphatically analyzed,and then the differences in these characteristics between the downburst and typhoon are further compared.(2)Reconstruction of the downburst outflows based on the improved analytical model.According to the spatial distribution of the outflow winds of the multicellular downburst,the traditional empirical analytical model of the microburst proposed by Holmes has been improved to describe the more complex multicellular macroburst.Then the physical model that conforms to the measured macroburst spatial distribution has been established by optimizing the physical parameters of the improved analytical model with the heuristic algorithm.(3)Short-term prediction of downburst winds based on a double-step modification enhanced approach.According to the time series analysis theory,comparative studies of different time series models for short-term wind speed prediction of the downburst have been carried out.From this,intensive studies of correlation and characterization methods between prediction errors and historical wind speeds have been carried out,and then a dual modification enhanced approach for short-term wind prediction based on residual compensation has been proposed.(4)Modal parameter identification and wind correlation study of the long-span cable-stayed bridge under the action of downburst winds.Based on the measured vibrations of the long-span cable-stayed bridge when the downburst occurred in the bridge site,the measured buffeting responses are analyzed in a nonstationary perspective.Assuming the response in a short period is stationary,the modal frequencies and damping ratios of the bridge girder and cable have been identified by random decaying technique,and the evolutionary features of modal parameters with measured wind speeds are further analyzed.(5)Prediction of buffeting responses of the large-span cable-stayed bridge under the action of downburst winds.Based on the wind speeds and buffeting responses measured by SHMS installed on the SCB,the correlations between buffeting responses and wind characteristics are analyzed.Considering the multi-step evolutions of the buffeting responses,neural network models for the buffeting responses prediction have been constructed,which comprehensively considered the historical vibration status of the bridge and wind characteristics in the bridge site,and the reliabilities of different models are further verified by the measured data. |
PDF Full Text Request