Wind-induced Structural Performance Assessment Considering Uncertainty Based On Wind Field Measurement

With the rapid growth of economy,numerous major engineering projects including the super-tall buildings and ultra-high voltage transmission lines have been established which are highly sensitive to strong wind load.Located in complex environment,the large-scale projects are vulnerable to severe weather conditions under strong wind disasters.Therefore,to ensure the operational safety of the major projects it is economically significant to carry out the real-time estimation of the structural performance and timely warning before the potential disasters.The dissertation took a super-tall building and transmission line located in coastal area that are prone to typhoons as the research objects.Combining with field measurement,numerical simulation and theoretical analysis,the following studies involving the wind characteristics,modal identification,reliability assessment and warning were carried out:1.Strong wind turbulent characteristics analysis based on site measurement:A full-scale monitoring system for the transmission tower-line system was deployed to obtain the wind field in the southeast hilly coast of China.Based on the measured wind data,the stationary and Gaussian characteristics of the wind field is analyzed.Combining the stationary test method,a self-adaptive processing algorithm is proposed to deal with the non-stationary wind speed and then the S transform technique is applied to estimate the evolutionary spectrum.Considering the non-Gaussian characteristics of measured wind,a refined gust factor calculating method is provided based on the peak factor theory.2.WRF-based hybrid neural network for short-term wind speed prediction under typhoon condition:Based on the numerical weather prediction model WRF,a WRF-based hybrid prediction method is extended by integrating the data decomposition algorithm and neural network model.Targeting the measured typhoon wind speed,the best hybrid neural network method is selected from two data decomposition methods,wavelet decomposition and empirical mode decomposition,and four neural network models,BP,Elman,GRNN,ANFIS models.The optimized method is further utilized to predict the extreme wind speed under typhoon condition.3.Modal parameter identification and its improvement:Several typical methods of modal parameter identification,including the fast Bayesian FFT method,random decrement technique and stochastic subspace identification method are improved.Assuming that the wind load spectrum approximately follow the exponential form in a certain frequency band,the fast Bayesian FFT method is refined to identified the modal parameters under wind load excitation.By introducing the Gabor transform,two kinds of Bayesian identification methods are extended to infer the time-varying parameters,namely,fast Bayesian Gabor transform method and Bayesian evolution spectral density method.Based on the multivariate empirical mode decomposition(MEMD),a multi-dimensional signal modal parameter identification method based on MEMD is proposed.The improved methods were applied to different numerical examples and the results demonstrated that the improved methods are more robust and accurate than the primal methods.4.Modal parameter identification of high-rise buildings and transmission line:Based on the full-scale monitoring system,the dynamic responses of K11 Building in Hong Kong,Canton Tower and transmission line in Zhoushan Island are obtained under different excitations.Then by using the proposed identification methods,the dynamic characteristics of the monitored projects under seismic,typhoon and environmental noises are identified.Furthermore,utilizing modal parameter identification results under environmental and wind excitation,the aerodynamic damping of Canton tower and the transmission line is detailed analyzed.5.Wind-induced reliability analysis of high-rise buildings considering uncertainty:Based on the full-scale measured data,the typhoon field surrounding the building is simulated and further used to carry out the wind-induced vibration reanalysis.The Monte Carlo simulation method is used to quantify the structural parameters and wind load uncertainty in the structural reliability calculation by generating the samples of annual maximum wind speed and structural parameters.Then the wind-induced response is computed to assess the wind-vibration reliability of the high-rise building during the operational stage.Considering the dispersion of human sensitivity to vibration,fuzzy theory is introduced to characterize the uncertainty of comfort limit,and an integrated structural reliability assessment algorithm is established.6.Wind-induced fragility updating for transmission line based on Bayesian theory:Based on the full-scale measured wind field and wind-induced response of transmission line,a Bayesian framework to update the dynamic fragility is proposed for transmission line by integrating the Bayesian model updating,modal parameter identification and multi-scale wind simulation techniques.The framework is applied to the measured transmission line to update the wind-induced fragility after suffering from two super typhoons.The difference of wind-induced dynamic performance between operational and design stage of transmission lines is analyzed.Following the fragility framework,the influence of the wind spectrum and the environmental temperature on the wind swing of transmission lines is further discussed.