Typhoon Wind Hazard Analysis Based On Numerical Simulation And Fragility Of Light-Gauge Steel Structure In Southeast China Coastal Regions

Global damage and economic loss induced by typhoon are increasing. Enhancing the capability of infrastructure against typhoon is the technological measure to reduce the damage and loss. Both the wind-resistance design and planning for typhoon preparedness require risk analysis of typhoon. In addition, the insurance industry employs risk analysis results to determine their insurance policy and premium. Typhoon wind hazard and fragility are two important components of typhoon risk analysis. Owing to its infrequent occurrence and small size of typhoon, typhoon measured data is scarce, and the fragility of meteorological anemometer during typhoon passage makes the situation even worse, which makes typhoon wind hazard analysis based on the measured data impossible. Currently, the widely accepted practice has been to use a numerical simulation approach to perform typhoon wind hazard analysis.Southeast China coastal regions are frequently hit by typhoon, and there has been no validated method and accumulated data for typhoon wind hazard analysis method in this area. Prevailing light-gauge steel structures in the southeast China coastal regions are wind fragile structures, and there has been little attention on the research of typhoon fragility of this type of structures, especially quantitative results are in demand. With the goal to develop method for typhoon risk analysis for southeast China coastal regions, this thesis is to perform a systematic investigation on typhoon wind field model, probabilistic models for key typhoon parameters, numerical simulation for typhoon wind hazard analysis, and fragility analysis of light-gauge steel structures. The main contents and results of this research are outlined in the following.Based on the typhoon wind field model proposed by Thompson and Cardone (Thompson and Cardone wind field model), the capability of this model in the simulating the historical typhoons in southeast China coastal regions is studied by using the satellite observation data and field observatory data of three typhoons in 2008 and 2010 in Guangdong at several locations. A method to obtain the Holland pressure profile parameter (usually defined as B) is proposed by regression fitting the synoptic data. Several important problems in the validation of the model including the simulation of wind upon landing and sensitivity of the simulated wind speed and direction to the selection of parameters are discussed. The results indicate that there is reasonable agreement between the simulated wind field and the field measurement data for different B models. As for wind speed simulation, the simulation by the B determined by regression fitting the synoptic data produces the best results for the four observations. For the cases without synoptic data, the simulation by setting B=1 gives the smallest errors for the three observations. It is suggested that B model should be determined by regression fitting synoptic data to obtain a reliable pressure profile.The probability density functions (pdfs) of the key parameters of typhoons making landfall at the eleven cities in southeast China coastal region are derived by analyzing the Best Track data. The results indicate that these parameters are similar when the cities are close each other. Furthermore, by analyzing the historical typhoon data, it is found that there exists negative correlation between the radius to maximum wind speed (Rmax) and the central pressure difference, the Rmax and B. The correlation between Rmax and the latitude is very small and can be negligible.A systematic analysis of typhoon wind hazard is done for the southeast China coastal regions based on the Thompson and Cardone wind field model and the established pdfs of key typhoon parameters, and comparison of two simulation schemes and sensitivity analysis of the predicted extreme wind speeds to the Holland pressure profile parameter are performed. The results show that the scheme based on modeling the annual maximum wind speed is logically sound, and the predicted winds fit well with historical observations. The predicted extreme wind speed for Hong Kong using the B model proposed by Harper and Holland is within 5% deviation from that predicted by historical observations. By using the suggested random simulation scheme, the typhoon wind hazard results are obtained for the eleven cities in southeast China coastal region.Based on the wind tunnel test of scaled factory buildings, the fragility of the light-gauge steel structures using the Rackwitz-Fiessler procedure is analyzed. The results indicate that the pdf of area-averaged pressure coefficient is best modeled by shifted Gamma distribution, and the worst wind angle changes with regions and terrains. The fragility (i.e. the conditional failure probability) of structural component under the suburban terrain is less than the one under the open country terrain. The effect of non-Gauss distribution on fragility change with terrains: for open country terrain, the effect is insignificant, and for the suburban terrain, it is significant.