Research on Time-varying Reliability of Cable-stayed Bridge Based on Kriging Model Improved by U-learning Function
Fretting fatigue and steel wire corrosion are important factors affecting the safety of the cable structure of long-span cable-stayed bridges.Because it is difficult to directly obtain the structural function under corrosion damage,in order to explore the effect of corrosion fatigue coupling damage on the performance of the cable structure of existing cable-stayed bridges,a time-varying reliability analysis method of cable of cable-stayed bridges based on u-kriging considering the coupling damage state is proposed.Based on the principle of steel wire corrosion and fretting fatigue,the strength degradation law of in-service stay cables is derived,and the structural function of stay cables considering different damage modes is established.The Kriging algorithm improved by U-Learning function is used to fit the response characteristics of cable-stayed structures under different damage states.Taking a long-span in-service double tower cable-stayed bridge as an engineering example,the degradation trend of cable reliability index under different damage modes is discussed,and the correlation between some influencing factors and cable reliability index is analyzed.The results show that:compared with the initial Kriging model,the improved Kriging model with U-Learning function significantly improves the fitting accuracy of the response surface of cable-stayed structure,and the root mean square error of u-kriging model is reduced by 54.18％compared with the standard Kriging model;the results show that the corrosion damage will greatly accelerate the strength degradation rate of the stay cable,and the reliability index of the stay cable considering the coupling damage is about 1.0 lower than that of the stay cable considering only fretting fatigue damage after 50 years of service,and the time when the reliability index reaches the limit is about 10 years earlier;the effects of single damage and coupling damage on stay cables at different locations are different,and targeted maintenance measures should be considered in the operation stage.
cable stayed bridgestay cablescorrosion fatiguekriging modelU learning functionreliability analysis
万方数据
