Extreme response analysis of train-track-bridge-wind interaction system based on in-situ monitoring wind data

•The randomness of track irregularity and wind turbulence is considered to evaluate the extreme value distributions of high-speed train-bridge interaction system under crosswinds.•To obtain the accurate wind parameters used in the extreme analysis, the turbulence characteristic is analyzed based on the long-term wind field measurement near the bridge site.•The applicability of PDEM based virtual time process method to the extreme responses evaluation of train-bridge interaction (TBI) system is validated by comparing with Monto Carlo Method.•The Weibull distribution model (Type 3 of the generalized extreme value distribution) can be used for the extreme value evaluation of TBI system. This conclusion may be useful to the other similar analysis about extreme responses of TBI system. Deterministic safety curves of safe train running speeds versus mean wind speeds (MWSs) have been studied in the time domain, considering the train-bridge interaction (TBI) dynamics. However, the randomness of track irregularity and wind turbulence makes the extreme response of the TBI system indeterministic; and so does the safety curve. Therefore, an extreme evaluation theory based on the probability density evolution method (PDEM) is adopted in this study to investigate the extreme response distributions of TBI system under crosswinds with the consideration of the field measured turbulence characteristics. The applicability of the theory on train-bridge wind interaction system is verified first, followed by the discussion on the variations of extreme responses of train and bridge versus MWSs and running speeds of trains running on a high-speed railway cable-stayed bridge. A probabilistic safety curve is finally determined based on 40 cases with the wind speeds ranging from 10 m/s to 45 m/s and the train speeds ranging from 200 km/h to 400 km/h. The study analyses demonstrate the computational accuracy and efficiency of the PDEM-based extreme theory applied to the TBI system; and indicate that the extreme responses of high-speed trains and Quanzhou Bay Bridge increase with increasing MWSs and train speeds. In addition, most of performance indicators do not exceed the code-stipulated failure criteria except the train unloading factor and the lateral displacement at the mid-main span. The probabilistic safety curve for the studied bridge is thus deduced according to these two controlling indexes.