Study on regional temperature actions of railway steel-concrete composite girder bridges in China

This paper investigates regional differences in temperature actions on railway steel-concrete composite girder bridge in China. A refined thermal-structural coupling finite element model (FEM) of a typical Chinese railway steel-concrete composite girder was established and validated. Using 40 years of meteorological data from major cities (A stations), numerical simulations were performed, and each resulting two-dimensional temperature fields was decomposed into four temperature actions based on derived formulas. Daily extreme values of these temperature actions were extracted from the decomposed data sets. Representative values of temperature actions with a 50-year return period for A stations across China were then determined using extreme value analysis theory. Additionally, numerical simulations were conducted using 2021 data from 401 meteorological stations, and the corresponding annual extreme values of temperature actions were extracted. To capture the nonlinear relationship between the spatial parameters and the annual extreme temperature action value of the steel-concrete composite girder, these extreme data and their corresponding spatial parameters (longitude, latitude and altitude) were utilized to train a neural network model. Using this neural network model, annual extreme temperature actions data were obtained, and isotherm contour maps of the extreme temperature values for 2021 were then plotted. Based on these 2021 temperature action isotherm maps and a BP neural network model, representative values of temperature actions across various regions in China were determined, and corresponding isotherm maps were plotted. The results indicated significant differences in the representative values of the four-temperature actions within China. The temperature actions obtained in this paper exhibit enhanced geographical resolution compared to existing codes. Using these refined temperature action values, the mean normal strain of the railway steel-concrete composite beam section and the sectional bending curvature around the x-axis and y-axis can be quickly calculated. © 2024 Elsevier Ltd