Temperature field prediction for a PC beam bridge with corrugated steel webs using BP neural network and measured data

Due to the significant difference in specific heat capacity between concrete materials and steel, the prestressed concrete (PC) beam bridge with corrugated steel webs (CSWs) exhibits considerable non-uniform temperature fields under the combined effects of solar radiation, ambient temperature, and wind speed. To elucidate the distribution characteristics of non-uniform temperature fields in PC beam bridges with CSWs in service and improve the computational efficiency of temperature fields, this study proposes an efficient solution method for structural temperature fields based on the Back Propagation (BP) neural network. In this paper, a numerical model is established to simulate the heat conduction, radiation, and convection processes in PC beam bridges with CSWs. The temperature field characteristics of the box girder structure under the combined influence of solar radiation intensity, air temperature, and wind speed are analyzed and validated against measured data. Based on this, a BP neural network surrogate model is constructed to rapidly and accurately solve the temperature field of corrugated steel web PC girder bridges. The research results indicate that the BP neural network structure exhibits minimal errors between the predicted and actual temperatures of steel plates and concrete in composite girder bridges. The average coefficient of determination is 0.9985, indicating that the temperature prediction for concrete is superior to that for steel structures.