Analysis of Temperature-induced Deflection of Cable-stayed Bridge Based on BP Neural Network

Abstract

In order to study the relationship between ambient temperature and girder deflection quantitatively, realize the deflection prediction, a BP neural network deflection prediction method based on correlation analysis is proposed in this paper. The correlation between ambient temperature and girder deflection is analysed, and the BP neural network method is used to fit the samples with non-linear correlation quantitatively. Based on the quantitative relationship between ambient temperature and girder deflection, the prediction of girder deflection is realized. Taking Nanjing Yangtze River 3rd Bridge as an example, the feasibility of this method is verified based on monitoring data for four consecutive years. The results show that the non-linear mapping relationship between girder deflection at mid-span and ambient temperature is accurate and has good prediction effect. The method proposed in this paper provides a basis for the evaluation and early warning of the deflection.