Research on the Elevation Prediction Model of Suspended Continuous Beam Bridge based on PSO-BP Neural Network

Abstract To address the linear prediction problem encountered during the construction of continuous beam bridges with cantilever erection, an elevation prediction model based on PSO-BP combined neural network model was introduced. Model integrates Pearson correlation analysis, Generalized Matrix Inverted Grey Prediction (GM (1,1)) model, Particle Swarm Optimization (PSO) algorithm and Back propagation (BP) Neural Network. Through a literature review on past bridge data and Longxia Taojiang Bridge, predicted elevation values were compared with field measured values. Results demonstrated that predicted values of PSO-BP model are in coincidence with the measured values, meeting accuracy requirements. Therefore, this proposed model can serve as a valuable reference for the cantilever erection of other continuous beam bridges.