Neural Network Identification of the Parameters of Ultra-High-Performance Concrete Bridges

This study focuses on an ultra-high performance concrete bridge, utilizing the natural fundamental frequency and mid-span deflection as input data to construct a backpropagation neural network prediction model.The model inversely identifies two critical material parameters of ultra-high performance concrete -elastic modulus and unit weight -to minimize discrepancies between actual values and design values, thereby establishing a computational model that better approximates the real bridge structure.The results demonstrate the feasibility of the backpropagation neural network prediction model for identifying and correcting structural parameters, with the model exhibiting high prediction accuracy.This approach provides a valuable reference for acquiring actual structural parameters in bridge engineering.