Temperature Effect Analysis of Bridge Natural Frequency Based on Particle Swarm Optimized Neural Network

In order to eliminate the temperature effect on the tested natural frequency of bridge structures, a new method for temperature effect analysis of natural frequency of bridge was proposed based on the particle swarm optimized (PSO) neural network. First, the major frequency orders affected by temperature were derived from the tested multi-order natural frequencies through correlation analysis. Second, the PSO neural network algorithm was adopted to analyze the measured data, and the relationship between temperature and natural frequency was established, then the method on eliminating temperature effect was derived. Finally, the correctness and reliability of the presented method was verified by a three-span suspender arch bridge.