ABSTRACT
According to the Technical Condition Assessment Standard for Highway Bridges and the influence of special climate on bridge disease development in dry cold regions. 14 core indicators were selected as the judgment basis affecting bridge reliability, and 40 sets of small and medium-span bridge inspection data were collected as the base data of the model. Principal component analysis was used to reduce the dimensionality. After the training of the RBF neural network optimized by the particle swarm algorithm, the reliability level prediction model of small and medium span bridges in the dry cold regions was established. Finally, compared with the traditional RBF neural network and PSO-RBF neural network model, it can largely improve the network operation speed, the mean square error is small, and the multiple prediction results tend to be stable.
