Damage Identification of Simply Supported Bridge Based On RBF Neural Network

Efficient non-destructive damage detection procedures for bridge structures have become an important research topic. In this paper, the damage location is detected using modal shape curvature, while the damage severity is identified based on RBF neural network. A numerical example for a simply supported beam bridge with five girders is provided to verify the feasibility of the method. The contrast analysis between RBF and BP neural networks is conducted to confirm the superiority of RBF network. The results shown that the convergence speed of RBF is faster than BP, and the RBF network also possesses more favorable damage identification results.