Research of deflection data compensation by applying artificial neural networks

A great number of data have been obtained for a long time in the data acquiring subsystem of bridge health monitoring system. The data correctness will affect the bridge safety-judging result directly. In this paper, the characteristics of deflection data is analyzed, the definition of the correlation is brought forward, the relation among checking points is dredged up. According to the none-linear approximation of radial basis function (RBF) neural networks, the RBF neural network data prototype is established for compensating the abnormal data. The data acquired from the Dafosi Yangtze Bridge in Chongqing have been used to verify this prototype, the results show that it has good functional approximation and generalization ability.