Development of deformation prediction system of large bridge based on MATLAB and radial basis function neural network

Affected by many factors, deformation of large bridge will take place in the course of usage, so the development of deformation prediction system will contribute to the maintenance and management of large bridge. As the multiple-input and multiple-output characteristic of the RBF neural network, which also has the nature of universal approximation and optimal approximation, MATLAB combined with RBF network was choosed to develope the deformation prediction system of large bridge. The improved algorithm of RBF network parameters was studied, some key issues in the system development were solved, the choice of input factors and the preprocessing method of sampled data were discussed, the main function of the system was elaborated. This system has some popularization and application values to the deformation prediction and safety supervising of other similar projects.