Structural Safety Assessment of Bridge By Integration of Wavelet and D-S Evidential Theory

The high nonlinearity and nonstationarity has resulted in the huge challenge in structural safety assessment of bridge in the health monitoring by using the traditional method. In this paper, a novel technique is introduced to solve the problem in the structural safety assessment of bridge by the integration of the wavelet and the D-S evidential theory. In the context, the effects of live-load deflections and strains of the main girder are firstly extracted from the total measurement by wavelet analysis, and then they are used in the primary safety assessment associated with the change of inherent frequency of the bridge by using artificial neural networks. And futher the basic probability assignments are constructed according to results of the primary assessment through the relative artificial neural network. Finally, the integrated assessment of the bridge is completed by the information fusion based on the D-S evidential theory. The result from the application of a real bridge indicate that the proposed method is a promising way in the safety assessment of large scale bridge in the future. Keywords: bridge, safety assessment, wavelet, D-S evidential theory