Crack detection of arch dam using statistical neural network based on the reductions of natural frequencies

This paper presents the numerical simulation and the model experiment upon a hypothetical concrete arch dam for the research of crack detection based on the reduction of natural frequencies. The influence of cracks on the dynamic property of the arch dam is analyzed. A statistical neural network is proposed to detect the crack through measuring the reductions of natural frequencies. Numerical analysis and model experiment show that the crack occurring in the arch dam will reduce natural frequencies and can be detected by using the statistical neural network based on the information of such reduction.