Feedback and Sensitivity Analysis for Mass Data from Transverse Joints Monitoring System of Super Arch Dam in Sino Mainland

The fuzzy cluster analysis for monitoring data, based on SOM neural network, was applied to study the working behaviors of transverse joints of the super arch dam during construction phase. The parental samples were identified with the classification results of fuzzy-cluster researches. Thereby, the Elman network was used to diagnose test samples. Introducing the transverse joints monitoring data from super arch dam as an example, the results show that the proposed method has superb applicability and stability in both of error analysis and diagnosis accuracy.