Multi-parameter monitoring and pattern recognition applied to cable-stayed bridges model

A cable-stayed bridge model was established, and a multi-parameter experimental analysis on the bridge model was made. After feature extraction, the feature vectors were processed by applying information fusion technique. Based on the self-organizing neural network and the topography search method in knowledge processing, the patterns of deteriorated condition of the model were clustered on a two-dimensional plane automatically by three steps initializing-calibration-identifying.