LAUSR

Abstract Modal identification has been playing an important role in structural health monitoring. Traditional modal identification process requires too much user interaction and many manually defined thresholds. In this study, a new fully automated modal identification approach is proposed to automatically identify the modal parameters through automatic interpretation of stabilization diagram. The proposed approach works efficiently without any manually tuned threshold or index except some widely used thresholds, is robust to the mere input, and can be used to identify the closely spaced modes. A modified Fuzzy C-means clustering algorithm is proposed to automatically interpret the stabilization diagram, without any specifically tuned index or threshold but only with a default maximum clustering number. The optimal cluster number and desired clustering result can be obtained through an iterative intelligent graph partitioning algorithm. The boxplot is introduced in the last step to assess the precision of the clustering results from a statistical perspective, and outliers can be found and eliminated. The proposed approach is applied to analyze the real monitoring data of a long-span suspension bridge and a benchmark footbridge with closely spaced modes to validate the feasibility, and parametric analysis is also conducted to assess the robustness. It can be concluded from the results that the proposed approach can automatically extract the modal parameters, including the closely spaced modes, with a high accuracy and robustness. The proposed approach provides a reliable and promising method for analyzing the monitoring data from long term structural health monitoring.