A study on the extraction of the abnormal behavior point using Hilbert-Huang transform

The recent constructed structures are featured by the combination of their functions and shapes as well as by their enlarged dimensions, which increase the demand for Structural Health Monitoring technology. Since every structure bears unique dynamic characteristics and is exposed to diverse external forces, various methods are studied to monitor the health of the structure. This study applies the Hilbert-Huang transform, the variance analysis and the edge detection method on the acceleration response of the structure to identify the initiation time of the abnormal behavior in which the structure experiences abnormal vibration. A scaled cable-supported bridge model is fabricated and subjected to cable failure test from which data before and after the occurrence of the abnormal behavior are acquired and compared to validate the proposed anomaly-extraction technique.