Real-Time Structural Health Monitoring and Damage Detection

Abstract

Structural health monitoring (SHM) contains continuous structural vibration monitoring, extraction of damage sensitive features of structure from measurements, and statistical analysis of those features to detect and locate the damage in structures. In other words, SHM involves data collection, continuous monitoring and analyzing them in real-time. Changes in modal properties of structures during their service life are strongly related to damage in structures, which makes accurate estimation of modal properties an essential step in SHM. Therefore, both monitoring and accurate identification of real-time modal properties (modal frequency, damping ratio, and mode shape) are of crucial importance in order to have a good estimate in SHM. In this paper, a real-time modal identification techniques along with a damage detection algorithm based on inter-story drift calculation has been developed for SHM. The modal identification technique is based on the modification of standard spectral analysis tools for real-time data, and utilizes running time windows to keep track of time variations of structures' modal properties. On the other hand, the damage detection algorithm makes use of inter-story drifts, which is calculated by narrow-band filtering the recorded data around modal frequency and very sensitive to structural damage, by estimating the contribution of each identified mode of structure. A software package called REC_MIDS is developed for real-time modal identification. The software includes various user-selectable algorithms to identify modal properties, as well as options to plot their time variations and animations. The software has been tested with the ambient vibration data recorded from the Hagia Sophia Museum, a 1500 year-old historical structure in Istanbul, Turkey. Modal properties of the structure have been identified accurately in real-time. Results of the Hagia Sophia test have been compared with the previous studies conducted by different researchers. Comparison shows that the results of the REC_MIDS are in good agreement with that of the previous studies.