Improved modal identification using wireless continuous dynamic monitoring systems without real time synchronization

Abstract

In the context of vibration based Structural Health Monitoring using automated operational modal analysis, data synchronization methods permit a reduction in the cost of wireless monitoring systems, since the use of extra hardware for perfect data synchronization can be avoided. Furthermore, the relaxation of data synchronization requirements permits a more flexible topology (higher distance between sensors, no need of external antennas for data synchronization using GPS, lower power consumption), which is particularly advantageous in the dynamic monitoring of large structures. In this paper, continuous dynamic monitoring without real-time synchronization is studied. Firstly, using numerical simulations, the effects of synchronicity faults on operational modal analysis is deeply analysed, showing that time delays have a great impact on modal identification. Secondly, a new data synchronization method based on mode shapes identified with operational modal analysis methods is proposed to improve the accuracy of modal identification. This does not need the assumption of proportional damping. Therefore, it overcomes this limitation that up to the knowledge of the authors is present in all other available methods. The proposed processing strategy is validated in the processing of monitoring data continuously collected at a long span arch bridge (Infante D. Henrique Bridge) with two wireless acceleration nodes. The full-scale application shows high time synchronization accuracy and proves that the proposed synchronization method is adequate for continuous dynamic monitoring, providing reliable modal parameters estimates with any assumptions about the damping, permitting a reduction in the monitoring costs, easier installation and power consumption reductions (very important to increase the autonomy of wireless nodes). In short, hardware costs are replaced by a smarter processing.