Impact of time synchronization error on the mode-shape identification and damage detection/localization in WSNs for structural health monitoring.

Time synchronization in wireless sensor networks (WSNs) is a critical challenge for any distributed system such as WSNs for structural health monitoring (SHM). In SHM, mode shape identification, damage detection and damage localization are sensitive to time synchronization errors (TSEs). Indeed, the errors, due to the time shift between the incoming raw data from each sensor node, may hugely affect the data integrity and then the mode shape identification of the structure under analysis. In this paper, we characterize the impact of TSE on the modal analysis, damage detection and damage localization using frequency domain decomposition (FDD) implemented in a semi-local manner. In order to decrease the size of the transmitted data by the sensor nodes and reduce the processing load and the needed storage capacity on the central unit, we adopt a semi-local processing approach where each sensor node partially processed data and transmit it to a central unit for further processing such as mode shape identification, damage detection and damage localization. We adopt the model where each sensor node performs the Fast Fourier Transform (FFT) of the measured vibration signal and the transmission of the FFT values to a central unit or to a cluster head for further processing. The results show that TSE has a strong impact on the mode shape identification, damage detection and damage localization. Furthermore, results show that semi-local processing is more sensitive to TSE compared to centralized processing.