Vibration source identification method based on multisensor vibration signal fusion analysis

The accelerated development of urban areas has resulted in an increase in environmental vibration factors in the vicinity of buildings. For buildings, particularly ancient ones, vibrations that exceed their limits can potentially cause irreversible damage. The identification of the source of the vibrations and ensuring that the identification process and results are physically meaningful is necessary to implement timely and targeted measures to mitigate the impact of vibration and noise on the building structure. We proposed a vibration source identification method based on the correlation analysis of multi-source vibration response signals. This method uses two parameters, correlation coefficient and signed time delay, to describe the similarity and directionality of different vibration response signals propagating through various parts of the structure. Moreover, we innovatively integrate these two parameters into a new indicator vibration propagation correlation coefficient (VPCC), which could simultaneously quantify the similarity and direction of the propagation of different vibration response signals, thereby achieving vibration source identification through the quantification of intervals. With the intuitive index construction, the proposed method achieves clear physical interpretability while ensuring low computational complexity, providing an efficient and physically insightful solution for vibration source identification. The results of validation demonstrated that this method is an effective means of identifying single load vibration sources, whether originating from wind or construction activity. However, its efficacy is somewhat diminished when applied to mixed load vibration sources, consisting of a combination of wind and construction. In these instances, the overall identification accuracy reached 96.7 %.