Research Of Decentralized Modal Parameter Identification Based On Wireless Smart Sensor Networks

Structure modal parameter identification is a major part of the structural dynamic recognition, which plays an important role in the structural health monitoring system. For large civil structural health monitoring, a large number of wireless sensor is necessary to arrange to ensure accuracy of parameter identification. However, the wireless data communication will be thousands of increments caused by the increase of nodes, which is difficult to realize for centralized data processing. The application of distributed hierarchical network, allowing mutual coordination data processing of wireless sensor and data integration of sub-structures nodes, can significantly reduce the wireless data communication and has a great advantage. In this paper, Decentralized in the wireless sensor networks combined with random decrement technique of structure modal parameter identification is carried out to study the random decrement technique, substructure modal integration and distributed modal parameter identification of improved random decrement/ITD, then the feasibility and effectiveness of the algorithm in wireless sensor networks are also compared and discussed. The current studies and results are summarized as following:In view of nonstationary ambient vibration, it introduces a kind of modified random decrement technique based on Brown motion function. By a simulation example, it's verified that the modified random decrement technique combined with time-domain ITD method can accurately identify structure modal parameters in incentive for the environment, having more general applicability. It discusses the different triggering conditions on the influence of the method. In the fusion of substructure modal shapes, it proposes Recursive Least Squares Method fusing substructure mode shapes, and then through a seven degree of freedom simulation model of the spring and cable-stayed bridge model, it's verified the correctness of the algorithm, and it is concluded that Larger groups with overlapping nodes are found to reduce errors in Larger groups with overlapping nodes were found to reduce errors in the assembled global mode shape. As a14across truss model for the simulation object, it uses the modified decentralized RTD/ITD algorithm to identify the modal parameters comparing with the decentralized NExT/ITD algorithm, discussing the accuracy of structure modal parameter identification and the wireless data communication of the truss simulation model in differences of trigger condition and differences of topology. It is concluded that the division of different topology has much effect on the wireless data communication. when the number of substructure nodes is the different and the number of overlapping nodes is the same, the more the number of substructure, the less the wireless data communication; when the number of substructure nodes is the same and the number of overlapping nodes is the different, the less the number of overlapping nodes, the less the wireless data communication. In the conclusion, the modified decentralized RTD/ITD of extremum triggering condition is the most suitable for modal parameter identification of the truss model. It not only can accurately identify modal parameters under nonstationary ambient vibration, but also compares the other triggering condition's modified decentralized RTD/ITD to reduce the wireless data communication.