Wireless Synchronous Measurement Error Analysis And Modeling Of Mechanical Vibration

Traditional wired structure health monitoring systems are difficult to apply to health monitoring,damage detection and location of large structures due to factors such as complex equipment installation and high maintenance costs.Therefore,structural health monitoring systems based on wireless sensor networks have received extensive attention.Considering the monitoring problem of mechanical vibration,the main focus of this thesis is the wireless vibration acquisition system.Due to the small vibration amplitude and high frequency,high-precision acquisition of sensors is particularly important.Considering the inherent uncertainty of wireless sensor networks,it is difficult for wireless clock synchronization algorithms to completely eliminate clock synchronization errors between different nodes.Based on signal processing,this thesis analyzes the effects of temporal jitter and space jitter on the identification of modal complex frequency,modal residue and mode shape by the mode decomposition in complex frequency domain and the properties of Laplace transform.The main contributions of this thesis are as follow: 1)Using the unified eigendecomposition method,the derivation and collation of the modal decomposition of the undamped system,the viscous proportional damping system and the general viscous damping system are completed.The characteristics and proof process of the mode shape of different damping type systems are detailed.The derivation process of the frequency domain mode decomposition is supplemented in consideration of the needs of modal analysis and testing.2)Starting from the transfer function matrix,the influence of time jitter on the identification of modal cmomplex frequency,damping factor and other modal parameters is analyzed by using the property of scale transformation.Starting from the relevant conclusions of complex frequency domain mode decomposition,the transfer function matrix based on spatial jitter sampling is calculated,and the residue of the transfer function is calculated to study the influence of spatial jitter on the identification modal complex frequency and mode shape.3)The steps of least squares complex frequency domain algorithm,stability graph algorithm and the identification algorithm of residue in frequency domain based on single reference are introduced in detail.The basis functions in the algorithm are modified for discrete time series.At the same time,we have made corresponding adjustments to the parameters related to the PLSCF algorithm for the mismatch caused by the spatial jitter between the data and the parametric model.4)The thesis have completed the implementation of the algorithm on the MATLAB platform.By comparing the recognition result of the algorithm with the given theoretical value,the validity of the algorithm and the rationality of the above theoretical analysis results are verified.