Monitoring Signal Analysis And Processing For Bridges And Time-varying Modal Parameter Identification Based On Hilbert-Huang Transform

One of difficulties in health monitoring for large bridges is to analyze and process the measured responses of bridges objectively and effectively. Because the measured responses are tend to be nonlinear and non-stationary, it is hard to process and explain them by the well developed theories of linearity and stationary. So, it is urgent to develop effective and practical signal analysis and processing methods.Hilbert-Huang transform (HHT) is a newly developed method for time-frequency analysis for nonlinear and non-stationary signals. On the basis of studying the basic theories of HHT, this paper applies HHT to the monitoring signal analysis and processing for large bridges by using the health monitoring system installed on the Nanjing Yangtze River Bridge. And besides , the paper studies the modal parameter identification problem of time- varying structures. By analyzing and processing the measured structural responses more objectively , effectively and scientifically, the inherent mechanism of the complicated and non-stationary responses of bridges is explored and studied.The main contents are detailed as following:(1) Propose a new technique to restrain the end effect——extremashift method, and compare HHT with Wavelet transform and Fourier transform. The end effect is a choke point in applying the empirical mode decomposition (EMD), and it is of great significance for HHT to solve the problem effectively. According to the features of end points and the extrema near the end, four shifted extrema are added to each end of the signal respectively, with the purpose of restrain the end points from divergence. Decomposition results of measured response indicate that the proposed method possesses high calculation efficiency, and can relief the influence of end effect remarkably. The compared analysis shows that HHT has great capability in nonlinear and non-stationary signal processing.(2) Propose a HHT-based method for eliminating the short-time strong noise disturbance existed in signals. The method is based on the fact that instantaneous amplitudes and instantaneous frequencies can show obvious abnormity when a signal contains short-time but strong noise disturbance, and the abnormity is removed by fitting the data at normal section, consequently the corresponding IMF can be obtained with good accuracy. The processed results of numerical simulation and measured signal of the Nanjing Yangtze River Bridge show the good effect of the proposed method.(3) Apply EMD to trend analysis of monitoring signals for bridges, and explore the inherent mechanism of the complicated responses of bridges under train excitations. Firstly, the displacement responses and stress responses of a bridge under given load pattern are calculated. Secondly, the response is decomposed into an oscillation component and a trend component, and the physical meanings of each component are investigated. Finally, based on the measured data obtained from the health monitoring system of the Nanjing Yangtze River Bridge, the inherent mechanism of the responses is further studied. The conclusions can be applied to the practical engineering of health monitoring and condition assessment for bridges.(4) Present a method for modal parameter identification of time-varying structures. Theoretical formulas for identifying the modal parameters using the displacement responses of a time-varying SDOF structure are deduced. Taking advantage of the modal filtering characteristics of EMD, the presented method is extended to identify the modal parameters of time-varying structures with multi-degrees of freedom. By processing the responses of time- varying structures with multi-degrees of freedom, the free vibration response of single mode included can be separated. Simulations results show the validity of the method.(5) Design a time-varying structure experiment to further study the method above-mentioned for modal parameter identification. The experimental device is a cantilever beam. By adjusting the adjunctive mass and stiffness , three kinds of time-varying structures with continuous mass change, continuous stiffness change and sudden stiffness change, respectively, can be realized. Modal parameters of the time-varying structures are identified from the free acceleration responses, and compared with the modal parameters identified under several specified conditions corresponding to time-invariant structures and obtained by theoretical calculation. The results show the feasibility and validity of the proposed method in practice.