Font Size: a A A

Research On Sturctural Health Monitoring System Of Cable-Stayed Bridges Based On Wavelet And Fractal Theory

Posted on:2009-03-27Degree:DoctorType:Dissertation
Country:ChinaCandidate:C R YaoFull Text:PDF
GTID:1118360245989045Subject:Bridge and tunnel project
Abstract/Summary:PDF Full Text Request
Currently, the structctural health monitoring (SHM) of long-span bridges is one of the reseach heat points in the field of bridge engineering. Also, SHM is a very difficult task which is related to many subjects and industries. Because of complexity and indetermination in bridge structures and environmental conditions, the research on the SHM is still at the exploring and tryout phase. Nowadays, structural health monitoring systems have been installed on some long-span bridges. However, these systems are mainly used for data acquisition, lacking an effective health assessment support. Based on characteristics in the wavelet and fractal theory for complex nonlinear issues, a basic frame of the bridge health monitoring system is suggested. Following research works have been carried out in this doctoral dissertation.1. A baseline finite element model that can reflect structural responses comprehensively and correctly is necessary to the health monitoring system. Defining the displacement by static tests and the natural frequency by dynamic tests as state variables (SVs), and taking the objective function as the sum of square of deviation ratios of each test-values and corresponding calculating values, adopting the optimization methods in ANSYS, the initial analysis model of long-span bridge is updated by modifying the calculation parameters in this dissertation. In updating, changing ranges of parameters and results are limited by constraint conditions; therefore, the updated parameters have respective explicit physical meanings. The updated finite element model provides a more practical mathematical model for the bridge health monitoring, and also validates the feasibility of updating model methods by ANSYS optimization design.2. Dynamic monitoring is an important way for the health monitoring system, analyzing the structural dynamic characters is necessary at the beginning of building up the health monitoring system. In this dissertation, the effects by coupling vibration of vehicle-bridge system, road surface roughness, wind and temperature are synthetically considered, and practicable datas for establishing the health monitoring strategy are offered.3. Based on the wavelet theory, the reasons of the frequency aliasing in the Mallat algorithm and frequency sub bands disordering of the wavelet package are analyzed profoundly. And an improved algorithm to avoid frequency aliasing and frequency sub bands disordering is presented.4. Applying wavelet transform, differential equation of multi-scale structural dynamic is obtained, which provides a solid theoretical foundation to structural health monitoring system. With application examples, it is shown that the wavelet analysis is very effective to solve structural modal parameters, denoise signal, compress and reconstruct datas, and detect signal singularity in bridge health monitoring system.5. Applying wavelet transform to the displacement mode shape, curvature mode shape and stress mode shape of cable-stayed bridge, the results show that this method can be used well for diagnosing the position and degree of sturcutral damages.6. The improved wavelet algorithm is applied to the signals of health monitoring. The result shows that the wavelet energy, like structural natural frequency, can be used to reflect the structural inherent characteristic and local characteristic as well. Based on the analysis of wavelet package, the concept of wavelet energy cross entropy is proposed to serve as an index for damage diagnosis. By adopting weighted value which represents the local characteristic of structure, the index can diagnoses the location and the degree of damage effectively.7. Based on powerful ability of the wavelet in signal local analyzing and the fratal in nonlinear processing, a structural damage monitoring method using correlation dimension of wavelet details is proposed. Firstly, the vibration detection signal is decomposed into varying scales of temporal resolution, then the correlation dimension of wavelet details of key frequency bands is calculated. thirdly, cross entropy of the correlation between undamaged and damaged stucture is carried out, which is a diagnosed index. This proposed approach is fully tested with examples, and the results show it possesses sensityivity to damage and strong immunity to noise.8. Applying the general fractal dimension to diagnose the damage of cable-stayed bridge, it is shown that the sensitive dimension varies directly as degree of damage, and the cross entropy between undamaged and damaged structure as well.9. A basic frame of health monitoring system for cable-stayed bridge is built up, which integrates with the wavelet and fractal theory.
Keywords/Search Tags:cable-stayed bridge, health monitoring, model updating, wavelet analysis, energy of wavelet package, entropy, fractal, correlation dimension, general dimension
PDF Full Text Request
Related items