The Applied Research Based On Wavelet Analysis In Bridge Health Monitoring

Bridge health monitoring is a combination of modern sensor technology,communications technology and evaluation system, which can provide real-time onlinemonitoring on the structural response in the environment and under the loads during thebridge operation process, and analyze the health status of bridge according to variousmonitoring indicators obtained, and send out timely warning in case of structural abnormality,to provide a scientific basis for the management and maintenance and repair of the bridge. Forthe existing and potential problems in damage and early warning during the bridge healthmonitoring, wavelet transform was adopted in this paper as the basic analytical tool, andapplied in the bridge health monitoring, thus presenting more systematic study in the bridgestructure test data de-noising, modal parameter identification, damage identification, warningsituation analysis, with research results having a certain of engineering application value.Due to a variety of factors, the actual bridge test data often contain noise data, whichgreatly affects the structural damage identification and status assessment, therefore,controlling noise from test data fundamentally, to reduce the impact of noise interference ondata in the structural analysis, is a viable approach. Firstly, the principle of de-noising bymeans of wavelet transform modulus maxim and wavelet threshold was expounded, and thena simply supported beam structure modal shape was made as a one-dimensional signal,decomposition and reconstruction was conducted on the signal in these two de-noisingmethods under different noise levels, and according to the de-noised signal-to-noise ratio SNRand root mean square error RMSE these two indicators, it was found that the two methodscould significantly reduce the noise strength of modal shape signal, and through comparativeanalysis, it was pointed out that the wavelet threshold de-noising was more simple andefficient than the wavelet transform modulus maxim de-noising, which is better suited to thede-noising of bridge health monitoring data.In this paper, it was explored that how the complex Morlet wavelet transform was usedin the modal parameter identification of structure under ambient excitation. Wavelet transformwas conducted in the structural dynamic response signal, and according to the informationcontained in the time-scale diagram of wavelet transform coefficients, its modulus and phasewere further obtained, so as to achieve the identification of structural modal parameters. Thenumerical simulation of a simply supported steam imposed with steady-state white noiseexcitation and an actual environment analysis vibration test were analyzed, to verify the validity of the method, and point out that when the structural modal parameters are moreintensive, to improve the resolution of the Morlet wavelet and then make the analysis canproduce good results.Based on modal identification, the feasibility and suitability of the wavelet coefficientdifference of structural curvature modalDd ias damage identification index are analyzed andresearched in details. The different damage degrees of simple structure and of complexstructure are analyzed under single damage mode and multi-damage mode. Mass analysisresults indicate that: the modals to structure damage identification index will be graduallyreduced with intensification of damage degree; as for boundary damage element, signalextension method is raised for boundary extension of curvature modal so as to obtain integraldamage identification index information of boundary element; in terms of multi-damage modeof complex structure, the damage has different influences on structure characteristicparameters of different parts, thus, it is easy to cause wrong judgment or leak judgment inoverall identification whenDd idamage identification index is used so that the stageidentification of complex structure based onDd idamage identification index is put forward.In addition, it is verified that the stage identification can avoid wrong judgment and leakjudgment to a large extent in the analysis. The analysis also shows that the noise robustness ofDd iindex will be enhanced after the parameters are processed in wavelet thresholdde-noising method.The test identification error of high level modal of the structure is relative big, whichmay bring certain influence on damage identification. Given application limitation ofDdiindex, a new damage identification index, namely, damage identification indexC diofflexibility curvature wavelet coefficient difference, is proposed for reference. The analysis onvalue and simulation example indicates that it is only required to form low level damageindex of the structure whenC diis used as damage identification index and structurevibration frequency differs much; for complex structure, the front modals with approximatevibration frequency are required to structure damage identification index; there is good noiserobustness of this index. On the basis of analysis and research on two damage identificationindexes, the selection principle ofDd idamage index andC didamage index are concludedin respect of damage degree and mechanical properties of the structure.According to analysis on the benefit——cost, purposes and functions of healthmonitoring system, design and establish the first health monitoring system——health monitoring system of one super major bridge, which is the first example put into service, andintroduce its composition and characteristics. Deeply explore warning situation analysismethod of bridge health monitoring and mention that the warning situation analysis consistsof damage warning situation analysis and operation warning situation analysis. The fullprocess analysis of damage warning situation of one super major bridge is conducted and thefeasibility of recommended damage identification system is verified.Research the monitoring data analysis of parameter index which can reflect bridge statusbased on wavelet transform, take bridge strain and deflection curve as one-dimensional signal,conduct multilevel wavelet decomposition of the signal and obtain general status tendency ofbridge structure in monitoring period through analysis on wavelet coefficient; analyzeeffective information of wavelet coefficient to reconstruct deflection curve signal of one supermajor bridge and combine ARIMA model of time series to build prediction model ofdeflection index so as to know operation status of bridge structure in a future period, which isquite significant for the monitoring and warning of structural operation status.