Model Updating And Damage Identification Based On Health Monitoring Of Super-large Steel Box Girder Cable-stayed Bridge In Northern Cold Region

In recent twenty years, The Number of long-span cable-stayed bridge constructionin our country at the world leading edge. The structure of the actual state of bridgebridge stress condition and operation condition is a hot topic at present largecable-stayed bridge research. This paper studies some key scientific and technicalproblems of large cable-stayed bridge, The research contents include: health monitoringsystem’s data extraction data filtering of structure health monitoring system, the damagealarming, structural modal parameter identification, finite element model establisheand the model updating. The main research contents are as follows:First, summarized the characteristic of measuring point arrangement and dataacquisition of Liaohe bridge health monitoring system.Based on the health monitoringsystem, I developed a set of method for massive data processing suited for all kinds ofbridge.Second, put forward a kind data filtering method, Which is suited for differenttypes of sensors. The method is first to recursive weighted average the Original data，Then filter out the processed data by wavelet analysis to abandon the high frequencypart save low frequency part of the data.Finally, reconstructed the low-frequency data,to obtain steady state data.The method can effectively filter out the noise data causedby white noise and accidental factors.Third, by using the large general finite element software ANSYSand the TongjiUniversity Dr. bridge software to establish the model of Liaohebridge and analyzed thetwo calculation results By comparison of the bending moment, shear, axial force anddisplacement of the static analysis and ANSYS analysis results is similar to bridgedoctor. Dynamic analysis results accord with the actual rules of the bridge aswell.Through the establishment of the finite element model provides guidance for themaintenance of liaohe super major bridge, also for bridge damage identification andsafety evaluation and early warning and laid the foundation..Fourth through the use of health monitoring system of super major Liaohebridge a flat steel box girder cross section’s temperature and stress has been monitored, we havestudied the steel box girder temperature load generated by the temperature effect inCold region. We have analysed temperature and longitudinal stress daily change trend,the overall variation and the effect of temperature on the longitudinal stress. Analysisresults indicate that: the maximum value of the steel box girder roof temperaturecalculated by the specification of JTG D60-2004is less than the actual monitoring value;steel box girder continuous temperature changes of24h obey sinusoidal distribution,and strain subject to linear distributions for the started6hours, and subject to gaussiandistribution curve for the last18hours; According to the analysis results of two bigsoftware extracted bridge steel box girder stress accumulation and actual data for healthmonitoring of bridges, get real stress change.Fifth, I use the rain flow meter method to translate strain gauge data into one byone stress amplitude andStatistics it. Father more I have analyzed the fatiguedamage of steel box girder roof and combined with the Britishfatigue specification.Finally the author puts forward a kind of massive data processing method which isbased on the bridge temperature and support displacement correlation acquisition ofhealth monitoring system, The method is use wavelet analysis to filter and refactor origndata, Then set up the correlation model between reconstruction displacement data andtemperature data.Finally use refactor data to construct the BP neural network and buildup the model of the artificial neural network.