Study On The Methods For Online Calibration Of Bridge Structural Monitoring Systems

The whole life cycle state monitoring of bridge structures is an effective means to ensure traffic safety and an important trend of modern transport development.Calibration of the measurement characteristics of state monitoring systems,such as basic error,linearity and repeatability,is the basis to ensure the accuracy and reliability of the systems.Due to the requirements of all-time and all-weather monitoring service,the online calibration under the condition of “zero disassembly,zero interruption and zero interference” is the most effective means to realize the measurement traceability of the monitoring system.Therefore,this paper conducts the theoretical and experimental researches to build the online calibration model and to establish the methods for measuring characteristics analysis and evaluation,which are suitable for passive excitation source and the complex environment.The main research works are summarized as follows:(1)An online calibration method of structural monitoring system is proposed based on data series analysis under common excitation source.A calibration model is established,by which we can obtain online calibration results via reference data series,comparison data series,and time-space distribution of the variables.With the proposed methods,the uncertainty of measurement characteristics caused by“incomplete controllable conditions” in online calibration is reduced through“sample redundancy”.The in-situ verification shows that the theoretical model is consistent with the engineering practice,laying the foundation for online calibration of the structural monitoring systems.(2)An experimental model for online calibration based on Measurement Objects,Excitation Source,Reference System(RS),and System Under Calibration(SUC)are designed and verified.Under the stimulation of the common passive excitation source,the physical response of the target structure is taken as the measured object,and the original data sequence is collected by SUC and RS in real time.The online calibration results are obtained through the algorithms of time series matching and statistical analysis of measurement characteristics.The feasibility of the model is verified by simulation experiment and actual engineering.(3)A data series matching algorithm based on merging characteristic points sectioning with Symbolic Aggregate appro Ximation(SAX)is proposed.This algorithm adopts the symbolic representation of discrete low-dimensional data to represent the morphological characteristics of typical time serie sub-segments,and is applied for automatic searching and rough matching of initial positions,which solves the technical problem of matching data series of different scales.Then,using the local peak and valley values of the data serie as feature points,the RS and SUC data series are sectioned and precise matching is carried out to obtain the synchronous measurement data pairs of the two sensing systems.The experiment shows that the algorithm effectively solves the mismatch problem of SUC and RS data with respect to the same excitation source.(4)Based on the statistical characteristics of long-term online monitoring data sources,the characterization,calculation and evaluation methods of key measurement characteristics of structural monitoring systems are proposed.These methods operate on the basis of the morphology similarity between SUC and RS data series collected under common excitation.By partitioning the data series to blocks and screening the key data blocks on the edges,a reference line is established with least square regression.The proposed methods make it possible to make up for the uncertainty from uncontrolled conditions through mining in the large sample data the measurement system features.The experiment shows that the standard deviation of the basic error calculated from the random sample data is 0.7%,which verifies the feasibility of the proposed methods.(5)The verification experiment based on Jiujiang Bridge in Guangdong province was carried out,which takes two kinds of strain sensing system as an example to verify the online calibration process,including test system setup,measurement characteristics calculation,and uncertainty evaluation.The results show that the proposed method is valid for the online calibration of structural monitoring systems.