Automated Modal Parameter Identification And Uncertainty Quantification Under Ambient Excitation

Stochastic Subspace Identification(SSI)is a widely used time-domain modal identification method under ambient excitation.This method does not need artificial excitation,and acts directly on the output response data.It has the advantages of simple operation,good reliability and stability.But in the process of modal identification,both the determination of system order and the elimination of false modes need manual intervention,and there are problems of missing real modes and introducing false modes,which makes it difficult to carry out real-time analysis and application of massive bridge health monitoring data,and hinders the development of online health monitoring technology.In addition,it is worth noting that the dynamic test of the structure will be affected by the precision of the instrument,measurement error,ambient noise and other factors,so that the modal parameter identification results have a certain degree of uncertainty,which further affects the accuracy of the structural load identification,damage diagnosis and state assessment results.Therefore,this paper studies the modal parameter automatic identification and uncertainty quantification under ambient excitation.The main contents and conclusions are as follows:(1)The existing researches of modal parameter identification under ambient excitation and automated modal parameter identification method in SSI are clarified briefly,points out some problems existing in this research field,and further summarizes the research contents and objectives of this paper.(2)The basic theories of SSI are introduced.Using the covariancedriven stochastic subspace method to identify the first three vertical modes of Jingyuan Yellow River bridge,and compares the identification results with those based on the Eigensystem Realization Algorithm(ERA).In addition,Bootstrap method was used to quantify the uncertainty of the modal parameter identification results from the overall and local perspectives.The statistical results show that the modal parameters identified have good accuracy and reliability.(3)Taking automated modal parameter identification of 6 degrees of freedom mass-spring model and Jingyuan Yellow River bridge as examples,the application of fuzzy C-means clustering and hierarchical clustering algorithm in automated modal parameter identification is compared and analyzed.The results show that automated modal parameters identification based on clustering algorithm has a good accuracy,and the result is basically consistent with that of ERA algorithm,but the setting of clustering parameters has a certain influence on the result of recognition.(4)Aiming at the uncertainty introduced by clustering and ambient factors,an automatic modal parameter identification and uncertainty quantification method based on Block-Bootstrap and multi-stage clustering is proposed.Firstly,the Block-Bootstrap was introduced to decompose the response signal of the structure into N blocks of data,and then randomly extracted with M times of put back.For M-block data,absolute false modes are eliminated based on soft index threshold,and the stability axis of mamplitude stabilization diagram is automatically picked up by fuzzy Cmeans clustering algorithm,so as to reduce the computation of subsequent hierarchical clustering.Secondly,the hierarchical clustering method is adopted to carry out the secondary clustering of the picked stable axis according to the defined distance threshold,and the false modes are eliminated according to the proposed true and false mode discrimination index MDI to obtain the initial modal parameters.Finally,the above steps were repeated B times,and the defined discrimination index was used to modify the initial modal parameters of group B.The mean value of elements in the cluster was taken as the recognition result of modal parameters,and the accuracy of the recognition result was measured by standard deviation.The results of numerical simulation and modal parameter identification of Jingyuan Yellow River bridge show that the proposed modal discrimination index has better effect of identifying true and false modal compared with the threshold value of traditional index.Also,the proposed method can eliminate the uncertainty introduced in the clustering process and the influence of ambient noise,improve the identification accuracy and have better anti-noise performance.