Bridge Structural On-Line Modal Parameter Identification

In order to meet the real-world engineering requirement of long span bridges, long-term online modal parameter identification, using three recursive modal parameter identification algorithms to identify the girder, tower and stay cable modal parameters of certain long span cable-stayed bridge were performed. After the comparative analysis of identified results derieved from different recursive modal parameter identification methods, the specific recursive modal parameter identification method which is more feasible in the real-world engineering is identified. Then, the time-varying characteristics of different members were analyzed by identifying their modal parameters. The main research in this thesis are listed as follows:(1) The research status of the modal parameter identification of stochastic subspace algorithm, the signal subspace recursive tracking theory algorithm, recursive stochastic subspace space theory is reviewed in detail, then the real-world engineering needs for a long-term online modal parameter identification of large span bridge is analysed.(2) The projection approximation subspace tracking algorithm (PAST), the instrumental variable approximation subspace tracking algorithm (IV_PAST), the expand instrumental variable approximation subspace tracking algorithm (EIV_PAST) are systematically introduced one by one. Based on the three subspace tracking algorithms and stochastic subspace algorithm, three kinds of signal subspace recursive estimation algorithm theories are expounded. At the same time, the computational code based on MATLAB platform for the three kinds of recursive stochastic subspace space modal parameter identification methods are presented.(3) The modal parameters of the girder, pylon and stay-cable are identified by the three kinds of signal subspace recursive estimation algorithms as mentioned above for a certain period of time i.e. one month data, to plot the time frequency diagrams. The computational time and the number of identified clear stable modal order are taken as a criterion, to select the most suitable method for different large cable-stayed bridge components.(4) Then using the most suitable recursive modal parameter identification methods for the large cable-stayed bridge of the different components by identifying their respective modal parameters and then draw the time-frequency characteristic diagram of each month during the whole year. Then to select the month whose modal tracking result is better and plot the control chart, thus the operational status of the components can be figured out from the control charts.(5) Studies have shown that the time required for the three recursive modal parameter identification methods:PAST based recursive modal parameter identification algorithm is the least time consuming, then is the IV based recursive modal parameter identification algorithm, the EIV based recursive modal parameter identification algorithm is the most time-consuming. Different components can use different recursive modal parameter identification algorithm, based on the its suitability criterian, and the operational status is in good condition.