Research On In-service Condition Assessment Methods Of Bridges Using FBG-based Strain Monitoring Data

Bridge health monitoring system can provide authentic and credible operational information of bridge structures and monitor the structural responses under a variety of loads. The long-term strain monitoring data obtained by the health monitoring system can directly reflect the local stress state of bridge structures under various loads, and the service status of bridge structures can be effectively evaluated by tracking the strain changes of bridge structures. In this thesis,the in-service condition assessment methods of bridge structures are proposed based on the strain monitoring data in combination of wavelet multi-resolution analysis, Kalman filter, kernel principal component analysis, and support vector machine regression, which are employed to Jiubao Bridge monitoring data for multi-component data analysis, bridge structure damage identification, bridge structure performance assessment, and safety pre-warning. The main research works are:(1) The purpose and significance of the in-service condition assessment of bridge structures based on strain monitoring data are discussed. The analytical methods for processing of strain monitoring data as well as the methods of damage detection,performance assessment and safety pre-warning based on strain monitoring data are summarized. The research ideas of this thesis are given.(2) A method for multi-component data separation is proposed based on wavelet multi-resolution analysis. The feasibility of the proposed method is verified by analyzing the field monitoring data. The accuracy and effectiveness of the proposed method are discussed by using correlation analysis.(3) A method of damage detection is proposed based on Kalman filter and neutral axis position. The feasibility of the neutral axis damage index is verified by numerical simulation. The neutral axis position of the Jiubao Bridge is determined based on the high-frequency components separated by wavelet multi-resolution analysis. The Kalman filter method is used to filter the neutral axis position of Jiubao Bridge for damage detection of bridge structures.(4) A method of assessing the performance of bridge structures is proposed based on load resistance factor. Based on the field monitoring data, the changes of load resistance factor under traffic load and different weather conditions are compared and analyzed. The dangerous position of Jiubao Bridge steel box beam is determined by the comparison analysis of performance evaluation of the same section and different sections.(5) Based on the kernel principal component analysis and support vector machine regression algorithm, the method of safety pre-warning of bridge structures is proposed. In combination of the two algorithms, the abnormal data extraction model is established. The validity of the proposed method is verified based on the strain monitoring data of bridges.