Environmental corrosion and external loads degrade the performance of a bridge over the course of its service life. Although dynamic fingerprints are damage-sensitive, they are rarely applied to bridges in-situ due to environmental noise. Machine learning techniques can facilitate effective structural damage detection. This paper proposes a detection method based on dynamic fingerprints and machine learning techniques for multi-damage problems in bridges. Vibration analysis is conducted to acquire the dynamic fingerprints, then the Bayesian fusion is used to integrate these features and preliminarily locate the damage. The RSNB method, which combines Rough Set theory and the Naive-Bayes classifier, is introduced as a robust classification tool for damage qualification. A continuous bridge is numerically simulated to validate the effectiveness of the proposed method. The RSNB method is compared with back propagation neural network, support vector machine, and decision tree techniques, it is found that the RSNB outperforms other three methods in terms of transparency, accuracy, efficiency, noise robustness, and stability.
1 May 2021 in Journal of Natural Gas Science and Engineering
Since the monitoring of cable tension are rather susceptible to environmental influence and external loads, the condition assessment of stay cables is vitally difficult because of these uncertainties. In this paper, regarding the health condition of stay cables, a multilevel assessment framework is presented, which can synthetically combined the evaluation results from numerical simulation, field monitoring and visual inspection. Based on these methods, three qualitative and three quantitative indices are selected as the evaluation indices. To reduce the uncertainties during the assessment procedure, an intelligent methodology based on cloud model and Dempster-Shafer (D-S) evidence theory is proposed. With the combination of forward cloud generator and backward cloud generator, the cloud parameters of in-situ data is transmitted to the cloud model of grade criteria, then the cloud evidence with relative weights are fused by Dempster combination, the condition grade of the cable is finally obtained. The Junshan Yangtze River Bridge is adopted to verify the effectiveness of the proposed methodology. The results show that the uncertainty degree can be obviously reduced from 55.7% to 6.7%, so that a scientific evaluation of cable conditions can be obtained. The multilevel assessment framework proposed in this study can serve as an effective basis for cable replacement and maintenance.