Comparison of deterministic subspace identification methods for model bridge shaking table test(PDF)
长安大学学报（自然科学版）[ISSN:1006-6977/CN:61-1281/TN]
- Issue:
- 2019年04期
- Page:
- 73-81
- Research Field:
- 道路工程
- Publishing date:
Info
- Title:
- Comparison of deterministic subspace identification methods for model bridge shaking table test
- Author(s):
- ZHOU Xiaohang; SHAN Deshan; TAN Kangxi; LI Qiao
- (School of Civil Engineering, Southwest Jiaotong University, Chengdu 610031, Sichuan, China)
- Keywords:
- bridge engineering; modal parameter identification; shaking table test; deterministic subspace; cablestayed bridge model
- PACS:
- -
- DOI:
- -
- Abstract:
- In order to investigate the performance of the deterministic subspace identification (DSI) and the combined deterministicstochastic subspace identification (CDSI) in practical applications, a comparison study of DSI and CDSI were carried out. First, the algorithm principles of DSI and CDSI were introduced, the different ways of considering the process noise and measurement noise were compared, and the advantages in the algorithm of CDSI were discussed. Next, a shaking table test of a largescale cablestayed bridge model was used as a case study, the acceleration responses of the model bridge under four different degrees of Chichi seismic excitation were obtained, and data were prepared for investigating the performance of DSI and CDSI. Then, the DSI and CDSI were conducted to identify the modal parameters of the model bridge. A stabilization diagram was used to present the modal parameter identification results and the number of stabilization points was counted, the stability of the identification results of the two methods were compared. Finally, the mode shape identification results of DSI and CDSI were exhibited and the modal assurance criterions (MAC) between their mode shapes were calculated, and the differences of the mode shape identification results of the two methods were compared. The results show that the identification results of frequency and damping ratio of DSI and CDSI are consistent, however, the identification results of CDSI possess a higher stability. The amplitudes of mode shapes of DSI are slightly greater than those of CDSI, and the amplitude of the mode shapes of the two methods are approximately the same after normalization processing. However, the difference between the mode shapes of the two methods increases with the increase of the mode order and excitation degree. Process noise and measurement noise are considered in CDSI, and its modal parameter identification results are slightly better than those of DSI. However, considering noise reduces the computational efficiency of CDSI, and thus, the advantages of CDSI do not seem significant. 2 tabs, 12 figs, 25 refs.
Last Update: 2019-07-29