隨機子空間識別方法在結構損壞預警之應用

The Application of the Damaged Structural Prediction by Stochastic Subspace Identification

唐嘉明(Chia-Ming Tang)

指導教授 ： 羅俊雄

國立臺灣大學/工學院/土木工程學研究所/碩士(2010年)

https://doi.org/10.6342/NTU.2010.02478

摘要

無資料

關鍵字

結構健康監測 ； 隨機子空間系統識別 ； 子空間識別 ； 遞迴式方法 ； 新奇指數 ； 預警系統

並列摘要

With the progress of signal processing technologies, structural health monitoring (SHM) has received more and more attentions. The damage detection method is proposed for structural health monitoring under varying environmental conditions. In the recent decades, the researchers already developed numerous system identification methods to monitor and detect if the damage occurs. A literature survey of system identification methods is examined in this research. It is found the structural health monitoring should possess a high-speed computation with reliable results and with accurate capacity of estimation.
In this research, the following structural system identification methods are introduced: (1) stochastic subspace identification, (2) subspace identification, (3) recursive stochastic subspace identification, (4) recursive subspace identification. Based on the results of identification the novelty indicator from Kalman filter is developed for damage estimation.
This research presents the application of these identification and damage detection methods through three experimental tests. In the 1st example, the test data was obtained from a series of shaking table test of a RC frame which is developed by NCREE (National Center for Research on Earthquake Engineering). And in the 2nd example, a bridge scouring test data was used to identify the change of system vibration characteristics due to scouring. The 3rd example is using the ambient vibration data of a RC frame to determine the vibration characteristic of the structure after a series of strong ground motion excitation back to back. The recursive SI and recursive SSI methods are applied to extract the dynamic characteristics of the structural response. The 3rd example confirms that recursive SSI can be applied to the practical situation for the system health monitoring in almost real-time. It is concluded that with a suitable design of model parameters in RSI or RSSI methods, the results can be used for early warning system.

並列關鍵字

Structural health monitoring ； Stochastic subspace identification ； Subspace identification ； Recursive technique ； Novelty index ； Prediction warning system

參考文獻

1. Ai-Min Yan, Pascal De Boe and Jean-Claude Golinval, Structural damage diagnosis by kalman model based on stochastic subspace identification, Sage publication, Vol 3(2):001-119, 2004.

2. A.-M. Yan, G. Kerschen, P. De Boe, J.-C. Golinval, Structural damage diagnosis under varying environmental conditions – part 1: a linear analysis, Mechanical systems and Signal Processing, 2005.

3. Bart Peeters, System identification and damage detection in civil engineering, December 2000.

4. Bart Peeters and Guido De Roeck, Reference-based stochastic subspace identification for output-only modal analysis, Mechanical Systems and Signal Processing 13(6) 855-878, 1999.

5. C.C. Chang and Z. Li, Recursive stochastic subspace identification for structural parameter estimation, Proc. of SPIE Vol.7292, 2009.