盲訊號識別法於結構健康診測之實驗研究

Experimental Study on Structural Health Monitoring using Blind Source Separation Method

陳聖夫(Sheng-Fu Chen)

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(2015年)

https://doi.org/10.6342/NTU.2015.01420

摘要

大部分情況下，結構物之受力情形難以量測，因此唯輸出 (Output-only) 系統識別方法近年來發展迅速，本研究目的為探討二階統計盲訊號識別法 (Second Order Blind Identification, SOBI) 於土木結構系統識別及損壞診斷的應用範疇。
本研究第一階段旨在檢測二階統計盲訊號識別法 (SOBI) 應用於非時變性與時變性系統上的表現並與隨機子空間識別法 (Covariance Driven Stochastic Subspace Identification, SSI-COV)比較其優缺點。本文將二階統計盲訊號識別法 (SOBI) 應用於數值模擬及實地沖刷試驗上：(1) 不同車輛載重下之車橋互制系統(非時變性) (2) 橋樑沖刷實驗之損壞診斷與預警之應用(時變性系統)。識別出動態參數後並利用不同損壞指標診斷出破壞位置。
本研究第二部分旨在檢測二階統計盲訊號識別法在輕微非線性系統上的表現。二階統計盲訊號識別法 (SOBI) 分別被應用在兩棟三層樓鋼構架上。試體一為對稱鋼構架，利用二階統計盲訊號識別法 (SOBI) 識別出動態特性並結合 ABAQUS CAE 執行動態分析預測應力，除此之外還可利用識別出的模態貢獻及應力判斷系統是否進入非線性並與遲滯圈比較。試體二則是替換第一層樓西北角之柱子為弱柱之反對稱試體，由於是反對稱結構，除了X方向模態還激發出Y方向及扭轉方向模態，另外還發現結構激發出雙重模態(Double modes)的情形，最後利用識別出的模態貢獻及應力判斷系統是否進入非線性。

關鍵字

盲訊號分離 ； 二階統計盲訊號識別法 ； 系統識別 ； 結構健康監測 ； 損傷檢測 ； ABAQUS CAE

並列摘要

In most civil structures, the input excitation is hard to be measured. Thus, the output-only system identification has gained lots of interest. In this research the application of output-only system identification technique known as Second Order Blind Identification (SOBI) algorithm in civil structures is carried out.
The objective of the first part of this thesis is to inspect whether SOBI algorithm is workable on time-invariant and time-variant system when subject to either white noise excitation or earthquake excitation and compared the results to Covariance Driven Stochastic Subspace Identification (SSI-COV). SOBI algorithm is applied on: (1) Vehicle-bridge interaction system under different traffic loading (time-invariant system). (2) Steel bridge under continuous scour (time-variant system). Through the results of identification the dynamic characteristics of the system are verified and damage detection using different damage index is also investigated.
The objective of second part of this thesis is to inspect the identification ability of using SOBI algorithm when system turned into lightly non-linearity. A shaking table test of two 3-story steel structures are carried out (one is the symmetric structure and the other is the unsymmetrical structure). For specimen 1, dynamic characteristics identified by SOBI algorithm are combined with ABAQUS CAE to estimate the local property, such as member stress. For specimen 2, additional lateral and torsion modes are identified using SOBI and the result is compared with the subspace identification technique.

並列關鍵字

Blind Source Separation ； Second Order Blind Identification ； System Identification ； Structural Health Monitoring ； Damage detection ； ABAQUS CAE

參考文獻

[1] Bernal, D., Gunes, B., "Flexibility Based Approach for Damage Characterization: Benchmark Application," Journal of Engineering Mechanics, Volume 130, Issue 1, January, 2004.

[2] Belouchrani, A., Meraim, K. A., Cardoso, J.-F. and Moulines, E., "A blind source separation technique based on second-order statistics, " IEEE Transactions on Signal Processing, Volume 45, Issue 2, February. 1997.

[3] Chen, S.F., Hung, T.Z. and Loh, C.H. "Application of traffic-induced vibration and damage detection by blind source separation with application to bridge monitoring," SPIE 9435, Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2015, March 27, 2015.

[4] Chao, S.H., Loh, C.H. "Output Only Dynamic System Identification Using Blind Source Separation Technique," Proceedings of the 5th World Conference on Structural Control, Tokyo, Japan, July, 2010

[5] Cardoso, J.F., "Blind source separation: statistical principles, "Proceedings of the IEEE, Volume 86, Issue 10, August, 2015.