利用地震反應資料進行結構全域及局部性損害評估

Global and Local Damage Assessment of Structures by Seismic Response Data

林裕家(Yu-Chia Lin)

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(2007年)

https://doi.org/10.6342/NTU.2007.00489

摘要

這篇論文主要在探討兩種不同的損害檢測方法評估全域性和局部性結構損害。本文中的傷害評估方法大致可分三個步驟：第一是利用量測資料識別狀態空間的結構系統矩陣，並利用結構系統矩陣轉換出動態的物理參數，第二是利用 DLV方法進行全域性損害檢測，第三是使用模態振形曲率的變化量並定義傷害指數評估局部性結構性損害。
狀態空間識別( SRIM )是利用量測資料識別結構系統矩陣，並且利用識別的結構系統矩陣識別二階系統的動態結構矩陣。就傷害檢測而言，結構物( 線性系統 )常被分為破壞前與破壞後兩階段。本論文首先利用四個數值模型為例進行全域性與局部性結構損害檢測，這四個例子分別為：( 1 )三個自由度的剪力構架，( 2 )二維桁架結構，( 3 )兩跨度梁的有限元素模型及( 4 )三層樓房屋構架的有限元素模型。
此外，上述方法也應用於實驗資料。本篇論文選用在國家地震中心的實尺寸三層樓結構進行實驗資料分析，此結構可分為兩種不同的型態：一個是在基準結構一樓加設支撐模擬健康狀態並且移除支撐模擬破壞狀態；另一種是基準結構一樓的柱子進行翼板切削以模擬破壞狀態。由於感測器安裝與數量上的限制，損壞樓層的位移和應變測量資料也被用來識別系統模態振形以進行局部性結構損傷檢測。

關鍵字

系統識別 ； 傷害檢測

並列摘要

This study presents two different damage detection algorithms: one is for global damage detection and the other is for local damage detection. For damage detection three steps are proposed: (1) the state-space model identification from input-output measurements to extract the physical parameters, (2) global damage detection using damage locating vector (DLV) method, (3) using the changes of curvature of mode shapes as damage index to localize the position of structural damage for more detail local damage detection.
For state space identification, SRIM method is used to identify the system matrix from input-output measurements and the second-order structural system matrix are also identified from realization model. For damage detection, the pre- and post- damaged structures are identified using the two separate equivalent linear models. In this study four numerical models are selected: (1) a three degree of freedom shear-type Structure, (2) a two-dimensional truss structure, (3) a two span continuous beam represented by finite element model and (4) a three-story finite element model, as examples for global and local damage detection in this study.
Besides, these damage detection methods are also applied to experimental data. Shaking table test of two different benchmark models for SHM research in NCREE is selected for the experimental study. One of the benchmark models is the 3-story steel frame with adding brace at its first floor to simulate the healthy state and removing bracing to simulate the damaged state; the other is the wedge-type cut on the flange of the column at first floor to simulate the damaged state. Because of the limit number of sensors in the implementation, the displacement and strain measurements are also used to extract the mode shape of damaged floor for local damage detection through the usage of damage index.

並列關鍵字

System Identification ； Damage Detection

參考文獻

[1] Alvin, K. F., and Park, K. C., ‘‘Second-Order Structural Identification Procedure Via State-Space-Based System Identification,’’ AIAA J., Vol. 32, No. 2, pp. 397–406. 1994.

[2] Alvandi, A., and Cremona, C., “Assessment of vibration-based damage identification techniques,” Journal of Sound and Vibration Vol. 292, pp. 179–202, 2006.

[3] Betti, R., Lus, H., Franco, G., Imbimbo, M., and Yu, Y., “Identification of strucural damage using dynamic input-output measurements,” Proc. of SPIE, Vol. 6174, 61741K, 2006.

[4] Bernal, D., “Load Vector for Damage Localization,” Journal of Engineering Mechanics, Vol. 128, No. 1, pp. 7-14, 2002.

[5] Bernal, D., and Gunes, B., “Flexibility Based Approach for Damage Characterization: Benchmark Application,” Journal of Engineering Mechanics., Vol. 130, No.1, pp.61-70, 2004,