利用改良型基因演算法於實際建築物之系統識別及初步健康診斷

Hybrid Genetic Algorithm to System Identification and Preliminary Damage Assessment of Real Buildings

蔡惠萍(Huei-Ping Cai)

指導教授 ： 王 淑 娟

朝陽科技大學/理工學院/營建工程系/碩士(2009年)

摘要

系統識別首先是應用於航空工程與機械工程，後來才被引用至土木工程上。系統識別是利用數值方法分析系統產生之輸出入量測以建立系統之模型的一種技巧。藉由系統識別之技巧除了可預測結構物未來可能產生之反應外，亦可監測系統參數或特性之變化，甚至根據這些變化了結構受損之情形，決定後續必要之處置。近二十年來，很多學者從事結構健康診斷之研究。所謂結構健康診斷是利用比較結構目前之狀態與未受損狀態或基線狀態(Baseline state)之過程來決定結構物受地震或其他型式外力作用下是否有損壞產生，從而決定損壞之位置及程度。
自民國八十二年起迄今，交通部中央氣象局地震測報中心之強地動觀測計劃，業已於臺灣地區完成結構物監測系統六十一座，其中四十四座為建築物，十七座為橋樑。截至921集集大震前已蒐集了為數相當可觀之地震紀錄。惜地震之強度不夠大，結構本身尚未進入非彈性範圍。921集集大震發生後，上述監測系統已蒐集了為數相當可觀之強震紀錄。
本文之主要目是以兩棟實際結構的地震紀錄來進行識別，其一為了解台電大樓主樓經長期地震過後，結構物雖無明顯之損壞，但振態參數可能會有變化或是結構可能有肉眼觀測不到的輕微損壞，利用單向擾動模式，進行該結構物在歷次地震之振態參數識別，因此本文將利用識別之振態參數結果，了解該結構物受長期間地震侵襲下，進行初步損壞評估。其二在中興大學方面又分為兩部分，第一部分利用單向擾動模式，進行中興大學土木環工大樓在歷次地震之振態參數識別，該大樓在921集集大地震中雖有受損，但為了解該大樓經長期地震過後，振態參數之變化程度，進行該結構物之初步損壞評估。
第二部分則針對921集集大地震進行系統識別，該地震規模較大，結構物已受損，研判該結構物已進入非線性反應行為，因此利用線性系統模式是無法識別合理之參數值，所以本文利用多個樓層反應當作輸出的等值線性系統模式進行分段識別，以了解每段所識別之振態參數變化程度，進行初步評估受損狀態之依據。

關鍵字

系統識別 ； 基因演算法 ； 損壞評估

並列摘要

Field of system identification has become important discipline due to the increasing need to estimate the behavior of a system with partially known dynamics. Identification is basically a process of developing or improving a mathematical model of a dynamic system through the use of measured experimental data. In addition to updating the structural parameters for better response prediction, system identification techniques made possible to monitor the current state or damage state of the structures.
Since 1993, the Central Weather Bureau has installed 61 Strong-Motion Systems on various kinds of structures, including 44 buildings and 17 bridges in Taiwan area. However, the data collected from the accelerographs installed on buildings before the Chi-Chi earthquake are still remained in elastic range since the intensities of the earthquakes before the Chi-Chi Taiwan earthquake, are not strong enough to trigger the inelastic response. The damage of the buildings induced by the Chi-Chi earthquake, provides a solid evidence that some of the structures has experienced inelastic response. As a consequence, this thesis is focused on the system identification of linear as well as nonlinear structural systems of buildings, using the data collected from the accelerographs installed under the Taiwan Strong-Motion Instrumented Program.
In the first part, it is intended to perform system identification of linear system to two real buildings using measured response data collected during real earthquakes. The first building identified is the Taiwan Electricity Main Building. This building seems to experience no visible damages under the attacks of the earthquakes. However, the dynamic parameters may be altered even though the damage of the structure is slight or invisible. To get knowledge of this damage state of the structure, single-input-single-output model is utilized to perform the identification of modal parameters of the structure. As a result, by monitoring the variation of the identified parameters, the preliminary damage assessment of the structure is performed and the damage state of the structure is evaluated. The second structure identified is the Civil-Environment building of Chung-Shin University, which experience moderate damage under the struck of the Chi-Chi earthquake. Again, the same model is utilized to perform the modal identification of the structure. The preliminary damage assessment of the structure is performed and the damage state of the structure is then evaluated.
In the second part, system identification of the second structure subjected to the Chi-Chi Earthquake is performed. Since the intensity Chi-Chi Earthquake is comparatively larger, the response of the structure is expected to experience a non-linear behavior. Consequently, model of a linear system may not be suitable to describe the behavior of the structure. To overcome this problem, the time history of the measurement is divided into a series of overlapping time intervals. Then, the model of equivalent linear system is employed to identify the modal parameters of the system for each time interval. Finally, the variation of the modal parameters can be monitored and the damage state of the structure can be assessed.

並列關鍵字

System Identification ； Genetic Algorithm ； Damage Assessment

參考文獻

【1】 McVerry, G. H., “Structural Identification in the Frequency Domain from Earthquake Records,” Int. J. of Earthquake Engineering and Structural Dynamics, Vol. 8, pp. 161-180(1980).

【2】 Beck, J. L. and Jennings, P. C., “Structural Identification Using Linear Models and Earthquake Records,” Int. J. of Earthquake Engineering and Structural Dynamics, Vol. 8, pp. 145-160(1980).

【10】 Koh, C. G. and Liaw C.Y., “Substructural and Progressive StructuralIdentification Methods,” Engineering Structures, Vol. 25, pp.1551-1563(2003).。

【11】 Jeong, I. K. and Lee, J. J., “Adaptive Simulated Annealing Genetic Algorithm for System Identification,” Engineering Applications of Artificial Intelligence, Vol. 9, No. 5, pp. 523-532(1996).

【12】 Koh, C. G. and Chen, Y. F. and Liaw C.Y., “A Hybrid ComputationalStrategy for Identification of Structural Parameters,” Computers and Structures, Vol. 81, pp.107-117(2003).