應用頻率域改良型基因演算法與遞迴式改良型基因演算法於結構動力系統識別

Frequency-Domain Hybrid GA and Recursive Hybrid GA to Structural Dynamic Parameter Identification

蔡永勤(Yong-Cin Tsai)

指導教授 ： 王淑娟

朝陽科技大學/理工學院/營建工程系碩士在職專班/碩士(2011年)

摘要

目前台灣地區在各地重要之大樓、學校及橋樑等土木建築物上皆裝有強震儀，可隨時記錄地震發生時之相關資料，利用其地震紀錄來識別對應之結構系統參數，據此進行建築物安全評估，並根據地震後建築物的破壞情形，進行修補工作。過去幾十年已發展出許多系統識別方法，這些方法可應用於往後類似建築物之系統識別上，便於快速了解該建築物受震後之結構系統特性，甚至根據這些參數變化了解結構受損之情形，決定後續必要之處置。近二十年來，很多學者利用系統識別的結果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)之過程來決定建築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。
為了解決時間域改良型基因演算法在運算反應上需花費較多的時間。因此本研究首先提出頻率域改良型基因演算法，頻率域改良型基因演算法主要是利用傅立葉轉換將地震紀錄轉換至頻率域上求解，頻率域上運算僅需要利用代數方式進行運算，相較於時間域上利用微分方式進行運算較為快速，故能大幅減少運算時間。在此將應用於數值模擬系統之動力特性識別，驗証其可行性；為了更加確定應用於真實建築物之可行性，亦對於含有雜訊的數值模擬地震紀錄及反應進行識別以探討新式識別法之識別效果。最後將本方法應用於真實大樓-台電大樓，並且比較時間域改良型基因演算法所識別之結果。
不論是時間域或頻率域改良型基因演算法皆無法應用於非線性系統上，因此本研究接著提出遞迴式改良型基因演算法應用到非線性系統識別，遞迴式改良型基因演算法是利用分段方式進行識別，其識別結果可看出當建築物受到地震作用影響時之參數變化。本法之驗證也是類似於頻率域改良型基因演算法之作法，是利用數值模擬系統與含有雜訊的數值模擬地震紀錄及反應驗証其可行性。最後將遞迴式改良型基因演算法應用於真實大樓-國家地震工程研究中心與台東消防分隊大樓，並且利用損壞評估的方式來判斷建築物是否有損壞。

關鍵字

系統識別 ； 改良型基因演算法 ； 損壞評估

並列摘要

Currently, in various important buildings, schools, bridges and other civil buildings in Taiwan are equipped with Strong-motion seismographs. It can recording the relevant information when the earthquake occurred. We can use those data to identify the system parameters of structure and building’s damage assessment. And the structural reinforcement is also according to the assessment result. Over the past few decades there are already developed many system identification methods that can be applied to buildings system identification and quickly to understand the post-earthquake system characteristics. In recent years, many scholars use the results of system identification to do the structural health monitoring research. Structural health monitoring is comparative the structure parameters or baseline state before and after earthquake to determine the location and extent of damage due to earthquake.
The time-domain hybrid genetic algorithm has to take more time to count the response. In order to accelerate the identification process, hybrid GA in the frequency domain is developed. That is using Fourier transform to convert the response, because frequency-domain operations only need to use algebraic approach, compared to time-domain differential approach to computing is more rapid, it can significantly reduce the computation time. This will be applied in the numerical simulation of the dynamic characteristics of the system to identify, verify its feasibility; used in order to be more realistic to determine the feasibility of the building, but also for the numerical simulation of the earthquake with the noise and response records to identify new identification method to study identification. Finally, the method is applied to real building - Taiwan Electricity Main Building. The comparison is made between the results in the frequency domain and the time domain.
Either time-domain or frequency-domain hybrid genetic algorithm, nonlinear systems are not used. In order to implement the hybrid GA to nonlinear system, the time history of the measurement is divided into a series of time intervals. Then, the model of equivalent linear system is employed to identify the modal parameters of the system for each time interval. this research and then propose a recursive application of hybrid genetic algorithm to nonlinear system identification, recursive hybrid genetic algorithms method is used to identify sub-way, the recognition results can be seen when the building affected by the earthquake when the parameters change. Verification of this law is similar to the frequency-domain hybrid genetic algorithm approach is the use of numerical simulation system and containing the numerical simulation of seismic noise records and the reaction verify its feasibility. Finally, recursive hybrid genetic algorithm applied to real building - National Center for Research on Earthquake Engineering, and Taitung Fire Brigade building, and algorithm is applied to these structures and the damage indices are then computed according to the identified parameters. By monitoring the variation of the identified parameters, the damage assessment of these structures is performed and the damage states of these structures are evaluated.

並列關鍵字

Hybrid GA ； System Identification ； Damage Assessment

參考文獻

【35】 林沛暘、羅俊雄，「標竿鋼構樓房震動台試驗」，國家地震工程研究中心報告，NCREE-06-020，臺北，2005。

【1】 McVerry, G. H., “Structural Identification in the Frequency Domain from Earthquake Records,” Int. Journal 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. Journal of Earthquake Engineering and Structural Dynamics, Vol. 8, pp.145-160, 1980.

【3】 Yun, C. B., and Shinozuka, M., “Identification of Nonlinear Structural Dynamic Systems,”Structural Mechanics, Vol. 8, No. ST2, pp.187-203, 1980.

【7】 Chaudhary, M.T.A., Abe, M., Fujino, Y., and Yoshida, J., “System Identification of Two Base-Isolated Bridges Using Seismic Records,” Journal of Structural Engineering, Vol. 126, No. 10, pp.1187-1195, 2000.