指導教授：羅俊雄臺灣大學：土木工程學研究所吳宗翰Wu, Tzung-HanTzung-HanWu2014-11-252018-07-092014-11-252018-07-092014http://ntur.lib.ntu.edu.tw//handle/246246/260756在結構健康監測及損害識別當中，作用在結構物上的外力扮演著重要的角色。因此，外力識別的技術被提出以了解作用在結構物上的外力。外力識別的演算法 為外力識別技術的核心。本文中介紹了幾種外力識別的演算法，包含了修正回歸平方卡爾曼濾波器法(modified recursive square Kalman filter method)、未知外力卡爾曼濾波器法(unknown input Kalman filter method)、降階觀測器法(reduced order observer method)以及理論柏努力梁方法(heoretical Bernoulli beam method)。為了驗證這些外力識別演算法，本文中建立了四個模型來模擬不同的結構物及外力情況。此外，本文中也將外力識別的技術運用到了實際的案例及實驗當中。首先，外力識別技術被運用到簡支梁試驗當中來驗證外力識別技術的實用性。接著，一個六層樓的框構架被放置在振動台上進行震動試驗並從量測到的反驗來重建地震力。這個試驗驗證了外力識別技術於地震力識別的可行性。另外，廣州鐵塔(Canton tower)受Burma地震力作用的紀錄被運用來嘗試重建地震力。最後，本文中透過沖刷實驗並配合外力識別技術來了解擾動力並建立出波速、擾動力及沖刷深度三者的關係。For purposes of monitoring and damage prognosis it is important to know the external loads which act on a structural system. Therefore, the force identification technique is developed for understanding the external loads. The force identification algorithm is the core among the force identification technique. In this paper, several force identification algorithms including modified recursive square Kalman filter method (MRLSKF), unknown input Kalman filter method (UIKF), reduced order observer method (ROOM) and the theoretical Bernoulli beam method (TBB) are introduced. Four simulated cases are used for verifying force identification algorithms. Also, the force identification technique is applied to experimental cases and real world cases. First a simply supported beam is set for proving the practicability of force identification technique. Second, a 6 floor frame structure by shaking table test is introduced for verifying that the force identification can be used in earthquake excitation case. Then, a real world case about the Canton tower under the Burma earthquake is introduced. In this case, the input force is tried to reconstruct by measurement response. Finally, a scouring experiment is applied to understand the turbulence force and establish the relationship among the wave velocity, turbulence force and the depth of scoring.口試委員審定書 i Acknowledgement ii Abstract (in Chinese) iii Acknowledgement (in English) iv Contents v Table list viii Figure list ix Chapter 1 Introduction 1 1.1 Background 1 1.2 Research Objectives 3 Chapter 2 Force Identification Algorithm 5 2.1 Introduction 5 2.2 Modified Recursive Least Square Kalman Filter Method (MRLSKF) 5 2.2.1 State Space System 5 2.2.2 Theory of Kalman Filter 7 2.2.3 Recursive Least Square Method 10 2.2.4 Process Noise Covariance Estimator 13 2.2.5 Procedure 17 2.3 Unknown Input Kalman Filter Method (UIKF) 19 2.3.1 State Space System 19 2.3.2 Process noise covariance estimator 20 2.3.3 Procedure 21 2.4 Reduced Order Observer Method (ROOM) 22 2.4.1 State Space System 22 2.4.2 Theory 23 2.4.3 Procedure 29 2.5 Theoretical Bernoulli Beam Method (TBB) 30 2.5.1 Theory 30 Chapter 3 Simulation 33 3.1 Introduction 33 3.2 Case 1: Simulated Simply Supported Beam 33 3.2.1 Impact Load Case 34 3.2.2 Random Force 37 3.2.3 Multiple load case 38 3.3 Case 2: Earthquake Force Applying on Multi-DOFs Structure 38 3.4 Case 3: Under-water Structure with Water-turbulence Force 40 3.4.1 Theory for Generating Water Turbulence Force 41 3.4.2 Simulation case 48 3.5 Importance of Process Noise Covariance 49 Chapter 4 Case Study 52 4.1 Introduction 52 4.2 Case 1: Simply Supported Beam 53 4.3 Case 2: Verification of Input Force Identification Using Shaking Table Test for 6 Floor Frame Structure 56 4.4 Case 3: Applying Force Identification Technique to Canton Tower Under Burma Earthquake 58 4.5 Case 4: Identification of Water-turbulence Force by Sub-beam 60 Chapter 5 Conclusions 67 5.1 Research Conclusion 67 5.2 Recommendations for Future Work 69 References 71 Appendix A: Bernoulli beam equation 75 Appendix B: Stochastic Subspace identification (SSI) 77 Appendix C: Singular Spectrum Analysis (SSA) 797052100 bytesapplication/pdf論文公開時間：2014/08/16論文使用權限：同意有償授權(權利金給回饋學校)外力識別演算法修正回歸平方卡爾曼濾波器法未知外力卡爾曼濾波器法降階觀測器法理論柏努力梁方法莫瑞森方程式廣州鐵塔沖刷thesishttp://ntur.lib.ntu.edu.tw/bitstream/246246/260756/1/ntu-103-R01521213-1.pdf