Vol.39/No.3 (153) (2024)

Vol.39/No.3 (153) (2024)

TitleDevelopment of Electromagnetic Variable Damping Seismic Isolation System
AuthorGing-Long Lin, Yi-Chun Huang, Ming-Bin Chang
Keywordselectromagnetic damping, variable damping, semi-active control, seismic isolation,
near-fault earthquake, shaking table test
AbstractThis study aims to develop an electromagnetic variable-damping seismic isolation system(EM-VDSIS), which incorporates a variable electromagnetic damper into a sliding isolation system. This allows the damping ratio of the EM-VDSIS to be a controllable parameter that can change in real-time according to a defined control law, enhancing the effectiveness of seismic isolation. The principle of the EM-VDSIS is first introduced in this study, and a mathematical model is constructed based on the mechanical behavior of the EM-VDSIS to derive its equations of motion and numerical simulation methods. Subsequently, numerical simulations are used to verify the variable damping function of the EM-VDSIS, with the potential to meet the seismic isolation requirements of both near-field and far-field ground motions. In terms of implementation, the EM-VDSIS (I) was first designed and manufactured, with a controllable resistance provided by an electronic load machine to alter the damping ratio of the isolation system. However, the electronic load machine fails to achieve ideal control of electromagnetic damping. Therefore, the EM-VDSIS (II) upgraded the power of the electromagnetic damper and used a mechanical variable resistor mechanism to provide controllable resistance values. The results of open-loop control with EM-VDSIS (II) using a shaking table demonstrate that the mechanical variable resistor mechanism can change the damping ratio of the isolation system in real-time according to control commands. Additionally, the theoretical analysis of the isolation system matches the experimental dynamic responses, confirming the accuracy of the theoretical model and related formulas in this study.
TitleSemi-Active Sloped Rolling-Type Isolators Based on Earthquake Early Warning Techniques
AuthorZi-Ting Chuang, Shieh-Kung Huang, Ting-Yu Hsu, Shiang-Jung Wang
Keywordsslope rolling-type seismic isolators, semi-active control, earthquake early warning, magnetorheological damper 
AbstractPassive rolling seismic isolators, once designed, manufactured, and installed on-site, have fixed parameters. They can effectively control the maximum acceleration response of the protected object within a certain range when subjected to typical far-field seismic waves, showing remarkable performance. However, when subjected to near-fault seismic waves with velocity pulses, their displacement response may exceed the limits, leading to collisions and damage to the protected object. Therefore, this study proposes the development of a semi-active rolling seismic isolator that integrates sloped rolling-type isolators (SRI) with magnetorheological dampers (MRDampers) and utilizes earthquake early warning technology. This research involves the development of a convolutional neural network (CNN) prediction model to estimate the peak ground velocity (PGV) based on the characteristics of the initial arriving wave. Additionally, control laws are established to determine the required voltage forthe MR Damper based on the predicted PGV. By measuring the initial wave arrival information, the system can predict the PGV using the CNN model and apply the control laws to obtain the required voltage for the MR Damper. This enables the adjusting the damping force of the rolling seismic isolator to prevent its displacement response from exceeding the limits during near-fault strong motions with velocity pulses. The feasibility of this proposed approach is verified through experimental tests of a SRI system. The experimental results demonstrate that this system effectively limits the displacement of SRI below the threshold value, validating the concept and feasibility of the proposed method.
Title Development and Performance Analysis of Seismic Isolation Bearings With Angled Viscous Damper
Author Chieh-Yu Liu, Chia-Ming Chang
Keywords seismic isolation, geometric nonlinearity, multiple performance objectives
Abstract Earthquakes pose a significant impact on machinery requiring high-precision manufacturing in advanced facilities. Meanwhile, severe earthquakes cause enormous economic losses and threaten lives. Base isolation is a popular method for controlling seismic impact, extensively employed to mitigate structural response and lessen seismic risk. However, certain studies suggest that base isolation could lead to excessive displacement during severe earthquakes. To enhance safety and functionality, supplemental damping is recommended to be integrated into the isolation system to mitigate large displacements. However, isolation parameters are typically designed for design-level earthquakes, resulting in increased absolute acceleration during small-to-moderate earthquakes and limited displacement control capacity during large earthquakes due to the lack of adaptability in damping. This study proposes an isolation system with geometrically nonlinear damping and first examines the dynamic characteristics. Subsequently, the relationship between seismic input and isolation responses in the frequency domain is analyzed using the averaging method. The seismic performance of the proposed system is then evaluated using earthquake records from the 2016 Kaohsiung Meinong earthquake to assess time-domain performance set at various initial inclining angles. Through a series of investigations, it is observed that the geometrically nonlinear damping configuration offers advantages by providing adaptive damping forces to isolation bearings and achieving multiple performance objectives across different earthquake magnitudes. Additionally, the proposed isolation system with a geometrically nonlinear viscous damper is experimentally validated to confirm the displacement-force relationship through shake table testing. In the experimental setup, the isolation system comprises three single-curvature grooves moving on fixed ball bearings alongside an angled linearly viscous damper. During the test, this isolation system is subjected to harmonic excitation on a uniaxial shake table to obtain force-displacement behaviors. The results demonstrate a close behavior between the simulated force displacement relationship and the experiment, thus indirectly carry out the multiple performance objectives of the proposed system against earthquakes.
Title Automatic Generation of an Active Structural Controller Using Direct Excitation With Machine Learning
Author Che-Wei Chou, Wei-Jung Wang, Pei-Ching Chen
Keywords direct excitation method, machine learning, autoregressive with exogenous inputs, optimal control, active mass damper, shake table testing
Abstract For active control structural systems, it is necessary to obtain a simplified numerical model of the structure through system identification. Controller design and analysis for vibration control are conducted based on this numerical model. Therefore, the representativeness and accuracy of the numerical model directly affect the performance of active structural control. Additionally, common structural controllers such as the linear-quadratic regulator (LQR) require the additional design of an observer to estimate the state of the structure for feedback control. However, both system identification and the design of structural controller and observer rely on the experience of engineers, thus increasing the practical application barrier of active structural control. In view of this, this study proposes a method for automatically generating structural controllers to mitigate seismic responses of structures. By using active control devices to generate small excitations on the structure and measuring the associated acceleration response, the inverse relationship between excitation force and structural acceleration response can be obtained through machine learning with a recurrent dynamic neural network called the autoregressive with exogenous inputs (ARX) model. Two structural models with 9-story, and 27-story configurations were assumed for numerical simulation. An active mass damper (AMD) was installed at the top of each structural model. Time history analyses were performed using 14 earthquake acceleration records to compare the control performance of the controller generated by the proposed method and LQR with optimized weighting matrices. Finally, a three-story shear building specimen was fabricated in the structural laboratory for shake table verification testing. An AMD driven by a servo motor was installed at the top floor. The experimental results show that the automatically-generated structural controller can effectively reduce the displacement and acceleration responses of the specimen and has similar structural control performance to structural controllers obtained through conventional design approaches.
TitlePerformance-Based Design and Assessment of Friction Dampers for Seismic Retrofit of a Reinforced-Concrete Structure
AuthorShih-Wei Yeh, Lyan-Ywan Lu, Fu-Pei Hsiao, Sheng-Qin Quo, Pin-Tsun Chen, Chia-Shang Chang Chien
Keywordsseismic retrofit, friction damper, performance-based design, seismic assessment, Bouc-Wen-Baber-Noori model, reinforced-concrete structure
AbstractThe use of dampers for seismic retrofit of a building structure is an advanced retrofitting technique, and many researchers have proposed various damper design methods. However, viscous and metallic-type dampers are more commonly used in practice, and studies on the design and assessment of friction dampers for seismic retrofit of structures are limited. To this end, this paper proposes a performance-based design method for friction dampers. The method, which combines the capacity-spectrum method with the codified damper design formulas, can improve the retrofitted building to a desired performance level. To validate the proposed performance-based design method, this paper employs the design procedure to determine the design parameters of a friction damper used in seismic retrofit of a seven-story reinforced concrete (RC) building. The seismic performance of the RC building with the friction damper is then assessed through the nonlinear time-history analysis using 11 spectrum-compatible ground motions. This paper adopts the Bouc-Wen-Baber-Noori model in the numerical model to accurately simulate the post-yield behavior of the RC columns in the first story. Under the DBE (design basis earthquake) intensity ground motions, the numerical results indicate an improvement in the seismic performance level of the RC building from the CP (collapse prevention) level to the LS (life safety) level, which meets the pre-set performance design objective. This validates the effectiveness of the proposed friction damper design method. Additionally, the numerical
simulation also demonstrates that the friction damper achieves a reduction rate of 40% on the peak inter story drift ratio of the first story and a reduction rate of 6% on the peak acceleration of the top floor. 
TitleDesign Passive Tuned Mass Damper With Optimal Target Response Using Static Output Feedback and Parameter Updating Iterative Method
AuthorYong-An Lai, Chi-Hung Chang, Xian-Zheng Hong
Keywordstuned mass damper (TMD), optimal passive control, optimal stiffness and damping coefficient, optimal mass, static output feedback, wind and seismic loads
AbstractThis study proposes a comprehensive passive tuned mass damper (TMD) optimization design method to minimize structural mean square responses. The optimization design problem for passive TMD is reformulated as an optimal control problem, specifically, the optimal gain matrix design problem in static output feedback (or direct output feedback). By solving for the optimal gain matrix, the optimal stiffness and damping coefficients, or optimal mass, of the passive tuned mass damper can be obtained. The proposed method is applicable to both single degree-of-freedom (SDOF) and multi-degree-of-freedom (MDOF) structures, whether damped or undamped structures, and subjected to wind or seismic loads. Moreover, for different vibration reduction objectives, only different output matrices need to be selected, and the corresponding weighting matrices can be combined for the solving process, making it intuitive and straightforward. In the case of SDOF structures, numerical simulations validate that the optimal design parameters of the passive TMD obtained through this method are identical to the analytical solutions derived from random vibration theory, or closely approach to the approximate solutions, confirming the correctness and feasibility of the proposed design method. Additionally, using the proposed method, the optimal TMD frequency ratio and TMD damping ratio for minimizing the mean square response of velocity or absolute acceleration of SDOF structures under seismic forces are presented, providing reference for engineers in design. Finally, demonstrations are conducted with a passive TMD installed on a five-story MDOF structures and a ten-story ETABS structure, respectively, to design TMD optimal stiffness and damping coefficients, or optimal TMD mass. The results confirm that the proposed method is applicable to MDOF structural systems.