An Algorithm for Parameter Identifications of Active Mass Damper and Primary Systems Based on Acceleration Measurements

Active control systems have already been installed in real structures and are able to decrease the wind- and earthquake-induced responses, while the active mass damper (AMD) is one of the most popular types of such systems. In practice, an AMD is generally assembled in-situ along with the construction of a building. In such a case, the AMD and the building is coupled as an entire system. After the construction is completed, the dynamic properties of the AMD subsystem and the primary building itself are unknown and cannot be identified individually to verify their design demands. For this purpose, a methodology is developed to obtain the feedback gain of the AMD controller and the dynamic properties of the primary building based on the complex eigen-parameters of the coupled building-AMD system. By means of the theoretical derivation in state-space, the non-classical damping feature of the system is characterized. This methodology can be combined with any state-space based system identification technique as a procedure to achieve the goal on the basis of the acceleration measurements of the building-AMD system. Results from numerical verifications show that the procedure is capable of extracting parameters and is applicable for AMD implementation practices. In addition, control force execution time delay cannot be avoided in real application of active control. Small delay time can degrade the control performance and may cause system instability. In this study, time delay effect of AMD system is considered in the proposed methodology to obtain the feedback gain of the AMD controller and the dynamic properties of the primary building.