Article ID: 24-00363
This paper proposes a novel quasi-preview control method for structural vibration control using real-time seismic wave data from multiple remote observation sites and structural response. The method uses an adaptive filter based on a polynomial dynamic model to estimate the future seismic waveform at the controlled structural system. The coefficients of the adaptive filter are updated with the normalized least mean square (NLMS) method in every sample period. The estimated waveform is used for feedforward control, while the structural response is used for feedback control. The control strategy is referred to as the quasi-preview control. We also incorporate a switching mechanism into the quasi-preview control to select the optimal control input from multiple candidates of control inputs obtained from the multiple remote observation sites based on a performance index. The performance index involves the one-step-ahead prediction of the structural response obtained by the model of the structural system and the control input candidate to perform the switching action. All design parameters in the quasi-preview control are optimized with the particle swarm optimization (PSO) algorithm. We demonstrate the proposed quasi-preview control method through simulations with recorded data of a major seismic event. The proposed method achieves superior control performance compared to a conventional LQR feedback control method. The proposed method also enhances the robustness of the control system with the switching mechanism by using multiple observation sites.
