Vibration suppression integrating semi-active control and model predictive control based on harmonic input

抄録

In this paper, we propose a novel control method that incorporates model predictive control (MPC) and semi-active control. Semi-active vibration control is excellent in damping effect, energy consumption, and system stability. With the degree extension of the MDOF system, parameter tuning might be complicated in the conventional semi-active methods. Therefore, in order to extend the diversity of the input decisions, MPC and semi-active control are integrated. MPC can design control input trajectories arbitrary. A predictive controller has an internal model and predict future states of a system among a prediction horizon. The future states are optimized by selecting the most desirable input trajectory. Future states of a system are predicted based on two assumptions. The first assumption considers future control inputs are expressed as summations of harmonic functions. The second considers that periodic disturbances that cause resonance are applied to a structure. Switching is conducted to match the polarity of the charge of the piezoelectric transducer with target inputs derived based on the proposed method. Switching based on the prediction can achieve more effective vibration suppression and easier tuning than conventional semi-active methods. Computational loads are reduced by appropriately designing input trajectory. The beneficial features of the proposed method are demonstrated in simulations and experiments.