Predictive Semi-active Vibration Control with the Tree-based Formulation and Optimization Algorithm

Abstract

This paper introduces a novel strategy that combines model predictive control (MPC) and semi-active vibration control with a piezoelectric transducer. Semi-active vibration control is an energy-efficient method that realizes effective suppression performance and stability. Formulating an effective control strategy for the semi-active vibration control further improves its suppression performance, and incorporating MPC is a promising solution. Since the semi-active input changes discontinuously as the switching of the control circuit, it is challenging to apply the commonly-used MPC strategy for prediction and optimization. The proposed method constructs a future prediction and optimization algorithm based on the semi-active input by using a tree data structure. The algorithm achieves flexible prediction and optimization of a semi-active input trajectory directly with a simple tree traversal. Furthermore, the switching criterion introduced in our method reduces the computational cost by only performing future predictions at which a promising performance is expected. The proposed predictive switching vibration control with tree-based formulation and optimization or PSTFO method achieves a direct and flexible optimization of the semi-active input trajectory while reducing the computational burden. The control performance of the conventional and proposed method was compared with the experiment, validating the effectiveness of the proposed method.