During ultra-high-speed elevator operation, the car experiences severe horizontal vibrations due to multi-source external excitations and system nonlinearity, posing risks to system safety and ride comfort. This paper proposes a state estimation-based model predictive control (MPC) method for multi-objective optimization of horizontal vibration control in ultra-high-speed elevators. First, to address system nonlinearity and the coupling of multi-source external excitations, a nonlinear full-chain-coupled dynamic model for ultra-high-speed elevators subjected to multi-source excitation has been established, and three typical external excitations are simulated. Then, a multi-objective optimization problem is formulated to suppress car vibrations and mitigate abrupt actuator forces, and an MPC method based on an unscented Kalman filter (UKF) hybrid state estimator is proposed. Subsequently, the impact of actuator quantity and spatial arrangement on control performance and computational complexity is investigated, revealing that a symmetric dual-actuator configuration exhibits unique advantages. Finally, simulations and real-elevator experiments demonstrate that, compared to the linear quadratic regulator (LQR) algorithm, the proposed method reduces the A95 peak value of horizontal acceleration by 59. 58%, the mean deflection angular acceleration by 76. 87%, and actuator forces were strictly constrained within \ (300\, N\) without any large abrupt change, verifying the effectiveness of the proposed method.
Cai et al. (Sun,) studied this question.