A disturbance-aware model predictive control method for suppressing the transverse vibration of ultra-high-speed elevator

Abstract

This paper proposes a disturbance-aware model predictive control (DAMPC) method to suppress transverse vibrations in ultra-high-speed elevator cars, addressing nonlinear dynamics that exceed the capabilities of conventional linear controllers. Key external excitations, including rail unevenness (modeled by filtered Gaussian white noise) and airflow-induced aerodynamic forces (expressed quadratically with velocity), are analyzed and mathematically characterized. A gas-solid coupled 4-DOF transverse vibration model is developed. By constructing an over-determined system of equations and employing the singular value decomposition method to perceive disturbances, the DAMPC method incorporates these disturbances into the model predictive controller to compute the optimal control input by minimizing a predefined cost function. Simulation experiments demonstrate that DAMPC is superior, achieving average vibration reduction improvements of 30.6%, 61.7%, and 79.4% over MPC, LQR, and uncontrolled cases, respectively. The method offers a scalable framework for enhancing ride comfort, safety, and operational durability in ultra-high-speed elevators, with broader applicability to active vibration control in high-speed transport systems.