Input-output Behavioral Data-driven LQG Control of Axial-gap Self-bearing Motor

This paper addresses the displacement control of the tilt-controlling axial-gap self-bearing motor via input-output data-driven Linear-Quadratic-Gaussian (LQG) approach. First, the dynamic model is stated and the linearized state-space model at the operating point is obtained. Although LQG method has been widely used in many practical problems, it is still hard to effectively address nonlinear processes, high-order dynamic modeling, and the negative effects caused by the complex structure of motors and noise. Therefore, we fully utilize actual input-output data from the system to reconstruct the system’s dynamic behavior and compute the static feedback control rates that can compete with the dynamic feedback gains obtained from model-free data-driven methods, which effectively improve the displacement control performance of the axial-gap magnetic self-bearing motor.