Particle Swarm Optimization Algorithm-based Identification of Fractional-order Magnetic Levitation Systems

Magnetic levitation systems are strongly nonlinear systems with time-varying model parameters, many external disturbances, and open-loop instability. Most of the linearized modeling methods ignore the higher-order nonlinearity of the system to some extent and the physical characteristics of the system such as the leakage flux of the winding, the reluctance of the magnet core and the guideway, making the established model less accurate. In this paper, a particle swarm algorithm-based fractional order suspension system identification method is proposed to identify the unknown system into a more accurate fractional order model. The obtained fractional-order model is more accurate in depicting the actual system than the integer-order model, which is more conducive to the analysis of the system and the design of the controller.