Robust Adaptive Cerebellar Model Articulation Controller for 1-DOF Nonlaminated Active Magnetic Bearings

This paper presents a robust adaptive cerebellar model articulation controller (RACMAC) for 1-DOF nonlaminated active magnetic bearings (AMBs) to achieve desired positions for the rotor using a robust sliding mode control based. The dynamic model of 1-DOF nonlaminated AMB is introduced in fractional order equations. However, it is challenging to design a controller based on the model's parameters due to undefined components and external disturbances such as eddy current losses in the actuator, external disturbance, variant parameters of the model while operating. In order to tackle the problem, RACMAC, which has a cerebellar model to estimate nonlinear disturbances, is investigated to resolve this problem. Based on this estimation, a robust adaptive controller that approximates the ideal and compensation controllers is calculated. The online parameters of the neural network are adjusted using Lyapunov's stability theory to ensure the stability of system. Simulation results are presented to demonstrate the effectiveness of the proposed controller.The simulation results indicate that the CMAC multiple nonlinear multiple estimators are close to the actual nonlinear disturbance value, and the effectiveness of the proposed RACMAC method compared with the FOPID and SMC controllers has been studied previously.