Dynamic decoupling control of AC-DC hybrid magnetic bearing based on neural network inverse method

A dynamic decoupling control approach based on neural network inverse system theory is developed for the AC-DC 3 degrees of freedom hybrid magnetic bearing (AC-DC 3-DOF HMB), which is a multivariable, nonlinear, strong coupled system. The configuration of AC-DC 3-DOF HMB is introduced briefly. The mathematics equations of radial and axial suspension forces are deduced. The analytical inverse system of the HMB is obtained by analyzing the reversibility of the mathematics model. The static neural networks and integrators are used to construct neural network inverse system. Then neural network inverse system and original system are in series to constitute pseudo linear system, and linear system theory is applied to the pseudo linear system to synthesize and simulate. The simulation results have shown that this kind of control strategy can realize dynamic decoupling control among 3 degrees of freedom of the system, and the whole control system has good dynamic and static performance.