Stabilization of Active Magnetic Bearing System Using Single Neuron Pid Controller

This paper presents the development and implementation of a hybrid controller to automatically levitate and precisely regulate the position of a suspended rotor in Active Magnetic Bearing system. The inherent nonlinearity and open-loop instability of AMBs present challenges for design a controller and the problem can be formulated as a dynamic, multivariable optimization problem. A single neuron adaptive PID controller which is different from conventional PID controller with constant gains is designed by combining single neuron structure with PID mechanism. The ability of selflearning and adaptive of ANN and the robustness of PID control retained in the proposed SNPID controller. It can adjust weighting parameters to retain the desired performance automatically. The simulation results show that using SNPID controller designed via PSO algorithm exhibits superior robustness compared to PID controlled systems.