Research On Algorithm Of Fuzzy Control And Digital Filter For AMB

Active magnetic bearing (AMB), which makes rotor suspend stably by controllable magnetic field force, is novel and of high performance. It will take the place of traditional roll bearing and sliding bearing, becoming the most powerful leading bearing of the 21th century. Magnetic bearings can only work well with the help of control system. So the performance of controller affects the dynamic performance of magnetic bearings and the control precision of the rotor directly. This article emphasizes on the control and filter arithmetic.AMB is an intrinsic instable system, so, it is indispensable to introduce differential part. In real implementation, usually, PID control is adopted. In this article, several improved PID control method and Simulink solution for the integral saturation phenomenon which is usually ignored, have been introduced. As for the deficiency of traditional tuning approach of PID, three PID parameters have been optimized by using the simplex method of linear programming and utilizing the improved ITAE performance index.Traditional PID control is designed according to the linear system theory, but AMB has the nonlinear characteristic, as a result, a rotor may not levitate efficiently far away from the balance point. Combining the high precision of PID and robustness of fuzzy control and rapid response performance of proportional, the P-Fuzzy-PID control has been proposed. Simulation result shows that this control can make the rotor levitate quickly without steady state error.In order to overcome the defect of deciding fuzzy rules only by experience, this article analyzes and designs fuzzy controllers with 4 adjustable coefficients and a fitting adjustable coefficient. The ITAE performance index that has been improved is used to evaluate the performance of AMB control systems. As for the fuzzy controller with a fitting adjustable coefficient, the fitting adjustable coefficient is approximated by a three order polynomial with least-squares solution. Simulation result shows that, with fuzzy controller with a fitting adjustable coefficient, excellent response performance can be gained.Filter is used to process signal in frequency domain and able to improve the performance of AMB control system. In this article, the low pass IIR digital filter is realized by digitizing the analog filter by bilinear transform method. Utilizing the command functions of the MATLAB signal processing toolbox, let the sampled actual displace signal passes the designed low pass filter, and the simulation results of both time and frequency domains demonstrate that the designed IIR filter has good performance.