| Self-sensing active magnetic bearing has a series of unique performance advantages including:integrated hardware, low cost, simple designed, actuator acts as sensor and etc, so it is the most innovative field in the scope of magnetic bearing.This dissertation intended to research the self-sensing technology based on the PWM signal on the active magnetic bearing platform which employs fixed frequency current mode power amplifier. The self-sensing technology employing PWM signal as high frequency injection had been thorough and detailed studied with respect to theoretical analysis, simulated research and experimental verifications. Some independent-innovative achievements on the basic theory and key technology are obtained:1. The inductor model of the active magnetic bearing was created, and then, the functional relationship between coil equivalent inductance and air gap was obtained under the assumptions that magnetic saturation, flux leakage, hysteresis and eddy current are all ignored. The air gap can be estimated by the variation of the coil inductance, and this conclusion is an important foundation for the self-sensing active magnetic bearing.2. Topology and operation principle of the fixed frequency current mode power amplifier were studied. Harmonic analysis on coil voltage and current was carried out using Fourier series. The functional relationship between current ripple amplitude, air gap and PWM duty cycle was obtained.3. Jacobi-Angle identity, the first kind of Bessel function and other mathematical tools were employed to research the influence on the frequency characteristic of current ripple from the force disturbance. Theoretical analysis and experimental results showed that time-varying duty cycle is generated when power amplifier tracks the command signal which caused by force feed through. The amplitude of the current ripple and its frequency component are both influenced by the time-varying duty cycle. Infinite side frequency spectrum with the same interval appears around the spectrum component at switching frequency, and the frequency interval equals to the frequency of current reference signal. The amplitude of each side frequency component is given by the first kind of Bessel function.4. Considering of the special frequency spectrum characteristics of the current ripple, the effect of pass band gain and pass band width, the main parameters of the band pass filter in the demodulator, on the demodulator’s performance was discussed in detail based on the Parseval theorem. The proper range of each parameter was given. An improved demodulator was designed using UAF42, simulation and experiment result verified its performance.5. A self-sensing strategy on magnetic bearing by duty cycle compensation was proposed. The hardware complexity was reduced to half by utilizing the on-chip resource of DSP28335. The position estimation from the duty cycle compensation strategy was used as feedback signal, and the four degree of freedom radial active magnetic bearing was steadily suspended within0-3000rpm.6. A self-sensing strategy on magnetic bearing using nonlinear parameter estimation was proposed. The closed-loop estimation of the rotor position was achieved by employing a simple PI controller and a magnetic bearing model which has the same parameters to the real one. Simulation results proved that, comparing to other open loop position estimation strategy, this method can greatly improve the accuracy and bandwidth of the estimator. The position estimation from the nonlinear estimator was used as feedback signal, and the four degree of freedom radial active magnetic bearing system achieved satisfactory performance within0-3000rpm, which obviously superior to that of duty cycle compensation strategy. |
PDF Full Text Request