Research On Sensorless Control Strategy Of PMSM Drives Based On Carrier-wave Injection

Permanent magnet synchronous motor(PMSM) has been widely used in industrial field due to its advantages of high efficiency, high power density and small volume, etc. Since the adoption of encoders in close-loop system leads to a decrease in system reliability and a rise in cost, sensorless control aimed at eliminating encoders has become a research focus. High frequency(HF) signal injection is an effective way to obtain position information in low speed range. However, observer and filters employed in the speed feedback loop limited the bandwidth of the whole system. And the method itself brings in noise issues as well. One way to alleviate this problem is to decrease the amplitudes of injected signals. Nevertheless, the signal noise ratio(SNR) of the induced signals will also degrade, making them harder to be extracted. This thesis focuses on the dynamic performance and low SNR signal extraction methodologies based on HF injection sensorless interior permanent magnet synchronous motor(IPMSM) drive system.To begin with, the restraining factors of the whole sensorless control system are investigated. Combining with the practical application of IPMSM drive system, the mathematical model of the double closed-loop system is established. The effects of the control delay and the adoption of filters in current feedback loop on the current loop bandwidth are analyzed. By taking the speed loop characteristics into consideration, how filters in the speed feedback loop affect the bandwidth of speed loop is studied. Finally, the theoretical analysis is verified by simulation and experiments.On the basis of above, research on improving the bandwidth of IPMSM sensorless control system is carried out. First of all, by establishing high frequency model of IPMSM, the principles for sensorless control which injects HF signal as an excitation are elaborated. Instead of using the traditional digital filters, an advanced mechanism for separating the fundamental components and the high frequency ones is employed. And at last, with a proper adjustment in the acquirement of speed estimate, higher bandwidth and better dynamic performance can be achieved.Then an extraction approach using discrete wavelet transformation which meets the requirement of low SNR circumstance is studied. It allows a smaller high frequency excitation signal which benefits in reducing additional noise. Depending on the advantages of wavelet transformation in processing low SNR and non-stable signals, low SNR rotor position information can be extracted, which means wavelet transformation is superior to traditional Butterworth filters in obtaining rotor position information under smaller amplitude voltage injection.Finally, based on the theoretical analysis, the simulation with Matlab/Simulink was given to verify the proposed methods. And the effectiveness of the proposed sensorless strategies is tested at an IPMSM drive experimental platform with an ARM chip.