Research On The Critical Technology Of PMSM Sensorless Control System

Permanent Magnet Synchronous Motors (PMSM) have been applied widely in electrical drive systems, for merits such as compactness, high power density, high efficiency and high power factor. Recently, sensorless control of PMSM drive system has become focus in this field. Based on reading and analyzing relevant information at home and abroad, current situation of sensorless control was reviewed; frontiers and hot issues in this field were explored. Then, this dissertation points out the difficulty of PMSM sensorless control: control in low-speed area where the key is to estimate speed accurately; the developing direction of full-speed range control is compound control. How to realize control method smooth transition between low-speed range and high-speed range is the key whether compound control can realize. Due to the above problems, this dissertation deeply investigates key issues of PMSM sensorless control by adopting high frequency signal injection and MT coordinate magnetic flux closed-loop estimation. The main research and innovative achievements are as follows:Based on analyzing mathematic model of salient-pole PMSM, this dissertation explained the principle of estimating rotor magnetic pole position through high frequency signal injection and analyzed reasons affecting the accuracy of signal estimation. Based on simplified Extended Kalman Filter (EKF), a novel signal extraction method was presented, which can not only extract rotor position information but also filter fundamental frequency control current. Compared with traditional IIR and FIR filter, the novel EKF, adopting iterative algorithm, has faster filter speed and better filter performance. It can eliminate low-pass filter in control feedback loop. So the dynamic performance and stability of the system is enhanced and the design of system controller is simplified. At the same time, since there is no high frequency signal component in control circuit, the controller affection to high frequency position estimation is decreased. The design of position estimator is unaffected by control regulator and simplified. Simulation and experiment results show the validity of this method.Because of the interference of saturation, non-linear of motor, dead-time of inverter, and bandwidth of controller, the speed estimated by angle differential is not accurate, even has large lag, when adopting high frequency signal injection method. It is harmful for real time speed control system. Even the whole control system diverges. This dissertation simulated variety of existing solutions, developed a novel MT coordinate magnetic flux closed-loop observer based on non-salient model. Meanwhile, rotor position information received by high frequency signal injection method was used to correct the result of the above method. The final obtained torque flux was used to estimate speed. The merits of this method are fast system dynamic response and considerably weakened interference component in rotor position signal by integral action of voltage model.The embedded PMSM development platform, based on both dSPACE 1103 and FPGA was designed, among which the dSPACE 1103 is the main controller. The control model can be built under Matlab/Simulink. After simulating successfully, through simple signal module replacement, it can be downloaded in the dSPACE controller to realize control ability in the system. Xilinx XUP VII PRO development board which is high frequency signal controller realizes extraction of rotor position signal. With System Generator of Xilinx, simulation model in Matlab/Simulink can be translated into HDL language recognized by FPGA to realize hardware-in-loop simulation. Hence, Hardware development cycle is largely simplified. Adopting presented high frequency signal injection and MT coordinat magnetic flux closed-loop method, PMSM simulation was built, experiment results were presented. Experiment results show that presented compound control method is able to reliably estimate rotor magnetic pole position and speed under low-speed and standstill range, and is not sensitive of PMSM parameters.