Sensorless Control Of PMSMs Based On Nonlinear Observer

High-performance permanent magnet synchronous motor (PMSM) drive systems have been more and more widely used in aerospace, robotics. CNC machine tools, electric vehicles and other areas. High resolution speed/position sensor is one of the key parts of PMSM drive systems, but the high price and strict environmental requirements of the sensor limit the application of PMSM drive systems. Thus sensorless control of PMSMs has great practical value.The key problem of sensorless control is how to accurately estimate the position and speed of rotor. This thesis summarized several common algorithms of sensorless control, and conducted deep researches on sensorless control based on nonlinear observer. The nonlinear observer was designed to add a correction term of estimation errors to the dynamic equations of the studying object. The main task of the nonlinear observer is to design the correction term.There're mainly two types of PMSMs:surface mounted PMSMs and interior PMSMs, and this thesis presented the design method and application method of nonlinear observer for both of the two types of PMSMs separately. For surface mounted PMSMs described inαβstationary, its observability was discussed and the nonlinear observer was presented, the stable criteria of the observer were stated based on Lyapunov stable theory. The load torque was ignored in observer design, which would cause an error in modeling, thus the load disturbance observer was added to form an extended nonlinear observer. It can be seen through simulation that the extended nonlinear observer is more accurate and insensitive to load disturbance. Because the form of the dynamic equations of interior PMSMs is complex, an simplification to the dynamic equations was introduced, which would make the dynamic equations as simple as it of surface mounted PMSM. The nonlinear observer and load disturbance observer was presented following the same procedure as it for the surface mounted PMSMs.Because the nonlinear observer is designed based on PMSM's model, it's accuracy will be influenced by the change of PMSM's parameters. So an online parameter adjustment was introduced. And the modeling and the sampling was not perfect, there existed errors between the real states and the estimated states, so compensation to the estimated states was introduced. Experiments was carried out on PMSM control platform based on TMS320F2812 microcontroller to test the system's performances under stable, dynamic, load disturbance situation. Results showed that the nonlinear observer could be applied to interior PMSMs as well as it for surface mounted PMSMs, the estimated rotor position and speed could converge to the real value quickly and accurately.