Nonlinear and adaptive backstepping speed control of permanent magnet synchronous motor

Permanent Magnet Synchronous Motors (PMSMs) have high efficiency, long life and high power density over the other kinds of electric motors such as direct current motors and induction motors. By virtue of these superiorities, PMSMs are widely used in, primarily position and speed control, various fields such as home applications, robotics and electric vehicles. However, as all electromechanical systems, PMSMs are also subject to parameter variations and immeasurable external disturbances in the course of operation. During operation, electrical parameters vary with temperature and magnetic saturation effects, as for mechanical parameters, they vary with applied load torque. From this perspective, it is clearly understood that PMSMs have the nonlinear dynamics containing parameter and external disturbance uncertainties.Since fixed gain linear controllers are sensitive to parameter variations and external disturbances, they may fail to meet the high performance requirements for position and/or speed tracking control of PMSMs. In this thesis study, two different and novel nonlinear and adaptive backstepping controller designs are proposed for high performance speed tracking control of an uncertain PMSM. The first controller design takes four parameter uncertainties into account and five parameter uncertainties are considered in the second controller design. Load torque is supposed to be uncertain in both controller designs. The second controller does not possess some drawbacks that the first controller has, for this reason; real time implementation of the second controller is only carried out. In this controller design, except for the pole pairs, all the other parameters together with load torque are assumed uncertain and all the nonlinear dynamics are taken into consideration. For both controller designs, in the closed loop control system, three phase currents feeding the PMSM and rotor speed are available for feedback. Control inputs and parameter estimation laws are designed by choosing appropriate Lyapunov functions. Proposed controllers do not have the drawbacks of singularity and overparameterization. In addition, regression matrices are not used in the design of the controllers and thereby the design complexity decreases. According to the result of the stability analysis, the closed loop control systems are globally asymptotically stable and all the signals in the closed loop control system stay bounded. The design of PI controller is carried out based on decoupling and pole-zero cancellation, and afterwards a performance analysis between these two controllers, the proposed and PI, is also concluded. According to the results of the performance analysis based on simulation studies, it is understood that the proposed controller requires less model knowledge, ensures more robustness against uncertainties and hence has a higher performance in general.The results of the experimental studies demonstrate, whatever the initial conditions, that the proposed controller ensures high performance asymptotic tracking of a time varying reference speed profile, robustness against all the uncertainties in both PMSM and load dynamics respectively and hence feasibility of the proposed controller.