The State Estimation Of A Sensorless Brushless Motor Its Control Under Disturbance Uncertainties

Brushless motors are widely applied due to their high torque-speed ratio, fastdynamic response, high efficiency and stability. Typically, they can be divided into twocategories as permanent-magnet synchronous motors (PMSM) and brushless DC motors,both of which share considerable similarities in structures, working principles and controlmodes. In practice, however, either sort is influenced by problems such as disturbanceuncertainties in the design of control system. The stability of the control system underextreme environment could be improved by removing the position sensors and estimatingthe states of position and speed of the motor through software, which saves space andreduce costs meanwhile. Therefore, how to cancel the disturbance uncertainties andestimate the states in the control system of a sensorless brushless motor has become ahotspot.In this paper, a composite control strategy is proposed, taking a representativePMSMas an example and considering the disturbance uncertainties existing in applications. Thestrategy consists of a disturbanceobserver based on linear extended state observer whichcan suppress disturbances by feedforward compensation and a Lyapunov-stability-basedposition controller. A non-singular terminal sliding-mode position controller is appliedtofurther improve the performance of the control system under sinusoidal perturbations,allowing state errors to converge to zero in finite time. Besides, special forms of load torque,hinge moment, and sinusoidal torque disturbances have been taken into account. Theproposed strategy is verified via simulations, which yield good performance with respect toposition tracking, robustness to load disturbances and dealing with parametric uncertaintiesof the motor.Then a sensorless speed control strategy for a PMSMbased on an adaptive nonlinearextend state observer (ANLESO) is presented against states estimation of the controlsystem of sensorless PMSM.An extend state observer (ESO), which is simpler in structureand higher in accuracy is introduced to estimate the position and speed of the rotor. Aftercomparing linear ESO (LESO) and nonlinear ESO (NLESO) in terms of estimating the back-EMF of PMSM, the latter is chosenbecause ofa better convergence. Moreover, theconvergence characteristics of the observer in error estimation is analyzed by Lyapunovtheory. Taking both stability and steady-state error into consideration, an adaptiveNLESOwhich adjusts the parameters of NLESO adaptively to a compromised value isproposed, whose performance is then proved through simulations and experiments.Finally, an experiment system based on the processor F28335MCU is built, where theapproach designed above is tested. The results demonstrate that the NLESO is superior tothe conventional slide-mode state estimating algorithm in that it is smaller in buffeting andfree of delay. That lay a foundation for further theoretical research as well as engineeringimplementation.