Unscented Kalman Filter Based Rotor Position Estimation And Variable-Structure Control For PM Brushless DC Motor

Owing to its intrinsic advantages such as high mechanical power densityelectrical effectiveness, small size, low weight and easy to maintain, permanentmagnet brushless DC motor (BLDCM) has been widely used in every field ofindustrialcontrolandhas abroadprospectinapplication.TraditionalBLDCMcontrolobtains accurate rotor position signals through position sensors for commutation andspeed regulation. However, the existence of position sensors brings many adverseeffects on the cost, size and stability of motor. Consequently, position sensorlesscontrol of BLDCM is of great practical significance and has become one of the hotspots in this field. Meanwhile, motor's parameters and load disturbance will affect itsperformance,andcontrollersalsolimitcontrolaccuracy.In this paper, a position sensorless control method based on Unscented KalmanFiltering (UKF) for BLDCM is presented, in which UKF is employed to estimate therotor position and speed. This paper designs a variable-structure control systemimprovedbyKalmanFilter.Firstly,thepaperintroduces thebasicprinciples andcontrol methods ofBLDCM,and compares the common rotor position detection methods. Then this paper putsforward a motor state observation method with UKF. Anonlinear state equation is setup through the basic equations of BLDCM, and the rotor position and speed signalsare observed with UKF algorithm. With this method, the adverse influences of modelerror, measurement error, outside interference and load changes can be eliminated.This paper also analyzes the effects of the motor parameters statistical characteristicsonobservationaccuracy.Secondly,thispaperintroducesthevariable-structurecontrolanditschattering.Avariable-structure controller based on reaching law is employed to speed controlsystem of BLDCM. And Kalman Filter improves the controller in the situation ofnoise disturbance. The simulation results show that this method highlights therobustness of variable-structure control for internal parameters and externaldisturbance. And the filter can effectively weaken controller's chattering and improvetheprecisionofsystem.At last, this paper describes the hardware and software of DSP controller for BLDCM. The experiment results and analysis are provided. And the results show theeffectivenessofthepresentedmethod.