Research On Control Strategy For Reducing Torque Ripple Of Switched Reluctance Motor

With advantages of simple structure, low manufacturing cost, high system reliability,high transfer efficiency of energy and wide speed adjusting range, SRM is considered asone of the ideal driving motors for the future electric vehicle industry. SRM has beeneffectively applied in aviation, mining and textile industry, etc. However, the widespreadapplication of SRM is severely limited in high controlling demand fields, for the reasons oflarge low-speed torque ripple and vibration noise caused by torque ripple. Because of thespecific double-salient structure and power supply by switching mode, the electromagneticproperty of the motor is characterized by strong nonlinear for which the precisemathematic mode of motor can’t be efficiently established. And also because of thesatisfactory effect can’t be obtained for the strong nonlinear object with traditional controlalgorithm, the designing of motor control methods for decreasing the torque ripple facesgreat difficulties. In order to reduce the low-speed torque ripple of SRM, two controlmethods are proposed in this paper as follows:(1) Current distribution control of SRM based on brain emotional learning model isproposed in the paper, in which the torque is controlled by adjusting the current indirectly.The outer-loop brain emotional learning model controller is used to transform the speeddeviation into bus bar reference current. Three-phase reference current can be obtainedfrom bus bar reference current by current distribution function. The three-phase controlsignals, which are got from the inner-loop current hysteresis control unit according to thethree-phase current deviation, smooth commutation of the motor, which reduce the torqueripple of the motor effectively.(2)SRM direct instantaneous torque control strategy based on constructed flexibleneural network is proposed, which refers to the ideas of traditional direct instantaneoustorque control to solve the strong nonlinear and high coupling of the SRM. In the improvedcontrol strategy, the outer-loop incomplete derivative fuzzy-PID is used to adjust the motorspeed, and the inner-loop flexible neural network self-adaptive PID is employed to regulatethe motor torque by using the square of the torque error as performance index function.The control system obtains a greater control effect.In the MATLAB/SIMULINK environment, simulation results prove that the twocontrol strategies can both effectively decrease the motor torque ripple. On the basis ofsimulation research, current distribution control of SRM based on brain emotional learningmodel, SRM direct instantaneous torque control method and SRM voltage chopping control method are tested in the experiment platform of SRM. The experiment resultsprove that the first two control methods are more effective than traditional voltagechopping control method for suppressing the torque ripple of SRM.