Modeling For Switched Reluctance Motor Torque Control Modelbased On Wavelet Neural Network

the structure of the switched reluctance motor determines the complexity of themathematical model, linear model and the quasi linear model can reflect the basic characteristicsof the motor, in order to acquire operating characteristics of the switched reluctance motor moreaccurately, it’s necessary to establish a new model combined with the theory of intelligentcontrol, based on the existing linear and quasi linear model.The paper firstly introduces the development of switched reluctance motor at home andabroad and the major research. secondly, the paper analysises the main structure and controlmethod of the switched reluctance motor, establishment of switched reluctance motor controlmodel, the relationship between the torque, speed and voltage of the motor; combined theadvantages of theory of wavelet network and the neural network, awavelet neural network fluxmodel is established,this model has a good generalization ability and it can operate in the everycontrol mode of the switched reluctance motor.in this paper, the linear model and the quasi linear model of the switched reluctance motoris isanalyzed, and the main factors affecting the motor’s speed are founded out. We limit themaximum and the minimum inductance, opening angle, turn off angle variables of the switchedreluctance motor,,we need to find out the relationship between the torque and the voltage atthe same speed,and a linear relationship between the speed and the torque.at the differentvoltage.Through the study on wavelet neural network a, a wavelet neural network model isestablished combined the advantages of them,. The training methods of the main parameters andlearning of wavelet neural network are studied, the wavelet neural network flux model isinitialized. the overall model of switched reluctance motor is established on this basis,and thesimulation research is carried out and analyzed in the SIMULINK environment, the results isthat flux graphics of the wavelet neural network and the graphics which can be calculatedthrough the experiment test data are similar.