Switched Reluctance Motors Reduce Torque Ripple

Switched reluctance motor has many advantages including simple structure,reliable performance,high efficiency,low cost,stable operation in extended speed range,and no need for rare-earth materials.However,there are some defects such as torque ripple and its noise,which restrict the application of switched reluctance drive.Therefore,reducing torque ripple is not only a hot spot but also a difficult point in the research of switched reluctance motor and its control system.Firstly,this paper introduces the research status of torque ripple of switched reluctance motor,and analyzes the working principle,mathematical model,control methods,modeling,and the structure of switched reluctance motor drive.Aiming at the problem that the electromagnetic property is highly non-linear and difficult to be accurately expressed by analytical expression,this paper builds the SRM model by using BP neural network with good nonlinear mapping ability.Based on the traditional structure of two input variables network,a third input variable reflecting the nonlinearity of the system is added by using the pretreatment method.The neural network model of switched reluctance motor with the pretreatment method is established,including flux linkage model,torque model,inductance model and current model.Compared with the traditional neural network model,the proposed model has the characteristics of smaller network size and higher model accuracy,which is beneficial to real time online control of switched reluctance motor drive.In order to reduce the torque ripple of switched reluctance motor,an instantaneous torque control scheme of switched reluctance motor based on online torque distribution function is proposed in this paper.In view of the fact that the actual torque existing is difficult to follow a given torque command in real time in the traditional torque distribution function,an online torque distribution function is proposed in this paper.Deviation is given between the given torque and the sum of instantaneous torque,and the deviation through the PID regulator output torque compensation amount.The online torque distribution function adjusts the torque instructions of each phase online in real time according to the given torque,rotor position angle and torque compensation amount,and realizes the actual torque to follow the given torque in real time.In order to improve the torque compensation effect and optimize the torque command output by the online torque distribution function,this paper uses BP neural network to design a PID with adaptive control parameters as a torque compensator.The motor speed ripple is caused by the torque ripple.The size of the speed ripple reflects the situation of torque ripple.In this paper,a speed ripple detection circuit using frequency-voltage conversion technology is designed to observe the torque ripple.The circuit uses the frequency voltage conversion chip to convert the pulse signal with a certain frequency into an analog voltage value that is linearly proportional to the pulse frequency and reflects the motor speed.By measuring the output voltage changes to detect the motor speed ripple,and then observe the torque ripple situation.