Control System Of Switched Reluctance Motor Drives

The current vigorous promotion of energy saving and emission reduction has led to the elimination of inefficient motor drive systems.At the same time,the rapid development of the national economy has brought about the upgrading of the installed capacity.Under this background,the switched reluctance motor,which has the advantages of high efficiency,sturdy and low cost,has a good opportunity for development.Dynamic performance is a precondition for the wide application of motor drive systems,and position sensorless control is the guarantee for motor drives to broaden their application.In view of the high performance speed control and sensorless control of switched reluctance motor,the relevant analysis and research are studied in this paper.The stable switching between three basic control methods of switched reluctance motor is studied.First,the traditional three-stage speed control strategy of switched reluctance motor is analyzed,and the jitter problem of switching is pointed out.And then,by analyzing the electrical characteristics of the chopped current control and chopped voltage control,a smooth switching control strategy from chopped current control to chopped voltage control based on the area equivalent principle is proposed,and on this basis,a phase-by-phase switching strategy is further proposed;and by analyzing the advantages and disadvantages of chopped voltage control and angular position control,the complementary control strategy of chopped voltage control and angular position control are proposed.Finally the control flows of switching control strategy and combining control strategy are given.The sensorless control of switched reluctance motor based on neural networks is studied.The principle of detecting rotor position of switched reluctance motor based on neural network and the training methods is analyzed,and training step of neural network for position detection is given.The obtained rotor position information is incompleted as the neural network detection cannot be performed when the windings are non-conducting,and needs to be completed.By analyzing the results of the neural network position detection of each phase winding,multi-neural networks completed rotor position information calculation method is proposed.In order to simplify the calculation and improve versatility,single-neural network completed rotor position information calculation method is proposed.This single-neural network method detects rotor position of one phase by neural network,and then selects a reference point which is accurately detected by the neutral network.With this reference point,position correction,speed calculation and integration reset are completed,and then the completed rotor position information obtain.A simulation model is built to verify the effectiveness of the scheme.Control strategies are verified on the prototype experimental platform.The experiment results show: The switching control strategy proposed in this paper can complete the stable switching from chopped current control to chopped voltage control.The combining control of chopped voltage control and angular position control has better ability to resist the torque disturbance.The neural network can detect rotor position under chopped current control,chopped voltage control and variable-angle chopped voltage control.The proposed single-neural network method for calculating the completed rotor position information can accurately calculate the rotor position,and then complete the phase commutation and achieve sensorless control of switched reluctance motor.