Research On Torque Ripple Reduction For Switched Reluctance Motor Based On Intelligent Control

In the 1980s, Switched Reluctance Drive (SRD) has developed as a new drivesystem rapidly, which is made of Switched Reluctance Motor (SRM), power electronic switched circuit, driving controller and detector. SRD has been widely used in household appliances, general industrial, servo speed control because of its broad scope of the speed and superior performance. But there are some obvious shortcomings: the torque ripple and noise, SRM precise mathematical model difficult to establish makes greater difficult control of SRD.In order to reduce torque ripple of SRM, domestic and foreign scholars have put forth a variety of control strategies, such as neural network control, optimizing the off-angle control, the torque distribution function control, micro-step moment control. This paper analyses the reasons of torque ripple of SRM, focus on two kinds of intelligent control strategy, the principal tasks are as follows:(1) Based on the four-phase (8/6) SRM linear mathematical model, the simulation of SRD is established and analysised. The linear and quasi-linear, non-linear mathematical model of SRM is analysised. In MATLAB, the simulation of SRD is established under current chopper control and voltage chopper control. The inductance, a phase current, torque and speed of SRD is analysised.(2) A fuzzy-logic-based turn-off angle optimization for torque ripple reduction is proposed. Based on fuzzy-logic turn-off angle optimization, the simulation of SRD is established under voltage chopper control. The real-time optimization to turn-off angle of SRM is carried out by using fuzzy logic controller to reduce torque ripple. The system has not torque feedback so that the system is simple and reliable, easy to implement.(3) The control based BP neural network for torque ripple reduction is proposed. Based on determining the magnetic properties for accurate data on the basis of samples and Levenberg-Marqvardt algorithm, the simulation of SRD is established under voltage chopper control. BP neural network with non-linear mapping capabilities is very suitable to reducing torque ripple. This network model is of fast training convergence,generalization and small network scale handy for real-time control.(4) The experiment of SRD is carried out. The design and manufacture of the hardware and the programming of the software for SRD was carried out based on the DSP.