Instantaneous Torque Control of Switched Reluctance Motors Based on RBF Neural Network

XIA Chang-liang,CHEN Zi-ran,LI Bin
DOI: https://doi.org/10.3321/j.issn:0258-8013.2006.19.023
2006-01-01
Abstract:The inherent simplicity,ruggedness,great efficiency and low cost of a switched reluctance motor(SRM)make it a potential candidate for adjustable speed application.However,owing to its nonlinear electromagnetism characteristic,the torque ripple of motor is much more severe than other drive system.For that reason,the torque ripple minimization in SRM has obtained great attention.To solve this problem,this paper presents a scheme of instantaneous torque control for switched reluctance motors based on radial basis function(RBF)neural network.The off-line training of RBF network is based on the data acquired by dynamic modeling.After trained from the optimized current profiles in different speed and torque,RBF network could be able to achieve the nonlinear mapping between torque,position and current.The approach of transient current tracing control is implemented in this system to adjust the current in order to control the torque of the motor.This method which has taken advantages of RBF neural network in approximation,generalization and calculation is simple and needs no on-line training.The results of experiment prove that this method could reduce SRM's torque ripple effectively and is adaptive to speed variety.
What problem does this paper attempt to address?