Torque Modeling of a Segmented-Rotor SRM Using Maximum-Correntropy-Criterion-Based LSSVR for Torque Calculation of EVs

High nonlinearities of switched reluctance motor (SRM) caused by its double salient structure limit its industrial application in electric vehicles (EVs). In this article, an algorithm called maximum-correntropy-criterion-based least-squares support vector regression (MCC-LSSVR) is applied to the nonlinear modeling of a segmented-rotor SRM (SSRM). First, the mathematical model of SSRM is established. Finite element analysis (FEA) is carried out to obtain the static flux linkage and torque. Then, the intelligent algorithm MCC-LSSVR using an adaptive weight to avoid the interference of outliers is introduced. It is verified and applied to SSRM modeling. The results show that MCC-LSSVR exhibits a better performance than other intelligent algorithms. Finally, simulation and experimental validation under various modes are given to verify the accuracy and effectiveness of the MCC-LSSVR model. It is shown that the simulation and experimental results are in good agreement.