Nonlinear Fast Modeling Method of Flux Linkage and Torque for a 12/8 Switched Reluctance Motors
Xiaodong Sun, Nuonuo Wang, Yunfei Cao, Dong Guo, Ming Yao, Yueping Sun, Yefei Xiong
Abstract
This paper presents a nonlinear modeling method for the flux linkage and torque of a switched reluctance motor (SRM). The method is based on Universal Weighted Least Squares Support Vector Machine Regression (WLSSVR), combined with Entropy Method (EM) and Improved Coyote Optimization Algorithm (COA) to optimize the kernel parameters. Among them, EM can well improve the error of the general WLSSVR in the selection of sample weights, and the improved COA can prevent the system from falling into local optimum and improve the iteration speed. This SRM modeling method, which combines WLSSVR based on entropy and improved COA, can improve the modeling speed on the premise of ensuring the modeling accuracy. mode speed. Finally, the effectiveness of the algorithm is verified by experiments