Optimization of Switched Reluctance Machine Drives Using Multi-Task Learning Approach

Alexandria Engineering Journal

Mohammad Tabrizian¹  Mojtaba Babaei²  Kasra Abolfathi³  Mohsen Alizadeh Bidgoli³ 

[1] Department of Electrical and Computer Engineering, Yadegar-e-Imam Khomeini (RAH) Shahre Rey Branch, Islamic Azad University, Tehran, Iran;Corresponding author.;Department of Electrical and Computer Engineering, Yadegar-e-Imam Khomeini (RAH) Shahre Rey Branch, Islamic Azad University, Tehran, Iran;

关键词: Switched Reluctance Motor (SRM);    Torque ripple;    Multi-objective optimization;    Optimization;    Multi-task learning;    Machine learning;   

DOI  :  

来源: DOAJ

【 摘 要 】

One of the major challenges for controlling the drive of switched reluctance machines (SRMs) is to have a proper conduction angle with the working point of the motor. This is due to the non-linear relationship of the flux-linkage with the position of the rotor. The optimization problem in SRM motors should be solved using multi-objective optimization methods because the objective functions are constantly in competition and a compromise should be established between them. In this study, we propose a multi-task learning (MTL) method to optimize this problem. The obtained results of the introduced algorithm were compared with the NSGA-II algorithm. This comparison was focused on two aspects of discipline and quality. Moreover, the covering rate of Pareto front for these two algorithms was evaluated. The accuracy of the proposed method was evaluated and the results showed that the proposed solution is efficient for the optimization problem of SRMs.

【 授权许可】

Unknown