Design, simulation and optimisation of a novel low ripple outer-rotor switched reluctance machine for variable speed application

• Published in: International journal of electronics Vol. 111; no. 4; pp. 579 - 598
• Main Authors: Khodadadeh, Omid, CheshmehBeigi, Hassan Moradi, Mousavi, Mohammad Hossein
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor & Francis 02-04-2024; Taylor & Francis LLC
• Subjects: Design optimization; Diameters; Electric bicycles; Electric vehicles; FEM analysis; Finite element method; genetic algorithm; Genetic algorithms; Outer-rotor switched reluctance machine; Parameter sensitivity; Poles; Reluctance machinery; Ripples; Rotors; Sensitivity analysis; Stators; Torque; torque ripple reduction; Wind turbines
• ISSN: 0020-7217; 1362-3060
• DOI: 10.1080/00207217.2022.2164083

Remarkable characteristics like simple structure, high starting torque, simple maintenance, good efficiency at all speeds and cost-effectiveness have made switched reluctance machines (SRMs) suitable candidates for electric cars, electric bicycles and variable speed applications (VSAs) such as wind turbines. In this paper, a three-phase Outer-Rotor SRM (ORSRM) with 6 stator poles and 10 rotor poles (6/10) for VSA has been investigated. Unlike conventional SRM structures such as 6/4 and 12/8, an SRM model with higher number of rotor poles can reduce torque ripple much more effective than conventional SRMs. First, the SRM design is done accurately, then, by changing some parameters, the sensitivity analysis is performed on torque ripple and average torque. Finally, the optimisation process on the geometry and structural dimensions of the SRM like rotor pole arc, stator pole arc, rotor yoke thickness, rotor pole height, rotor diameter and the number of turns per phase are executed to minimise the torque ripple and maximise the average torque via genetic algorithm (GA). The initial design, optimisation and performance of the proposed model are validated via a 2-D finite element method (FEM) technique by ANSYS-Maxwell.