Differential Evolution-Based Multiobjective Optimization of the Electrical Continuously Variable Transmission System

Abstract

With the increasing popularity of hybrid electric vehicles (HEVs), the demands of improving the fuel economy and energy efficiency of HEVs are enlarged. The technology electrical continuously variable transmission (E-CVT) system provides the opportunity to fulfill this requirement. However, conventional E-CVT with planetary gears suffer from frictional loss, high maintenance, and audible noise. Therefore, a series of gearless E-CVT propulsion systems are developed. Thereinto, a double-rotor single-stator brushless E-CVT having high torque density, reduced stator end winding length, and reduced copper loss was recently proposed. However, only torque of the E-CVT system is optimized by the genetic algorithm in the previous study. This paper adopts a differential evolution algorithm coupled with finite element method to optimize torque, energy efficiency, torque ripples, and multiobjective of the E-CVT system by reconfiguring the height of rotors, the areas of slots, and the height of yoke. It is simulated that by only tuning the weighting factors, the permanent magnetic machine in the E-CVT system can operate in the highest torque mode, highest energy efficiency mode, lowest torque ripples mode, and multiobjective optimization mode.