Sensitivity analysis and optimal design of a dual mechanical port bidirectional flux-modulated machine

Abstract

This paper presents an optimal design methodology of a dual-mechanical-port bidirectional flux-modulated machine for electric continuously variable transmission (E-CVT) in hybrid electrical vehicles (HEV). The key is to utilize bidirectional flux modulation effect to combine two rotors and one stator together, aiming to realize electrical and mechanical power flexible split and combination. Due to the complexity of the machine structure, conventional optimization methods using analytical model are inapplicable. Therefore, an effective and practical method which combines the genetic algorithm and finite element method (GA-FEM) is proposed to optimize the performance of the machine in this paper. Since the computational cost increases exponentially with the increasing of parameter numbers, to reduce the computational cost in the optimization, the design parameters are divided into two levels basing on a sensitivity analysis. And then the sensitive parameters are optimized using the GA-FEM coupled method. Finally, a prototype is fabricated to verify the effectiveness of the optimal design.