Cost – performance – size optimization for automotive induction machines: A fast and accurate FEM, analytical model and optimization mixed procedure

Abstract

Purpose

Seeks to posit a new procedure to optimize the geometry of an electric motor for hybrid vehicles (HV), based on an optimization model implementing the complex specifications, environment‐sizing constraints and vehicle‐operating points.

Design/methodology/approach

In this work, induction machines (IM) were studied for automotive traction applications, more specifically, for HV. The necessary models and methods to analyse and design such electric machines were developed. A limited number of finite element computations was used to establish a non‐linear electromagnetic model of the machine. This model was then adapted to study the sensitivity of the design for the most significant geometric variations. Thus, one can easily adapt a machine for new sizing requirements.

Findings

This modelling methodology was extended to take into account a great number of parametric variations. This model was used for constrained optimisation on the geometry of a machine. To achieve this, a new “reset model optimisation” method was proposed, combining fast computations, precision and simplicity. This methodology was used to design an IM with HV‐sizing considerations.

Practical implications

The methodology is based on an improved equivalent circuit of the IM where the FEM accuracy (saturation) is combined with the potential of analytical modelling (simple and quick).

Originality/value

This half‐analytical, half‐FEM methodology is well adapted for research projects where the specifications are frequently reviewed.

Citation

Pugsley, G., Chillet, C. and Fonseca, A. (2006), "Cost – performance – size optimization for automotive induction machines: A fast and accurate FEM, analytical model and optimization mixed procedure", COMPEL - The international journal for computation and mathematics in electrical and electronic engineering, Vol. 25 No. 2, pp. 297-308. https://doi.org/10.1108/03321640610648989