Airgap induction optimisation in permanent magnet synchronous machine

Numerical field calculations and mathematical programming methods are applied to solve constrained nonlinear optimisation problems in magnetostatic systems. The authors have used four different methods that seem to be adequate to solve optimisation problems in magnetostatic devices. These methods are based on the analysis of the sensitivity of an objective versus a design parameter; and they are used to solve the problems of obtaining a specific airgap induction distribution in a tetrapolar synchronous machine. Two solutions are proposed to modify the airgap induction shape: optimisation of the permanent magnet positions and optimisation of their remnant flux densities. These approaches allow approximation of a sinusoidal shape. The authors show the importance of the choice of the different optimisation parameters to insure a reliable use of the numerical methods